Search is not available for this dataset
article
stringlengths 4.36k
149k
| summary
stringlengths 32
3.35k
| section_headings
listlengths 1
91
| keywords
listlengths 0
141
| year
stringclasses 13
values | title
stringlengths 20
281
|
|---|---|---|---|---|---|
Dehydroepiandrosterone sulphate ( DHEAS ) is the most abundant circulating steroid secreted by adrenal glands—yet its function is unknown . Its serum concentration declines significantly with increasing age , which has led to speculation that a relative DHEAS deficiency may contribute to the development of common age-related diseases or diminished longevity . We conducted a meta-analysis of genome-wide association data with 14 , 846 individuals and identified eight independent common SNPs associated with serum DHEAS concentrations . Genes at or near the identified loci include ZKSCAN5 ( rs11761528; p = 3 . 15×10−36 ) , SULT2A1 ( rs2637125; p = 2 . 61×10−19 ) , ARPC1A ( rs740160; p = 1 . 56×10−16 ) , TRIM4 ( rs17277546; p = 4 . 50×10−11 ) , BMF ( rs7181230; p = 5 . 44×10−11 ) , HHEX ( rs2497306; p = 4 . 64×10−9 ) , BCL2L11 ( rs6738028; p = 1 . 72×10−8 ) , and CYP2C9 ( rs2185570; p = 2 . 29×10−8 ) . These genes are associated with type 2 diabetes , lymphoma , actin filament assembly , drug and xenobiotic metabolism , and zinc finger proteins . Several SNPs were associated with changes in gene expression levels , and the related genes are connected to biological pathways linking DHEAS with ageing . This study provides much needed insight into the function of DHEAS .
Dehydroepiandrosterone sulphate ( DHEAS ) , mainly secreted by the adrenal gland , is the most abundant circulating steroid in humans . It acts as an inactive precursor which is converted initially into DHEA and thereafter into active androgens and estrogens in peripheral target tissues [1] . In humans the serum concentration of circulating DHEAS is 100- to 500-fold or 1000 to 10 , 000 higher than that of testosterone and estradiol respectively . Unlike DHEA , which is swiftly cleared from the circulation and shows diurnal variation , serum DHEAS concentrations are stable and facilitate accurate measurement and diagnosis of pathology [2] . DHEAS is distinct from the other major adrenal steroids ( cortisol and aldosterone ) in showing a significant physiological decline after the age of 25 and diminishes about 95% by the age of 85 years [3] . This age-related decline has led to speculation that a relative DHEAS deficiency may contribute to the development of common age-related diseases or diminished longevity [4] , [5] . Low DHEAS concentrations are possibly associated with increased insulin resistance [6] , [7] and hypertension [8] , but not with incident metabolic syndrome [9] . It is strongly associated with osteoporosis in women [10] , [11] but not in men [12] . Concurrent change in DHEAS tracks with declines in gait speed , modified mini-mental state examination score ( 3MSE ) , and digit symbol substitution test ( DSST ) in very old women but not in men [13] . Low circulating DHEAS is also strongly associated with cardiovascular disease and mortality in men [14]–[18] but not in women [19] . A recent 15-year follow-up study showed that DHEAS was negatively related to all-cause , all cancers , and other medical mortality , whereas high DHEAS concentrations were protective [20] . This has led to its widespread and uncontrolled use as a controversial anti-ageing and sexual performance supplement in the USA and other western countries without any clear data about efficacy , potential risks or benefits [21] . Despite these observations , the physiological function of DHEAS and its importance in maintaining health are poorly understood . Although previous twin [22] , [23] and family-based studies [24] , [25] have shown that there is a substantial genetic effect with a heritability estimate of 60% [22] , no specific genes regulating serum DHEAS concentration in healthy individuals have been identified to date . Therefore , the current study meta-analyzed the results of genome-wide association studies ( GWAS ) performed in a total of 14 , 846 individuals from seven cohorts to identify common genetic variants associated with serum DHEAS concentrations . The findings not only advance understanding of how serum DHEAS concentration is regulated by genes but also provide clues as to its mechanism of action as well as Mendelian randomisation principles [26] .
We carried out a meta-analysis of 8 , 565 women and 6 , 281 men of European origin from collaborating studies: TwinsUK ( n = 4 , 906 ) , Framingham Heart Study ( FHS ) ( n = 3 , 183 ) , SHIP ( n = 1 , 832 ) , Rotterdam Study ( RS1 ) ( n = 1 , 597 ) , InCHIANTI ( n = 1 , 182 ) , Health ABC ( n = 1 , 222 ) , and GOOD ( n = 924 ) . Serum samples were collected either after overnight fasting or non-fasting in each cohort and DHEAS was measured by either immunoassay or liquid chromatography tandem mass spectrometry ( LC-MS/MS ) methods ( Table 1 ) . Mean age differed across the cohorts from 19 to 74 years in men and 50 to 74 years in women and corresponding mean DHEAS concentrations varied from 1 . 20 to 7 . 05 µmol/L ( Table 1 ) . Each cohort performed GWA tests for log transformed DHEAS on ∼2 . 5 million imputed single nucleotide polymorphisms ( SNPs ) in men and women separately with adjustment for age , and additionally for age and sex for those cohorts who had data in both men and women . Then Z-scores from each cohort were pooled for the meta-analysis at each SNP . In all our individual GWAS , λGC , which is defined as the median χ2 ( 1 degree of freedom ) association statistic across SNPs divided by its theoretical median under the null distribution [27] , ranged from 0 . 984 to 1 . 023 , indicating that there was no population stratification or it was very minor . Further , we corrected for population stratification by applying the genomic control method [27]; the λGC in the meta-analysis is 1 . 017 . In addition , the effect direction was consistent across all the cohorts and there is no between-study heterogeneity as indicated by I2 ranging between 0 and 0 . 12 ( Table 2 ) . We found 44 SNPs were associated with serum DHEAS concentrations in men at conventional genome-wide significance ( p<5×10−8 ) , which are all located on chromosome 7q22 . 1 ( Figure 1B; Table S1 ) . All these SNPs except for three were significant in women ( Figure 1A; Table S1 ) . In addition , 19 SNPs located on chromosome 19q13 . 3 were found in women to be associated with serum DHEAS concentrations with p<5×10−8 . In the sex-combined meta-analysis , the significance became stronger for all these SNPs ( Figure 1C; Table S1 ) . Further , we found 8 SNPs located on chromosome 10q23 . 33 which represents two regions more than 2 MB apart , 12 SNPs on chromosome 15q15 . 1 , and in addition , 4 SNPs on chromosome 19q13 . 3 were associated with serum DHEAS concentrations with p<5×10−8 . Together we found a total of 87 SNPs associated with serum DHEAS concentrations with p<5×10−8 , representing five chromosomal regions of less than 1 Mb each ( Table S1 ) . The most significantly associated SNPs in each of these five regions are presented in Table 2 . The minor allele of rs11761528 ( p = 3 . 15×10−36 ) on chromosome 7q22 . 1 , rs2637125 ( p = 2 . 61×10−19 ) on chromosome 19q13 . 3 , and rs2497306 ( p = 4 . 6×10−9 ) and rs2185570 ( p = 2 . 29×10−8 ) on chromosome 10q22 . 33 ( more than 2 Mb apart ) , were negatively associated with DHEAS concentrations . In comparison , the minor allele of rs7181230 ( p = 5 . 44×10−11 ) on chromosome 15q15 . 1 was positively associated with serum DHEAS concentrations . Based on the HapMap3 release2 CEU data , the significant 87 SNPs from within the five regions have low pair-wise r2 , indicating potentially multiple independent signals . To verify this , we performed a conditional meta-analysis with adjustment for the five most significant SNPs plus age and sex in each cohort . After this adjustment , all other SNPs on chromosome 10 , 15 , and 19 became non-significant ( Figure 1D ) . However , on chromosome 7 , we found two independent signals; one defined by rs11761528 and a second located 370 kb upstream in the 3′ UTR of the TRIM4 and CYP3A43 genes ( rs17277546 , p = 4 . 50×10−11 ) . Furthermore , we identified two additional significant loci associated with DHEAS , one on chromosome 2q13 ( rs6738028 , p = 1 . 72×10−8 ) , and another on chromosome 7 within the ARPC1A gene ( rs740160 located 161 kb downstream of rs11761528 , p = 1 . 56×10−16 ) ( Table 2; Figure 1D ) . In total , we found eight independent SNPs associated with serum DHEAS concentrations at conventional genome-wide significant level ( p<5×10−8 ) ( Table 2 ) . The effect was consistently in the same direction across all cohorts ( Table 2 ) . No heterogeneity among cohorts was observed ( Table 2 ) . These SNPs together explained ∼4% of the total and ∼7% of genetic variance of serum DHEAS concentrations ( based on TwinsUK data ) . To further look at whether the magnitude of these genetic association varies with age , we carried out an interaction analysis between age and each of these 8 SNPs on serum DHEAS concentrations by including an interaction term of age×SNP in the linear regression model in each cohort and then meta-analyzed the results . We found that there was no significant interaction between age and each of these SNPs ( all p values≥0 . 05 ) . The genes at , or near the identified SNPs , include BCL2L11 on chromosome 2 , ZKSCAN5 , ARPC1A , TRIM4 and CYP3A43 on chromosome 7 , HHEX and CYP2C9 on chromosome 10 , BMF on chromosome 15 , and SULT2A1 on chromosome 19 ( Figure 2 ) . To explore the potentially functional impacts and likely genetic mechanisms , we used two resources: Genome-wide expression data from the Multiple Tissue Human Expression Resource ( MuTHER ) [28] ( http://www . muther . ac . uk/ ) based on ∼777 unselected UK twins sampled for skin , adipose tissue , and lymphoblastoid cell lines ( LCLs ) ( more details in Text S1 ) ; and published gene expression data in human liver [29] . We found that 3 DHEAS-associated SNPs were clearly associated with the related gene expression levels in at least one tissue after accounting for multiple testing ( Table 3 ) . These specific transcript associations provide further evidence for the likely functional gene at each locus . Further , we carried out gene ontology and pathway analyses using a gene set enrichment analysis ( GSEA ) approach in MAGENTA [30] which consists of four main steps: First , DNA variants , e . g . SNP , are mapped onto genes . Second , each gene is assigned a gene association score that is a function of its regional SNP association p-values . Third , confounding effects on gene association scores are identified and corrected for , without requiring genotype data . Fourth , a GSEA-like statistical test is applied to predefined biologically relevant gene sets to determine whether any of the gene sets are enriched for highly ranked gene association scores compared to randomly sampled gene sets of identical size from the genome . More details of these four steps are described in the method section . In this analysis , we identified three pathways which passed our significance threshold ( false discovery rate ( FDR ) <0 . 05 ) ; xenobiotic metabolism with FDR = 0 . 001 ( pathway database: KEGG and Ingenuity ) , retinoid X receptor ( RXR ) function with FDR = 0 . 003 ( pathway database: Ingenuity ) , and linoleic acid metabolism with FDR = 0 . 02 ( pathway database: KEGG ) ( Figure S1 ) . The top significant genes with p<5 . 0×10−8 include CYP3A4 , CYP3A43 , CYP3A5 , and CYP3A7 on chromosome 7 , and CYP2C8 and CYP2C9 on chromosome 10 for all three pathways , and SULT2A1 for RXR pathway . The best index SNPs are rs17277546 for CYP3A4 and CYP3A43 , rs4646450 for CYP3A5 and CYP3A7 , rs2185570 for CYP2C9 , rs11572169 for CYP2C8 , and rs2637125 for SULT2A1 . The full list of the genes in each of the three pathways and the best index SNPs for each gene are listed in Table S2 . Three SNPs – rs17277546 , rs2185570 , and rs2637125 are the DHEAS-associated SNPs found in our meta-analysis . Both rs4646450 and rs11572169 were associated with DHEAS with p values of 8 . 8×10−17 and 4 . 8×10−8 , respectively , but become non-significant in the conditional meta-analysis because rs4646450 is in linkage disequilibrium ( LD , r2 = 0 . 429 ) with rs11761528 which is the most significant DHEAS-associated SNP while rs11572169 is in high LD ( r2 = 0 . 778 ) with rs2185570 . Intriguingly , two pathways - xenobiotic metabolism and linoleic acid metabolism , have been linked to ageing in model organisms [31]–[36] .
This is the first meta-analysis of GWA studies on serum DHEAS in 14 , 846 Caucasian subjects . We found 8 common SNPs that implicate nearby genes that are independently associated with serum DHEAS concentrations and provide clues to its role in ageing . Among the genes identified , SULT2A1 , a specialized sulpho-transferase which converts DHEA to DHEAS in the adrenal cortex , is an obvious candidate gene [3] . SULT2A1 has a broad substrate specificity , which includes conversion of pregnenolone , 17α-hydroxypregnenolone , and DHEA to their respective sulphated products [37] . Once sulphated by SULT2A1 , pregnenolone and 17α-hydroxypregnenolone are no longer available as substrates for HSD3B2 . Therefore , SULT2A1 sulphation of pregnenolone and 17α-hydroxypregnenolone removes these substrates from the mineralocorticoid and glucocorticoid biosynthetic pathways . This suggests that high levels of SULT2A1 would ensure the formation of DHEAS [3] . Variation in SULT2A1 expression has previously been associated with variation of DHEAS concentration [38] . The SULT2A1 gene is predominantly expressed in the adrenal cortex and to a lesser extent in the liver . We found that rs2547231 ( p = 1 . 76×10−17 ) , located 12 kb downstream of SULT2A1 , was strongly associated with expression levels of SULT2A1 in human liver tissues . Although this SNP is not the most strongly associated with serum DHEAS , it is itself in strong LD with the most significant SNP rs2637125 ( r2 = 0 . 658 ) . However , we did not find a significant association with SULT2A1 expression levels in LCL , skin , and adipose tissues , suggesting a tissue specific effect . The SULT2B1b is also reported to play a role in sulphation of DHEA , but in comparison , the strongest signal from that genomic region was rs10417472 with a p = 0 . 06 . In contrast , enzymatic removal of the sulphate group from DHEAS to form DHEA is performed by steroid sulphatase gene ( STS ) , but that gene is on the X chromosome and so was not assessed in this meta-analysis . CYP2C9 is an important cytochrome P450 enzyme , accounts for approximately 17–20% of the total P450 content in human liver , and catalyzes many reactions involved in drug metabolism as well as synthesis of cholesterol , steroids and other lipids [39] . We found that rs2185570 located in the CYP2C9 gene region is associated with serum DHEAS concentrations . This SNP is in strong LD with rs4086116 and rs4917639 ( r2 = 0 . 67 for both ) which were found to be associated with acenocoumarol [40] and warfarin maintenance dosage [41] respectively in recent GWAS . Two other cytochrome P450 enzymes – CYP11A1 and CYP17A1 , are two important enzymes which are required in the synthesis of DHEAS in the adrenal gland [3] , however , the strongest signals in the genomic region were rs2930306 with p = 0 . 29 for CYP11A1 and rs4919686 with p = 0 . 04 for CYP17A1 . The decline in serum DHEAS concentrations with increasing age has been proposed as a putative biomarker of ageing [21] . We found that two putative ageing genes – BCL2L11 and BMF [42] are associated with serum DHEAS concentrations . Both of them encode proteins which belong to the BCL2 family and act as anti- or pro-apoptotic regulators that are involved in a wide variety of cellular activities . BCL2L11 has been implicated in chronic lymphocytic leukaemia ( rs17483466 , P = 2 . 36×10−10 ) [43] and follicular lymphoma ( rs3789068 , P for trend = 0 . 0004 ) [44] . The DHEAS-associated SNP rs6738028 is not however one of the same SNPs associated with lymphocytic leukaemia and follicular lymphoma nor is it in LD with them . Nevertheless , rs6738028 is strongly associated with BCL2L11 gene expression levels in both LCL and adipose tissues , suggesting its putative regulatory role . There is relatively little data on the human BMF gene or the protein product , but Bmf−/− mice show altered immune and hematopoietic phenotypes as well as defects in uterovaginal development . However , we were not able to detect any association between rs7181230 and the expression levels of BMF in the tissues we studied . HHEX encodes a member of the homeobox family of transcription factors , many of which are involved in developmental processes . This gene has been found to be associated with type 2 diabetes by several recent GWAS [45]–[51] . The risk alleles of the diabetes-associated SNPs rs1111875 and rs5015480 are associated with low serum DHEAS concentrations although the p values ( p = 0 . 0009 for both SNPs ) didn't reach to the GWAS significance level . This is consistent with the observation in which the low serum DHEAS concentrations were associated with insulin resistance [6] , [7] . The identified DHEAS-associated SNP rs2497306 is in moderate LD with rs1111875 and rs5015480 ( r2 = 0 . 38 ) . And the major allele of rs2497306 is associated with increasing serum DHEAS concentrations . The reason for the observed association is unknown . Studies showed that insulin infusion increases the metabolic clearance of DHEA and DHEAS [52] , [53] , resulting in decreased DHEA and DHEAS concentrations , and DHEA administration significantly enhances insulin sensitivity attenuating the age-related decline in glucose tolerance [54] , partly explaining why the diabetes-associated gene is also associated with DHEAS . Interestingly , HHEX null mice show cardiovascular , endocrine , liver , muscle , nervous system , and metabolic phenotypes , suggesting extensive multisystem roles for the protein product of this gene . The findings could help dissect causal pathways for the observed associations between DHEAS , insulin resistance , age-related decline in glucose tolerance [54] , and other age related phenotypes [55] . Three identified DHEAS-associated SNPs on chromosome 7 ( Figure S2 ) , which were independent , and 161 kb downstream ( rs740160 ) and 370 kb upstream ( rs17277546 ) apart from rs11761528 which is located in the middle of the region , are located in four genes - ZKSCAN5 , ARPC1A , and TRIM4/CYP3A43 . ZKSCAN5 encodes a zinc finger protein of the Kruppel family and is expressed ubiquitously in adult and fetal tissues with the strongest expression in testis [56] . rs11761528 is located in the intron of the ZKSCAN5 gene . It is the strongest signal we found and explains 1% of the total variance of serum DHEAS concentration alone . ARPC1A encodes one of seven subunits of the human Arp2/3 protein complex which has been implicated in actin polymerization and filament assembly in cells [57] . TRIM4 encodes a member of the tripartite motif ( TRIM ) family whereas CYP3A43 is another cytochrome P450 enzyme . The potential mechanisms for the association are unknown , but we found that rs17277546 is strongly associated with expression levels of TRIM4 not CYP3A43 , suggesting TRIM4 is the possible candidate for DHEAS . However , rs17277546 is the best index SNP for both CYP3A43 and CYP3A4 genes in the pathway analysis , indicating a fine mapping in this region is needed to reveal the potential mechanism for the association . Further , the region harbours many other genes including CYP3A7 which has been reported to increase the clearance of DHEA and DHEAS [58] and a common haplotype polymorphism in the gene has been associated with DHEAS [59] , [60] . However , none of the DHEAS-associated SNPs are associated with its expression levels in the tissues we studied , and the best index SNP rs4646450 for CYP3A7 found in our pathway analysis is in LD with rs11761528 and become non-significant in the conditional analysis . In the pathway analysis , two DHEAS-associated SNPs ( rs2185570 and rs17277546 ) were contained in all three pathways we found and one SNP ( rs2637125 ) was contained in the RXR function pathway . Intriguingly , components of the xenobiotic metabolism pathway have been linked to ageing in model organisms , for example , age-associated changes in expression of genes involved in xenobiotic metabolism have been identified in rats [31] , [32] , up-regulation of xenobiotic detoxification genes has been observed in long-lived mutant mice [33] , and adrenal xenobiotic-metabolizing activities increase with ageing in guinea pigs [34] . Furthermore , linoleic acid metabolism has also been linked to changes with ageing in rat cardiac muscle [35] and in human skin fibroblasts [36] . Taken together , these findings suggest that molecular pathways involved in ageing and longevity may also underlie DHEAS regulation , suggesting shared genetic components in both processes and corroborating a role for DHEAS as a marker of biological ageing . In summary , this first GWAS identified eight independent SNPs associated with serum DHEAS concentrations . The related genes have various associations with steroid hormone metabolism , co-morbidities of ageing including type 2 diabetes , lymphoma , actin filament assembly , drug and xenobiotic metabolism , and zinc fingers - suggesting a wider functional role for DHEAS than previously thought .
Seven study samples contributed to this meta-analysis of GWA studies on serum DHEAS concentrations , comprising a total of 14 , 846 men and women of Caucasian origin . The consortium was made up of populations from TwinsUK ( n = 4 , 906 ) , Framingham Heart Study ( FHS ) ( n = 3 , 183 ) , SHIP ( n = 1 , 832 ) , Rotterdam Study ( RS1 ) ( n = 1 , 597 ) , InCHIANTI ( n = 1 , 182 ) , Health ABC ( n = 1 , 222 ) , and GOOD ( n = 924 ) . Full details can be found in Text S1 . Blood samples were collected from each of the study participants either after overnight fasting or non-fasting and the serum levels of DHEAS were measured by either immunoassay or liquid chromatography tandem mass spectrometry ( LC-MS/MS ) methods ( Text S1 ) . Because the distribution of the serum DHEAS levels was skewed , we log transformed the concentrations and the transformed data used in the subsequent analysis . Seven study populations were genotyped using a variety of genotyping platforms including Illumina ( HumanHap 317k , 550k , 610k , 1M-Duo BeadChip ) and Affymetrix ( array 500K , 6 . 0 ) . Each cohort followed a strict quality control on the genotyping data . More details on the quality control and filtering criteria can be found in Text S1 . In order to increase genomic coverage and allow the evaluation of the same SNPs across as many study populations as possible , each study imputed genotype data based on the HapMap CEU Build 36 . Algorithms were used to infer unobserved genotypes in a probabilistic manner in either MACH ( http://www . sph . umich . edu/csg/abecasis/MACH ) , or IMPUTE [61] . We exclude non-genotyped SNPs with an imputation quality score <0 . 2 and SNPs with allele frequency <0 . 01 from meta-analysis . Each study performed genome-wide association testing for serum concentrations of DHEAS across approximately 2 . 5 million SNPs under an additive genetic model separately in men and women ( Text S1 ) . The analyses were adjusted for age . In addition , the association testing was performed in the combined men and women data with adjustment for age and sex . Studies used PLINK , GenABEL , SNPTEST , QUICKTEST , or MERLIN for the association testing . The summary results from each cohort were meta-analyzed by Z-score pooling method implemented in Metal ( http://www . sph . umich . edu/csg/abecasis/metal/ ) . We chose this method to minimize the impact of the different assays used for serum DHEAS measurements . Specifically , for each study , we converted the two-sided P value after adjustment for population stratification by the genomic control method to a Z statistic that was signed to reflect the direction of the association given the reference allele . Each Z score was then weighted; the squared weights were chosen to sum to 1 , and each sample-specific weight was proportional to the square root of the effective number of individuals in the sample . We summed the weighted Z statistics across studies and converted the summary Z score to a two-sided P value . We also used I2 index to assess between-study heterogeneity and the inverse variance weighted method to estimate the effect size . Genome-wide significance was defined as p<5×10−8 . The association between the DHEAS-associated SNPs and the related gene expression levels in MuTHER data were examined by mixed linear regression modelling which takes both family structure and batch effects into account . The significance was defined as p<0 . 006 after accounting for multiple testing ( Bonferroni method , correcting 9 independent tests ) . All studies were approved by local ethics committees and all participants provided written informed consent as stated in Text S1 .
|
Dehydroepiandrosterone sulphate ( DHEAS ) , mainly secreted by the adrenal gland , is the most abundant circulating steroid in humans . It shows a significant physiological decline after the age of 25 and diminishes about 95% by the age of 85 years , which has led to speculation that a relative DHEAS deficiency may contribute to the development of common age-related diseases or diminished longevity . Twin- and family-based studies have shown that there is a substantial genetic effect with heritability estimate of 60% , but no specific genes regulating serum DHEAS concentration have been identified to date . Here we take advantage of recent technical and methodological advances to examine the effects of common genetic variants on serum DHEAS concentrations . By examining 14 , 846 Caucasian individuals , we show that eight common genetic variants are associated with serum DHEAS concentrations . Genes at or near these genetic variants include BCL2L11 , ARPC1A , ZKSCAN5 , TRIM4 , HHEX , CYP2C9 , BMF , and SULT2A1 . These genes have various associations with steroid hormone metabolism—co-morbidities of ageing including type 2 diabetes , lymphoma , actin filament assembly , drug and xenobiotic metabolism , and zinc finger proteins—suggesting a wider functional role for DHEAS than previously thought .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"medicine",
"public",
"health",
"and",
"epidemiology",
"geriatrics",
"epidemiology",
"endocrinology",
"genetics",
"biology",
"genomics",
"diabetes",
"and",
"endocrinology",
"genetics",
"and",
"genomics"
] |
2011
|
Eight Common Genetic Variants Associated with Serum DHEAS Levels Suggest a Key Role in Ageing Mechanisms
|
Decision making has been studied with a wide array of tasks . Here we examine the theoretical structure of bandit , information sampling and foraging tasks . These tasks move beyond tasks where the choice in the current trial does not affect future expected rewards . We have modeled these tasks using Markov decision processes ( MDPs ) . MDPs provide a general framework for modeling tasks in which decisions affect the information on which future choices will be made . Under the assumption that agents are maximizing expected rewards , MDPs provide normative solutions . We find that all three classes of tasks pose choices among actions which trade-off immediate and future expected rewards . The tasks drive these trade-offs in unique ways , however . For bandit and information sampling tasks , increasing uncertainty or the time horizon shifts value to actions that pay-off in the future . Correspondingly , decreasing uncertainty increases the relative value of actions that pay-off immediately . For foraging tasks the time-horizon plays the dominant role , as choices do not affect future uncertainty in these tasks .
Decision making has been studied with a wide array of tasks . Choices in many of these tasks either do not affect future choices or are modeled as if they do not affect future choices . For example , when asked to choose between gambles ( e . g . 50% chance of $20 or 100% chance of $11 ) , the choice in the current trial does not affect the gambles presented in the next trial , or the information on which one decides in the next trial . Correspondingly , even reinforcement learning tasks , where choices do affect the information that will be available for future choices , are often modeled using delta rule reinforcement learning ( DRRL ) or logistic regression , neither of which provides a normative description of the task . These modeling approaches assume that current choices should be driven entirely by past outcomes without considering how they will affect the future . Many interesting decision making problems , however , require consideration of how current choices will affect the future [1–7] . For example , there has been interest in the explore-exploit tradeoff [8–17] , information sampling [4 , 6] , and foraging [18 , 19] . Explore-exploit trade-offs exist in any real-world decision making context where one has to choose between continuing to exploit a known option , for example a familiar restaurant , vs . exploring an unknown or novel restaurant . Similarly , information sampling underlies many deliberative choice processes where one collects information before committing to a decision . For example , one might study product reviews or ratings before making a large purchase . These tasks require more sophisticated choice strategies because choices can be driven by future expected values . In other words , the best choice may not be the one that delivers the largest immediate reward . The best choice may lead to larger rewards in the future at the expense of smaller immediate rewards . Choices in these tasks can be modeled with markov decision processes ( MDPs ) . MDPs provide a general modeling framework , useful in tasks where the future depends upon what one chooses in the present . If one assumes that an agent is maximizing the expected ( discounted or undiscounted ) total reward , MDPs can be used to provide normative , or at least approximately normative , solutions to most current decision problems . While the choice behavior of subjects often deviates from normative behavior [4] , particularly in patient groups [5 , 20 , 21] , normative models are still important . Specifically , normative models identify the information on which decisions should be based , and the computations that must be carried out on that information . These two points can be conceptualized as the strategy optimal for the task . Further , normative models can be parameterized to fit the behavior of individual subjects [4 , 5 , 22] . This approach can provide insight into how subjects are deviating from the normative model and therefore it can suggest specific deficits or biases , as opposed to an overall change in task performance . Here , we used MDPs to model n-armed bandit , information sampling , and foraging tasks . Normative solutions to some of these tasks , to our knowledge , do not currently exist in the literature . There is , however , a long theoretical literature on binary bandit tasks [23 , 24] , and some foraging tasks have been modeled using the marginal value theorem [25] . For MDPs , the development of approximation techniques using basis functions has opened up the solution of a much larger class of problems than was tractable previously [26] . The normative solutions provide insight into the optimal strategies . We also used the models to examine several specific questions . For example , when is it useful to explore in a bandit task , and which features of the task can increase the value of exploring ? How can non-stationarity drive exploration ? Furthermore , once the tasks were mapped into the MDP framework we could examine their similarities and differences . This showed that decisions in all of these tasks pose a trade-off between immediate and future expected rewards . Further , we identified two factors that are important to this trade-off in these tasks . The first is uncertainty and the second is the time horizon . In bandit and information sampling tasks future expected values are relatively higher for options about which there is more uncertainty . When there is less uncertainty , action values are driven more by immediate expected reward . Further , uncertainty , and the value of exploring uncertain options is more valuable when the time horizon is longer . We also show that reward rate maximization in foraging tasks with an undiscounted , infinite time horizon is insensitive to travel delays to patches . In general with MDPs , infinite horizon undiscounted models are insensitive to finite delays to rewards .
When the environment is unknown , and model-free reinforcement learning ( RL ) is used to learn the environment [27] , exploration can be used to drive the RL algorithm to sample from the complete space of possible options . Here we deal with tasks where the environment is specified and MDPs ( or POMDPs ) can be used to calculate expected values for each state . Therefore heuristic exploration does not have to be used to make choices . Exploration , if it is defined as selecting options which have a smaller IEV but a larger action value , however , can still be optimal [27] . If an agent is maximizing total expected reward , an option with a smaller IEV can be selected if its FEV is relatively larger . Thus , immediate rewards can be foregone to obtain more total rewards over the relevant time horizon . We began by examining the explore-exploit trade-off in a stationary 2-armed bandit task , in which both bandits paid-off with the same fixed reward . The bandits varied , however , in the fraction of times they delivered a reward if chosen . In this case the explore-exploit trade-off affects the first few choices , before both targets have been sampled a few times . We modeled this as a finite state , finite horizon , undiscounted POMDP , where the information states were the number of times each bandit was chosen , Ci and the number of times each bandit was rewarded , Ri . This information state space is formed by the sufficient statistics for the two bandit processes . Transitions through the information state space occur after each choice and its associated outcome and they correspond to belief updates for the process . To examine the information state space for the bandit task more quantitatively , we can examine the distributions over expected future reward values generated in the task . Each of the bandit options is represented by a tree of possible outcomes ( Fig . 1A ) . Each node in the tree defines the information state ( i . e . Ri , Ci ) for that option . The information state can be used to estimate the underlying reward probability , q for each bandit option , where q is the hidden state of the system . As one of the options is sampled , the tree is traversed . With a binomial likelihood function and a beta ( α , β ) prior , the posterior over reward probability is given by pqri , ci∝pri , ciqpq p ( q|ri , ci ) ∝qri ( 1−q ) ci−riqα−1 ( 1−q ) β−1 p ( q|ri , ci ) ∝qri+α-11-qci-ri+β-1 The beta prior is the natural conjugate prior for the binomial likelihood function . Therefore , the prior can be interpreted as data . The posterior expected value is q|Ri , Ci = RiCi = α+riα+β+ci , where we have defined the actual choices and rewards as ri , ci , and the posterior choices and rewards as the data plus the prior Ri = ri + α , Ci = ci +α + β . If we start with a beta ( α = 1 , β = 1 ) prior we have posterior values of Ri = 1 , Ci = 2 for each bandit arm before any options have been sampled ( Fig . 1A ) . The possible posterior expected values are given by the nodes of the tree ( Fig . 1A ) . These nodes are also the immediate expected value for a choice , i . e . <rst , a> = RiCi , and these values also define the transition probabilities . Thus , if one is in state , Ri , Ci one transitions to Ri+1 , Ci+1 with probability pj = Ri+1 , Ci+1st = Ri , Ci , a = RiCi and one transitions to Ri , Ci+1 with probability pj = Ri , Ci+1st = Ri , Ci , a = 1-RiCi . This defines two of the terms on the r . h . s . of equation 2 ( ignoring the cost to sample ) . The other term on the r . h . s . of equation 2 is the utility of the next state , ut+1 . These utilities are recursively related to future utilities , ut+2 , etc . However , in the final trial , assuming a task where there are a finite number of trials and the number is known a-priori , there is no FEV because there will be no choices in the following trial . Therefore , utilities in the final trial , t = N , are given by the IEV , <rt ( st , a ) > . The IEV for each state that can exist in the final trial can be directly calculated from these information states . Once these are calculated , one can calculate the utilities for t-1 , and continue backwards until the utilities for the current trial can be calculated . This is the backwards induction algorithm ( [28]; see methods ) . When an option has not been sampled , any point in the tree can potentially be reached , although not under the optimal policy , and the distribution over reward probabilities is broad . This tree , therefore , represents the possible outcomes if one of the options is chosen repeatedly ( Fig . 1A ) . The state space for the task is , however , the product space over the nodes of two of these trees ( Fig . 1B ) , as it is constructed of all combinations of possible outcomes from each individual tree . When the FEV is calculated for one of the options , it is only calculated across the nodes in the full tree that are visited by the optimal policy . This is because the max operator in equation 1 is an expectation over the policy that optimizes choices in each state . Thus , when the FEV is calculated the expectation is taken over the portion of the product space ( Fig . 1B ) where the expected action value of an option is greater than the other option ( thick lines in Fig . 1B ) . The expectation is not computed over the dotted lines ( Fig . 1B ) because an optimal policy does not choose these actions . If we examine the distribution of reward probabilities over a representative finite horizon ( Fig . 1C ) we see that options which have been sampled less have higher expected values , when IEVs are each 0 . 5 . In this example it is less likely that one will encounter a reward probability ( q ) of 0 . 5 for options that have been sampled less , and more likely that one will encounter options that have a reward probability greater than 0 . 8 . This increased mass over higher reward probability nodes in the tree drives exploration in bandit tasks . As an example , we examined a scenario in which bandit option 1 was sampled 6 times , and rewarded three times ( Fig . 2A ) . ( Note that in this example the agent is not following the optimal policy . Rather we have defined choices and outcomes to illustrate action values under particular scenarios . ) The action value for option 1 exceeds the action value for option 2 during the first three trials while it is being rewarded . The FEV , however , of option 2 is larger than the FEV of option 1 , even in the first 3 trials , during which option 1 is being rewarded . After option 1 is not rewarded once , it becomes more valuable to sample option 2 ( i . e . Q ( s , 2 ) > Q ( s , 1 ) in trial 5 ) . After option 1 had been sampled 6 times and rewarded three times , its IEV is the same as option 2 , which had an expected value of 0 . 5 because of its prior . However , the action value ( IEV + FEV ) favors option 2 at this point ( i . e . trial 7 ) . If option 2 is then sampled 6 times and rewarded 3 times , the action values of the two options are again the same ( i . e . trial 13 ) . The exploration bonus ( here taken as the difference in FEV between the two options on trial 4 ) is also larger when the time horizon is longer ( Fig . 2B ) . This is because option 2 can be exploited for a longer time horizon if it is sampled and found to be better . When the first option chosen is rewarded , and it continues to be chosen and rewarded , the action value of the second option will not exceed the value of the first option ( Fig . 2C ) , given these finite time horizons . The exploration bonus is driven by three factors . Continuing on the example above , assume option 1 has been sampled and option 2 has not been sampled . First , there is uncertainty about option 2 ( i . e . the prior distribution over possible reward probabilities for unsampled options is broad , assuming a vague prior ) . Therefore , option 2 might be better than option 1 . If option 1 cannot be better than option 2 , because of the structure of prior knowledge , there is no exploration bonus . The second factor , as shown above ( Fig . 2B ) is the time horizon [17] . If the time horizon is too short one cannot obtain enough additional rewards when option 2 is found to be better than option 1 , to make up for the scenarios ( i . e . other episodes of the task ) when option 2 is found to be not as good as option 1 . This factor relies on the assumption that option 2 might be better than option 1 . Third , if option 2 is sampled and it is not as good as option 1 , then one can switch back to option 1 . On the other hand , if option 2 is better than option 1 , then one can stick with option 2 . This preference for the option which will be found to be better in the future , drives choices in the present via the max operator over action values in the utility equation ( equation 1 ) , which operates on the distribution of future outcomes via the embedded recursion . We next examined a novelty task [5 , 8 , 29] . This is a 3-armed bandit task similar in several ways to the 2-armed bandit task described above . The size of the reward is the same for each bandit option , but the probability of receiving a reward when each option is selected differs . In addition to this , however , choice options are replaced by novel choice options at stochastic intervals . Thus , after subjects accumulate experience with the current set of 3 bandit options for a period of time , one of the options is replaced by a novel option . These replacements are stochastic and not known in advance , but they are indicated to the subject . We modeled this task with an infinite horizon , finite state , discounted POMDP . Consistent with the 2-armed bandit , the information state is defined by Ri , Ci for each option . The full information state is now a product space across 3 trees ( Fig . 1A ) , so it is larger . To examine this task we considered a scenario similar to the one examined for the 2-armed stationary bandit . The action value of the chosen option ( option 1 ) increased while it was being rewarded in trials 1–3 ( Fig . 3A , for the choices and rewards see Fig . 3D; note that these actions are not chosen by the optimal policy . Rather they were chosen to illustrate the effect of experience with an option ) . The FEV also increased for all 3 options because of the overall increase in the expected reward in the environment ( Fig . 3C ) . However , similar to what was seen in the 2-armed bandit ( Fig . 2A ) , the FEV was larger for unexplored options ( Fig . 3B ) . Further , when option 1 was replaced , after each of the options had been chosen a few times , its FEV increased relative to the other two options ( Fig . 3B , trial 15 ) . Similarly , when option two was replaced on trial 20 , its FEV increased ( Fig . 3B ) . As with the 2-armed bandit , when the discount parameter was increased towards 1 ( Fig . 3E ) , the exploration bonus increased ( Fig . 3F ) . Thus , when a long time-horizon is available to exploit a novel option if it is found to be more valuable , the FEV for exploring that option increases . Every time a novel option is introduced , it is equivalent to resetting that option to the beta ( 1 , 1 ) prior , resetting it to the start of the tree ( Fig . 1A ) . Thus , uncertainty drives an exploration bonus as long as a sufficient time horizon is available to exploit the novel option if it turns out to be better than the alternative options available . Correspondingly , the substitution rate of novel options also affects the novelty bonus , by effectively limiting the time horizon ( Fig . 3F ) . If the substitution rate is high , one likely will have less time to exploit novel options that turn out to be good , before they are again replaced . To examine exploration in related bandit tasks , we used an infinite horizon , discounted , continuous state , POMDP to model a non-stationary two-armed bandit task [9] . The information state in this model is given by the mean and variance of the bandits , which are the sufficient statistics for the two processes . The bandits in this task returned continuous valued rewards ( e . g . 0–100 ) . The means of the returned values for each bandit were non-stationary in time , following independent , random walks that decayed to 50 . The actual reward earned on an individual trial was given by a sample from a Gaussian distribution with the current mean , and a standard deviation of 4 . The IEV is given by the estimated mean of each bandit . The utility depends on the estimated means of the two options ( Fig . 4A ) as well as the estimated variance of the options ( Fig . 4B ) . The effect of variance on utility also depends on the time-horizon ( Fig . 4B ) . The variance has a larger effect when the time horizon is longer . The effect of the variance of the utility can be understood in the framework developed above for the stationary bandit ( Fig . 1 ) . Specifically , when an option is not sampled its variance grows because of the nonstationarity of the underlying generative model , effectively driving it backwards in the tree ( Fig . 1A ) . On the other hand , when an option is sampled its variance decreases , effectively driving it forwards in the tree ( Fig . 1A ) . Thus , an option which has not been sampled for several trials becomes similar to a novel option , and it should be explored . We examined the choice sequence of the algorithm for some examples . If we consider an artificial case where the means are locked at 45 and 55 ( but the algorithm still assumes the means are non-stationary ) , and compare the sampling under two different discount parameters ( effective time horizons ) we see that the algorithm periodically samples the option with a smaller estimated mean , as its variance grows ( Fig . 4C ) . In addition , when the discount parameter is larger ( γ = 0 . 90 vs γ = 0 . 99 ) the algorithm samples more often , consistent with the larger difference in utility for a given standard deviation for larger discount parameters ( Fig . 4B ) . This can also be seen clearly in the action values ( Fig . 4E and F—Note that the algorithm stochastically sampled option 1 first in panel E and option 2 first in panel F , which gives rise to the initial downward vs . upward fluctuation ) . With the means fixed , the action values depend only on the variance of the two processes , if we ignore the decay of the process to 50 . When an option is sampled its variance decreases and its utility decreases , and when an option is not sampled its variance increases and its utility increases . The combination of these eventually drives the action value of the recently unsampled option to exceed the action value of the option currently being sampled ( Fig . 4E and F ) , and the option which has not been recently sampled is then sampled . This can be seen in example sequences drawn from the actual generative process as well ( Fig . 4G-H , the actual process values are identical for these two examples ) . In this case when the algorithm is modeled with a longer time horizon it samples more ( Fig . 4H ) . We next examined an information sampling task , often referred to as the beads or urn task [4 , 5 , 20 , 21 , 30] . In this task subjects are shown a sequence of beads drawn from one of two possible urns ( Fig . 5A ) . One of the urns has q orange beads and 1-q blue beads and the other has q blue beads and 1-q orange beads . After each bead is drawn subjects have three choices . They can either draw another bead from the urn , guess that beads are being drawn from the predominantly blue urn , or guess that beads are being drawn from the predominantly orange urn . Sampling another bead usually involves an explicit cost-to-sample . In other words , subjects are charged for collecting more information . In this task , the value of choosing an urn is given by the IEV , because no more samples are allowed after an urn is chosen so FEV is zero , whereas the value of sampling another bead is given by the FEV ( minus the cost-to-sample ) , because there is no reward if one does not try to infer the urn . Thus , this task explicitly sets up a trade-off between immediate and future expected rewards , and in this sense it is similar to the explore-exploit trade-off in bandit tasks . In most cases subjects are told that they can draw only up to a maximum number of beads and after the last bead is drawn they have to guess an urn . As such , the task can be modeled as a finite horizon , finite state , undiscounted , POMDP . The information state space is simpler than the state space in the bandit task , as it is given by a single tree ( Fig . 5B ) , where instead of rewards and no rewards , the state is given by the number of blue ( or orange ) beads that have been drawn , and the total number of beads drawn . These form the sufficient statistics for the process . As one draws beads , one works through the state space , similar to the situation with the bandit tasks . For example , the first 3 bead draws for the example sequence shown in Fig . 5C would go through the set of states shown ( Fig . 5B ) . Unlike the bandit task , this task was modeled with an uninformative prior on bead draws , because it is normally implemented by showing subjects one bead before asking them to decide [21] . The action values for guessing either of the urns or sampling again show that the value of guessing an urn increases as evidence for that urn increases ( i . e . more beads drawn of the corresponding color ) , and decreases as evidence for the urn decreases , in a cumulative fashion ( Fig . 5C-F ) . The value of sampling again is initially above the value of guessing an urn , but at some point it drops slightly below . Note that without a cost-to-sample ( C ( st , a ) = -0 . 005 in panels C-E and C ( st , a ) = -0 . 025 in panel F ) it is always best to sample all of the available beads . To examine the effect of the cost-to-sample , we calculated values for two costs , on identical sequences of bead draws ( Fig . 5E-F ) . When the cost was lower ( C ( st , a ) = -0 . 005; Fig . 5E ) , it was optimal to delay the decision until after the 11th bead was drawn , whereas when the cost was higher ( C ( st , a ) = -0 . 025; Fig . 5F ) it was optimal to decide after 2 beads . This task can be considered a pure exploration task: how long does one explore before committing to ( exploiting ) one of the choices ? This is similar to exploring a novel option for several trials , while always considering whether to switch back to the known option , or sticking with the novel option . As the certainty about which urn is being drawn from increases , picking an urn ( which will deliver an IEV ) , as opposed to drawing again ( which is valuable because of the FEV ) , becomes more valuable . The final tasks we considered were foraging tasks . Much like the tasks examined above , these tasks trade-off immediate and future expected values . Should one stay in the current patch whose resources are being depleted ( i . e . choose IEV ) or travel to a new patch ( i . e . choose FEV ) [19] ? Or , should one sample again ( i . e . choose FEV ) or commit to the current gamble on offer ( i . e . choose IEV ) [18] ? The state spaces for these tasks differ in a fundamental way from the state spaces in the bandit and information sampling tasks ( Fig . 6A and 7A ) . The state spaces for the foraging tasks are recursive . Stated another way , the state spaces for the foraging tasks do not represent learning or information accumulation . Learning or information accumulation are not recursive because you do not return to the same state ( technically , this is not completely accurate , as one can with some probability , return to a previous state in either the non-stationary bandit or the novelty bandit ) . Rather , in the foraging tasks the current state is provided to the animal and the animal does not have to estimate beliefs or distributions over states . Therefore these tasks are MDPs , as opposed to POMDPs where the state is hidden . In the foraging tasks one observes the state directly . In the patch leaving time task the subjects chose between staying in the current patch or traveling to a new patch [19] in each trial . The state relevant to choices is defined by the current amount of juice and the travel delay . If they stay in the current patch , they receive a ( slightly delayed ) reward , and the amount of reward that they will receive in the next trial if they again choose to stay in the current patch is decreased . If they choose to leave the current patch they have to wait for a known travel delay and they receive no immediate reward . The amount of reward that will be received in the new patch is reset to a fixed level and the travel time to the next new patch is sampled from the distribution of possible travel times ( Fig . 6A ) . The patch leaving time task was modeled as an infinite horizon , discounted MDP . The relevant state variables when a decision is made are given by the current travel delay and the current reward estimate ( Fig . 6A ) . From the model one can calculate the difference in action value for staying in the patch vs . leaving for another patch ( Fig . 6B and 6C ) . It can be seen that the longer the travel time , the longer one stays in the patch ( Fig . 6D ) , consistent with what was shown previously with heuristic models [19] . However , this effect only occurs for discount parameters less than 1 . The undiscounted model is insensitive to finite travel times ( Fig . 6E and 6F ) . This is because undiscounted infinite horizon MDPs are insensitive to finite time delays . Stated another way , if K is the mean first passage time to a state st = j and from state j one follows the optimal policy , then with an infinite horizon the value function can be written [26]: vNπs = limN = ∞1N∑t = 1K-1rst , asπ+limN = ∞1N∑t = KN-1rst , asπ From this it can be seen that actions taken prior to entering state j , at time K , do not matter . This is because the first sum is finite if the rewards are finite , and so it goes to zero in the limit . In the foraging task , if K is the time to get to the inter-trial interval ( ITI ) after choosing to travel , it doesn’t matter how long K is for finite K . The final task is a variant on standard foraging tasks . The state for this task is given by the current gamble pair on offer and the state space includes all the possible gamble pairs . In most foraging tasks a decreasing marginal reward in the current patch eventually drives the action value to leave the patch above the action value to stay in the patch , because leaving has a fixed expected value . This task , however , used a paradigm in which one samples , in each trial , two gambles from a set of six possible individual gambles ( Fig . 7A ) . The six individual gambles from which the pairs were drawn were shown for the current foraging bout and their reward values were known ( e . g . gamble 1 may have had a value of 12 points ) . In each round , a pair of gambles from the set of individual gambles was sampled ( 15 possible pairs assuming sampling without replacement from the 6 , and symmetry of gambles ) . For example , if the gambles for a given session were g1 … g6 , a subject might be shown in a single trial g3 and g5 . They then have to decide whether to engage with that offer pair , or sample again . If they sampled again , a new pair was drawn from the current set of six possible gambles ( perhaps g2 and g3 ) . Every time the subjects sampled again they also incurred a cost-to-sample . ( Note that a cost-to-sample is paid at the time of sampling , and it does not decrease the value of future gambles , in an MDP . ) If they decided to accept the offer , they moved to a decision stage . In this stage the probability that the reward associated with each gamble would be delivered was revealed , and this probability was randomly assigned to each gamble every time the decision stage was entered . The subjects had to choose one of the two gambles in the decision stage based on its magnitude and the associated probability . For example they might be choosing between p1g1 and p4g4 where pi is the probability that the subjects will receive reward gi if they choose that gamble in the decision stage . The agent then selects the gamble that has the highest expected value . The value of sampling again is given by the FEV . The FEV is not equal to the average values of the individual gambles . The FEV is the expected value of future draws ( see methods ) , plus the cost-to-sample . The time horizon is long , and many future samples could be drawn . However , the cost-to-sample decreases the value of future samples linearly with time , when viewed from the present decision . This can be compared to exponential discounting which exponentially decreases the value of future samples . With a sufficient time horizon the FEV is fixed . The task provided no explicit time horizon so we modeled it as a finite ( although long ) time horizon MDP . Therefore one simply samples until the IEV of the offered pair exceeds the ( constant ) FEV ( Fig . 7B ) . It is important to point out that sampling more in this foraging task , unlike the beads task , does not improve the IEV . In other words , the IEV does not necessarily increase with samples , although one can sample a pair with a better IEV . This is related to the state space of the problem . Additionally , without a cost-to-sample , the optimum strategy would be to sample until the pair with the highest value is drawn . The cost-to-sample creates a situation where choice of a gamble pair that is not the largest is optimal , because it may cost too much to obtain a better pair .
We began by examining the explore exploit trade-off in a two-armed bandit task , in which the reward amount for both options was the same , but they differed on the fraction of times that they were rewarded . Bandit tasks have been used to study learning in healthy and clinical populations [31 , 32] . In the first few trials there is value to sampling both options , and unsampled options have a larger FEV . This future expected value depends on three factors . First , the distribution over possible reward probabilities for the unsampled option is broad , given by the prior . Thus , the unsampled option may be more rewarding than the options which have been sampled . Second , if the unsampled option is sampled , and it is not as good as the other options , the subject can switch back to the other options . However , if the ( previously ) unsampled option is better than the other options , the subject can stick with it . Finally , the time horizon must be long enough to reap the rewards of investing samples in the novel option . Heuristically , one could consider the following approximate example . Assume that one has sampled one of two available options ( call it option 1 ) and found that it is being rewarded 70% of the time , and that one now has 100 more trials . One could then sample the alternative option ( option 2 ) 10 times . If option 2 is rewarded 80% of the time one could then stick with that option , gaining on average 80 rewards over the 100 trial horizon . If it is found that option 2 is only rewarded 20% of the time , then one could switch back to option 1 , gaining 0 . 2*10 + 0 . 7*90 = 65 rewards on average . If option 2’s ( i . e . all option 2’s that one encounters , in repeated plays of the task ) are either rewarded 80% of the time , or 20% of the time , the average reward with this simplified strategy will be 72 . 5 over the 100 trials , whereas it would only be 70 if one always stuck with option 1 . The 2 . 5 additional rewards on average is the exploration bonus . It depends on the possibility that the novel option is better than the current option , the fact that one will switch back to the alternative if it is better than the novel option , and having a sufficient time horizon . We also examined two other tasks which are extensions of the bandit task . Specifically , a non-stationary bandit [9] , and a novelty task [5 , 8 , 29 , 33] . In the non-stationary bandit the mean reward magnitudes of the two options follow independent random walks . When an option is sampled several times , a relatively accurate ( i . e . low variance ) estimate of its mean can be derived . However , when an option is not sampled , the distribution of its mean becomes broad . When a random walk is not observed for a period of time , the variance of its estimate grows linearly with time . One way to conceptualize this , relative to the stationary bandit , is to say that when an option has not been sampled for some time , it becomes like a new option , and there is value in exploring it . This is true of any tasks that have an underlying non-stationarity in the reward [34] . It is , however , variance in the estimate of the mean that drives the exploration bonus . When the variance gets large , the option might be better than the current options , and exploration is advantageous . Similarly in the novelty task , when a novel option is substituted for one of the options that has been sampled , the reward probability for the novel option is unknown , and therefore it is valuable to explore it . Next we examined the beads or urn task [4 , 5 , 20–22 , 30 , 35] . This is an information sampling task , similar in structure to other sampling tasks [36] . The POMDP model for this task only optimized choices in single trials with an explicit cost-to-sample . It did not optimize reward rates over multiple trials . Subjects are given the option to sample as much information as they would like , before guessing an urn . The choice to sample rests on the belief that the FEV of sampling is greater than the IEV of guessing an urn . In this respect , sampling is similar to exploring , as it is a choice in favor of the FEV , relative to the IEV . It differs from exploration , however , in that exploration in bandit tasks usually has some IEV . That is , choice of the unknown option in bandit tasks usually leads to some reward . This does not need to be true in general . In information sampling tasks , however , choosing to sample usually leads to zero IEV ( or a slightly negative IEV , given by the cost to sample ) . In this way , sampling is more similar to foraging . It is also worth pointing out that reaction time versions of perceptual inference tasks can be modeled within a framework that is equivalent to the approach used here to model information sampling [37 , 38] . Perceptual inference tasks , as well as many other choice tasks , are often modeled using a drift-diffusion framework , and it is assumed that when an evidence bearing particle crosses a threshold a decision is made . The “threshold” crossing is a choice to stop sampling . It is often inferred for drift diffusion models in perceptual inference tasks , on the basis of behavioral reaction times . But with an MDP the threshold can be calculated dynamically , on the basis of current levels of belief , costs-to-sample , and transition probabilities [38] . Thus , an optimal threshold can be inferred for any tractable task . The final tasks we considered were foraging tasks . These tasks also trade-off immediate vs . future expected values . The choice to forage leads to a zero IEV . The action value of choosing to forage is entirely an FEV . Foraging tasks differ from the tasks considered above , because their state spaces have a recursive structure and the state is observed , not inferred from information bearing observations . The tasks loop through their recursive state spaces over and over again . The choice is defined as a comparison between current and future stochastic offers . The current offer can be to stay in the patch and collect an approximately known , decreasing reward , or take the current pair of gambles that have been offered . The future stochastic offer can be explicitly calculated from the information given . It is either the value of a new patch , given the current travel time , or the expected value of the decision stage for the set of gambles that can be drawn from the current set . These average values are fixed with a sufficient time horizon . Therefore the strategy is to either stay in the current patch until the reward value drops below the value of leaving , or to sample gambles until the sampled gamble is worth more than the expected value of future samples . In foraging tasks there is generally no updating of distribution estimates , and therefore foraging differs fundamentally from exploration , in this respect . There are two important factors that drive choice preferences across these tasks . The first is uncertainty , and the second is the time horizon . Uncertainty affects these models in two ways . First , in bandit tasks when novel options are available , or equivalently when non-stationary options have not been explored for some time , the distribution of possible reward values is broad and uncertainty is high . Therefore , sampling the options a few times to learn about them is valuable , given a sufficient time horizon . The value of this uncertainty is driven by the future expected value . The sampling itself , however , decreases uncertainty about the options . When one learns that an option either is , or is not valuable , then one can act accordingly . Thus , increased uncertainty drives value through the FEV . Because we used models that maximize expected reward , uncertainty does not affect IEV . However , as uncertainty can lead to a larger FEV , decreasing uncertainty and therefore decreasing FEV increases the relative importance of IEV on the total action value . The same reasoning applies to the information sampling tasks . As long as uncertainty is high , the FEV is high . When uncertainty is decreased , however , the IEV of guessing an urn becomes larger . Interestingly , and in contrast to this , increasing uncertainty in temporal-discounting tasks actually decreases preference for delayed , larger rewards [5] . ( Temporal-discounting tasks are tasks in which subjects are offered a choice between an immediate smaller reward and a delayed larger reward . ) This is because of the state space of temporal-discounting tasks . One can model temporal discounting tasks using an MDP which , at each time step , includes the possibility of exiting the path to the reward and terminating in a state with no reward , with some probability . If this probability of terminating in a no reward state increases , it becomes less likely that one will get to the reward , for a fixed delay to the reward . Interestingly , this is thought to be a fundamental factor that drives crime [39] . Time horizon is also important . In infinite horizon problems the time horizon is controlled by the discount parameter . In bandit problems , the time horizon affects the relative value of exploration . In stationary problems the time-horizon affects the relative value of exploring novel or unknown options . Longer time-horizons , or discount parameters closer to 1 , increase the value of exploration . In non-stationary environments this relationship is more complex , as the non-stationarity limits the effective time-horizon of any policy . In foraging tasks , however , time horizon is also important . In undiscounted infinite horizon problems , travel times are irrelevant . If one has an infinite , undiscounted time horizon , then any finite travel time does not affect value . In the non-stationary bandit task , when the discount parameter approaches one , the algorithm samples options with lower means more often . As another example , consider the simplified MDP shown in Fig . 8 . The undiscounted , infinite horizon solution to this problem does not favor action 1 over action 2 [40] because the relative value of this initial transient reward will be zero in the infinite time limit . Methods such as sensitive discount optimality exist to deal with such situations , although these can only be applied to tractable state spaces [40] . However , a discounted MDP favors action 1 , in this case . This suggests that temporal discounting , in some form , is ubiquitous because it is always biologically ( or computationally ) relevant . Whether discounting is specifically exponential or hyperbolic , or takes on some other form is less the issue . More important is some sort of monotonic decrease in the value of future rewards with distance into the future . The explore-exploit trade-off is often modeled with heuristics . A strong criticism of heuristics is that they explain no more than they assume , and tell us no more than the data does [28] . Heuristics , however , can provide reasonable solutions to engineering problems , often provide insight into patterns in the data , and may better approximate behavior than normative models [15] . For example , recent work has explicitly examined the role of noisy vs . directed exploration , and found that human subjects use both directed and noisy exploration strategies [31] . In some cases , however , heuristics can be difficult to interpret . For example , the beta or inverse temperature parameter in delta-rule reinforcement learning ( DRRL ) is often thought to control the “explore-exploit” trade-off . This parameter can only control noise in choice processes , however , and standard implementations of DRRL do not turn this noise down as reward values are learned . Therefore , exploration cannot be differentiated from noise in the choice process using this parameter and poor learning looks like exploration . Several more sophisticated variants including Thompson sampling [41 , 42] and related algorithms [43] , however , decrease exploration with learning and can achieve minimal regret . In an MDP framework exploration need not be undirected or noisy . Exploration can be an intentional , directed , normative strategy if there is sufficient knowledge of the environment and the agent has sufficient computational resources . One does not necessarily explore so much as one learns or accumulates information ( bandit or information sampling tasks ) until the additional information indicates that an alternative choice is better . Every choice delivers some information , because one is always transitioning through states as choices deliver information in these tasks . Equivalently , leaving the current patch in a foraging task is an explicit calculation of the relative value of traveling to a new patch , the expected value of which is characterized by some probability distribution over patch values . It is possible that animals have relatively unsophisticated strategies for dealing with these issues . It seems likely , however , that they have developed at least a good approximation to the underlying normative utilities , at least in tasks that match the animal’s ecological nitch or on which the animals have extensive experience .
We modeled the tasks using markov decision processes with either observable ( MDP ) or partially observable ( POMDP ) states . Tasks were modeled as finite or infinite horizon , discrete time , and discounted ( i . e . with a discount parameter γ < 1 ) or undiscounted ( i . e . with a discount parameter γ = 1 ) as indicated in the manuscript . Some models also included a cost-to-sample . For discrete state models the utility , u , of a state , s , at time t is ut ( st ) = maxa∈Ast{ r ( st , a ) +C ( st , a ) +γ∑j∈Sp ( j|st , a ) ut+1 ( j ) } where Ast is the set of available actions in state s at time t , r ( st , α ) is the reward that will be obtained in state s at time t if action a is taken . The variable C ( st , α ) is a cost-to-sample , which may be zero . The summation on j is taken over the set of possible subsequent states , S at time t+1 . It is the expected future utility , taken across the transition probability distribution p ( j|st , α ) . The transition probability is the probability of transitioning into each state j from the current state , st if one takes action a . The γ term represents a discount factor . The terms inside the curly brackets are the action value , Q ( st , α ) = r ( st , α ) + C ( st , α ) + γΣj∈sp ( j|st , α ) ut+1 ( j ) , for each available action . For continuous state models the utility is utst = maxa∈Astrst , a+C ( st , a ) +γ∫Sp ( j|st , a ) ut+1 ( j ) dj All state integrals over continuous states were calculated with discrete approximations . Equations 1 and 2 assume a reward maximizing agent , through the max operator . For discrete state , finite horizon models with tractable state spaces , we used the backward induction algorithm to calculate utilities and action values [28] . This was done for the 2-armed stationary bandit , beads and sampling foraging tasks . With a finite horizon the final state delivers a reward , but no further actions are possible . Therefore , if we start by defining the utilities of the final states , we can work backwards and define the utilities of all previous states . Specifically , the algorithm proceeds as follows [40] , where N is the final state . 1 . Set t = N uN ( sN ) =r ( sN ) for all sN ϵ N . 2 . Substitute t-1 for t and compute utst = maxa∈Astrst , a+C ( st , a ) +γ∑j∈Sp ( j|st , a ) ut+1 ( j ) Set Ast , t* = argmaxa∈Astrst , a+C ( st , a ) +γ∑j∈Sp ( j|st , a ) ut+1 ( j ) 3 . If t = 1 stop , otherwise return to 2 . The non-stationary 2-armed bandit , novelty and patch-leaving foraging tasks were modeled as infinite horizon POMDPs or MDPs . The utilities were fit using the value iteration algorithm [40] . This algorithm proceeds as follows . First , the vector of utilities across states , v0 , was initialized to random values . We set the iteration index , n = 0 . Then computed: vn+1 = maxa∈Ast{ r ( s , a ) +γ∑j∈Sp ( j|st , a ) vn ( j ) } . ( 3 ) After each iteration we calculated the change in the value estimate , Δv = vn+1-vn , and examined either ||Δv||<∈ or span ( Δv ) <∈ . The span is defined as . span ( v ) = maxs∈S v ( s ) —mins∈S v ( s ) For infinite horizon undiscounted models the value continues to grow with iterations of equation 3 , but the spans converge [40] . This is because the final state values are the average costs per stage plus a differential . This only applied to the foraging patch leaving example with discount parameter equal to 1 . We only examined differential values , in that case , so the average cost per stage is subtracted out , because it is added to all states . We also used approximate methods for the non-stationary 2-armed bandit and the novelty task , as their state spaces were intractable over relevant time horizons . For these POMDPs we defined a basis , and then approximated the utility with v^ ( s ) =∑i=1Maiϕi ( s ) . ( 4 ) In all cases we used fixed basis functions so we could calculate the basis coefficients , ai using least squares techniques . We assembled a matrix Φi , j = ϕi ( sj ) , which contains the values of the basis functions for specific states , sj . We then calculated a projection matrix H = Φ ( Φ'Φ ) -1Φ' ( 5 ) And calculated the approximation v^ = Hv . ( 6 ) The bold indicates the vector over states , or the sampled states at which we computed the approximation . When using the approximation in the value iteration algorithm , we first compute the approximation , v^ . We then plug the approximation into the right hand side of equation 3 , vn+1 = maxa∈Astr ( s , a ) +γ∑j∈Sp ( j|st , a ) v^n ( j ) . We then calculate approximations to the new values v^n+1 = Hvn+1 . This is repeated until convergence . For basis functions we used piece-wise polynomials and/or b-splines [44] . For b-splines see [44] . For piecewise polynomials , the first basis functions are given by hi ( x ) = xi-1 . For an order K spline ( i . e . for cubic K = 3 ) , i goes from 1 to K+1 . In addition to these global polynomials , we also add hj ( x ) = ( x-tj ) K for the J knots , tj . Because all of the state spaces were multidimensional and the piece-wise polynomial basis varied between knots , we also had to compute products of the basis functions across dimensions . Computing the full tensor product basis space was usually intractable . It created a projection matrix that either could not be stored in memory or iteration over the very large projection matrix was so slow that the algorithm would not converge in a reasonable amount of time . Therefore we started with linear terms and added interaction terms of increasing order ( i . e . second order , third order , … ) until the approximation stopped improving . We did not find an improvement by going beyond the quadratic terms . Knot locations were explored systematically to find locations that led to good approximations . Approximations were checked in several ways . First , we plotted vn+1 vs . v^n to see that they were consistent after convergence , as well as checking the variance of the residual . Second , we added knots to see if the fit was improved . Third , we increased the order of the polynomial to see if the fit was improved . Cubic polynomials ( i . e . K = 3 ) were used in all cases . When the order was increased beyond cubic the value iteration often diverged . Finally , performance of the approximate inference MDP for the novelty task could be compared to a corresponding finite horizon model , at least for short time horizons to see if they made consistent predictions . For the novelty task , the numerics were easier to implement if we approximated the number of samples for each option ( N ) and the probability that it was rewarded ( p ) . We used a 3rd order B-spline basis . Knot locations for N were 0 and 150 , and the algorithm was optimized at ( using Matlab colon operator notation ) N = e0:54:5 and p = 0: 0 . 25: 1 . The N values were not integers , but this does not affect evaluation of the value function . Interactions up to second order were included . For the non-stationary two-armed bandit , the means were fit with a 3rd order B-spline , and the standard deviations were fit with a 2nd order piece-wise polynomial . This approach gave well-behaved value functions . The node locations for the means were given by -30 50 and 130 . The means were evaluated at 0 , 12 . 5 , 25 , 37 . 5 , 50 , 62 . 5 , 75 , 87 . 5 , 100 . The node locations for the standard deviations were given by 0 . 25 , 1 , 3 , 5 , and 15 . The standard deviations were evaluated at 0 . 5 , 1 , 2 , 3 , 4 , 5 , 7 and 14 . Interactions between all basis functions up to second order were included in the model .
|
Numerous choice tasks have been used to study decision processes . Some of these choice tasks , specifically n-armed bandit , information sampling and foraging tasks , pose choices that trade-off immediate and future reward . Specifically , the best choice may not be the choice that pays off the highest reward immediately , and exploration of unknown options vs . exploiting known options can be a normatively useful strategy . We characterized the optimal choice strategies across these tasks using Markov Decision Processes ( MDPs ) . The MDP framework can characterize optimal choice strategies when choices are affected by the value of future rewards . We found that uncertainty and time horizon have important effects on the choice strategies in these tasks . Specifically , in bandit and information sampling tasks , increasing uncertainty increases the value of exploring choice options that tend to pay off in the future , while decreasing uncertainty increases the value of choice options that pay off immediately . These effects are increased when time horizons are longer . Foraging tasks differ in that uncertainty plays a minimal role . However , time horizon is still important in foraging . Specifically , for long time horizons , travel delays to rewards become less relevant .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[] |
2015
|
Theory of Choice in Bandit, Information Sampling and Foraging Tasks
|
Migratory waterfowl of the world are the natural reservoirs of influenza viruses of all known subtypes . However , it is unknown whether these waterfowl perpetuate highly pathogenic ( HP ) H5 and H7 avian influenza viruses . Here we report influenza virus surveillance from 2001 to 2006 in wild ducks in Alberta , Canada , and in shorebirds and gulls at Delaware Bay ( New Jersey ) , United States , and examine the frequency of exchange of influenza viruses between the Eurasian and American virus clades , or superfamilies . Influenza viruses belonging to each of the subtypes H1 through H13 and N1 through N9 were detected in these waterfowl , but H14 and H15 were not found . Viruses of the HP Asian H5N1 subtypes were not detected , and serologic studies in adult mallard ducks provided no evidence of their circulation . The recently described H16 subtype of influenza viruses was detected in American shorebirds and gulls but not in ducks . We also found an unusual cluster of H7N3 influenza viruses in shorebirds and gulls that was able to replicate well in chickens and kill chicken embryos . Genetic analysis of 6 , 767 avian influenza gene segments and 248 complete avian influenza viruses supported the notion that the exchange of entire influenza viruses between the Eurasian and American clades does not occur frequently . Overall , the available evidence does not support the perpetuation of HP H5N1 influenza in migratory birds and suggests that the introduction of HP Asian H5N1 to the Americas by migratory birds is likely to be a rare event .
Long-term surveillance of influenza in migratory waterfowl in North America [1–4] from 1976 to the present and more intensive surveillance in Europe from 1998 to the present [5 , 6] have established the importance of Anseriformes ( waterfowl ) and Charadriiformes ( gull and shorebird ) in the perpetuation of all known subtypes of influenza A viruses . The available evidence suggests that each of the 16 hemagglutinin ( HA ) and nine neuraminidase ( NA ) subtype combinations exist in harmony with their natural hosts , cause no overt disease , and are shed predominantly in the feces . Although highly pathogenic ( HP ) influenza viruses have occasionally been isolated from wild migratory birds , including H5N3 A/Tern/South Africa/61 [7] and H7N1 from an outbreak of HP avian influenza in Italy in 1999–2000 [8] , the usual finding is that each of the HP H5 and H7 lineages that emerge in gallinaceous poultry originate from different nonpathogenic precursors—e . g . , A/Chicken/Pennsylvania/83 ( H5N2 ) [9] , A/Chicken/Mexico/94 ( H5N2 ) [10] , A/Chicken/Netherlands/2003 ( H7N7 ) [11] , and A/Chicken/Canada/2004 ( H7N3 ) [12] . There has been overall agreement between the findings of various influenza surveillance studies in migratory birds in regard to the role the birds play in the emergence of pandemics in humans , lower animals , and domestic poultry . The only significant difference between the findings for influenza surveillance in aquatic birds in the Americas and in Europe is the role of shorebirds; in Europe , influenza viruses to date have rarely been isolated from shorebirds , while in the Americas , the available evidence supports the notion that shorebirds carry influenza viruses during their migration from South America to their North American breeding grounds each May . Our earlier studies also showed that shorebirds and gulls in the Americas are more frequently the source of the potential precursors to HP H5 and H7 avian influenza viruses [3] , while in Eurasia , the precursors of HP influenza viruses are usually from duck species [13 , 14] . An important unanswered question is how often the influenza viruses in wild migratory birds in Eurasia spread to the Americas and establish lineages on that continent , and vice versa . This is currently a question of great concern to both veterinary and public health officials in the Americas . The continuing circulation of the Asian HP H5N1 in several countries in Eurasia and the reemergence in the winter of 2006–2007 of HP H5N1 in South Korea , Japan , Thailand , Vietnam , and China supports the contention that this H5N1 virus is being perpetuated in this region . One of the many unanswered questions is whether the HP avian H5N1 virus is being perpetuated in domestic poultry or in wild bird species . Examination of the global migratory pathways of migratory waterfowl shows overlaps between eastern Eurasia and Alaska and between Europe and eastern North America [15] . This leads to the question of why the HP H5N1 viruses have not arrived in the Americas . Phylogenetic analysis of influenza viruses from migrating birds generally divides them into two large polyphyletic clades or superfamilies: one in Eurasia , the other in the Americas [6 , 16–18] . Whereas most of the known subtypes of influenza A viruses have been detected in each hemisphere , the rarely isolated H14 and H15 subtypes have , to date , been detected only in Eurasia . While influenza viruses or gene segments have been shown to exchange between parts of Eurasia and the Americas [19–24] , the frequency of exchange of all eight gene segments of influenza viruses is unknown . However , based on the clear phylogenetic separation into two superfamilies [25] , it might be predicted that the exchange of whole influenza viruses would be infrequent . Here , we utilize the genomic information from 248 complete influenza virus sequences and 6 , 767 gene segments [26] to estimate how often this occurs . In our studies of influenza surveillance in wild ducks in Alberta , Canada , and in shorebirds and gulls at Delaware Bay ( New Jersey ) , United States , between 2001 and 2006 , we confirmed the presence of the H16 subtype in the Americas and the presence of a high frequency of infection of H7N3 viruses from shorebirds and gulls . We continued to detect H1 through H13 HA subtypes and N1 through N9 subtypes of NA but not H14 or H15 in ducks , shorebirds , or gulls . We found no virologic evidence of the HP H5N1 virus in ducks , shorebirds , or gulls or serologic evidence in ducks sampled in Alberta . Our genomic analysis of the full sequence of 248 influenza genomes showed no viruses whose entire genome transferred between the two hemispheres . The size of our complete-genome virus data set , however , may be too small to detect a transferred virus before it had opportunities to reassort .
The rapid spread of HP H5N1 influenza from Qinghai Lake , China , to Europe and Africa raised the possibility that migratory birds were involved [15] . The reemergence of HP H5N1 influenza virus in domestic poultry in Japan , South Korea , and Thailand in January 2007 after successful eradication of these viruses in 2003 and 2004 raised the question of the mode of spread and raised the level of concern for further spread of these viruses . To determine the prevalence of different subtypes of influenza A viruses at two sites in North America , virologic surveillance was done in wild ducks in Alberta and in shorebirds and gulls at Delaware Bay from 2001 through 2006 . Prospective surveillance has been done for the past 30 years in wild ducks in Alberta and for 21 years in shorebirds and gulls at Delaware Bay . During the five years of our study , 590 cloacal samples from wild ducks yielded 98 influenza A viruses ( isolation rate , 16 . 6% ) , while 1970 fecal samples from shorebirds and gulls yielded 114 influenza A isolates ( isolation rate , 5 . 8% ) . Antigenic analysis of the influenza virus isolates from ducks , shorebirds , and gulls established the continuing circulation of HA subtypes H1 through H13 and NA subtypes N1 through N9 ( Figure 1 ) . We also report the isolation of the newly characterized H16 subtype in shorebirds . Subtypes H2 , H4 , and H8 were identified only in ducks , while subtypes H5 , H13 , H16 , and N2 were recovered only from shorebirds and gulls . Subtypes that occurred in both ducks and shorebirds , but with much higher prevalence in shorebirds , were H9 , H11 , H12 , and N9 . In contrast , although they were recovered from both ducks and shorebirds , N1 , N4 , and N6 were more prevalent in ducks . Five HA subtypes accounted for 75% of the isolates found in shorebirds: H1 ( 19 . 5% ) , H10 ( 17 . 5% ) , H9 ( 13 . 5% ) , H3 ( 12 . 4% ) , and H7 ( 11 . 6% ) , whereas four HA subtypes represented 82% of the viruses identified in ducks: H1 ( 24 . 0% ) , H4 ( 24 . 0% ) , H10 ( 17 . 7% ) , and H3 ( 16 . 7% ) . For the NA subtypes in shorebirds , the N7 ( 24 . 7% ) and N9 ( 24 . 3% ) isolates were dominant , and for ducks , N4 ( 25 . 0% ) and N6 ( 25 . 0% ) were the most prevalent subtypes . We isolated only three H5 viruses , all of which came from shorebirds in 2004 and were paired with either N7 or N8 . We identified a total of 49 HA and NA subtypes ( out of a possible 144 ) , with 25 subtypes being found in ducks and 31 subtypes being found in shorebirds . There were 17 subtypes that were isolated only from ducks and 24 subtypes that were recovered only from shorebirds . The most highly represented subtypes found in ducks were H4N6 ( 21 . 9% ) , H1N4 ( 18 . 8% ) , and H10N7 ( 16 . 7% ) , and the most numerous subtypes identified in shorebirds were H10N7 ( 17 . 5% ) , H1N9 ( 15 . 1% ) , and H7N3 ( 19 . 8% ) . No H5N1 was detected during the study period . The isolation of high numbers of H7N3 influenza viruses from gulls and shorebirds in May 2006 was different from the results of previous years . These H7N3 viruses are examined in more detail below . To determine the prevalence of infection of wild ducks with the influenza viruses that are considered more likely to cause a pandemic in domestic fowl or in humans , hemagglutination inhibition ( HI ) studies were done with H5 , H6 , H7 , and H9 influenza viruses ( Table 1 ) . The influenza viruses chosen for use were those that were of current relevance in Asia and the Americas . Low levels of HI antibodies to each of the viruses tested were detected each year with an overall higher rate of detection of H9 followed by H5 , H6 , and H7 antibodies . Most positive sera reacted with HI titers of 1/10 to 1/20 , indicating that the birds were probably infected with a virus of the homologous subtype but probably not with viruses identical to those used in the test . Studies done yearly from 2004 through 2006 on adult ducks showed no trend toward higher numbers of birds infected with H5 , H6 , H7 , or H9 . Thus , serologic studies confirmed the continued circulation of each of the influenza subtypes . A new subtype of influenza A viruses from black-headed gulls in Sweden was recently characterized and designated A/Black-headed gull/Sweden/2/99 ( H16N3 ) [27] . Large-scale sequence analysis of avian influenza viruses [26] established that viruses of the H16 subtype are present in the American superfamily . The first of these viruses , A/Black-legged Kittiwake/Alaska/295/75 ( H16N3 ) , was isolated in 1975 , and later isolates were detected in shorebirds in 1986 and in herring gulls in 1988 . In our study , three H16N3 influenza viruses ( A/Shorebird/DE/168/06 , A/Shorebird/DE/172/06 , and A/Shorebird/DE/195/06 ) were isolated from shorebirds . No H16 viruses were detected in samples from duck species . Phylogenetic analysis showed that the HA of these viruses could be divided into two subgroups , one from Europe and the other from North America ( Figure 2 ) . It is noteworthy that A/Black-headed gull/Sweden/5/99 ( H16N3 ) grouped with the North American H16 isolates , suggesting a possible transfer between North America and Europe . However , it is not possible to determine if the HA gene originated in North America or Europe . Antigenic analysis of the H16 influenza virus isolates from North America with postinfection ferret sera ( Table 2 ) showed that the viruses were antigenically distinguishable; two of the current H16 isolates ( SB/DE/172/06 and SB/DE/195/06 ) were indistinguishable from each other , but the SB/DE/168/06 and BHG/Sweden/5/99 reacted to 4-fold lower HI titers ( 1/80 versus 1/320 ) . The H16N3 isolate from 1975 also reacted to a titer of 1/80 , while the 1986 and 1988 viruses reacted to a titer of 1/40 . Thus , the H16 viruses showed antigenic and genetic diversity in their natural hosts . Two H7N3 influenza viruses of the American superfamily of influenza viruses have in the recent past evolved into HP viruses . One of these HP H7N3 viruses occurred in Chile , South America , A/Chicken/Chile/4322/02 ( H7N3 ) [28]; the other was in British Columbia , Canada , A/Chicken/British Columbia/NS-2035–12/04 ( H7N3 ) [12] . Each of these viruses evolved from nonpathogenic H7N3 precursor viruses , and the available evidence indicates they acquired the HP characteristic by recombination; they inserted additional amino acids at the cleavage site of the HA [29 , 30] from one of their other gene segments . In 2006 , we isolated 24 H7N3 , one H7N4 , and one H7N5 influenza viruses from shorebirds and gulls at Delaware Bay . Each of these viruses was passaged once in chicken eggs after the initial egg isolation , produced high HA titers ( 1:640–1:4 , 096 ) , and 20 of the viruses killed the chicken embryos by 48 h after injection . To determine whether these viruses could also cause death of chickens , the A/Laughing gull/DE/42/06 ( H7N3 ) isolate was inoculated into four young adult white leghorn chickens intranasally , intratracheally , and intraocularly with 108 . 75 egg infectious doses50 ( EID50 ) . The birds showed no loss of appetite or disease signs; virus was reisolated from tracheal or cloacal samples of three of the four chickens ( results not shown ) . The H7N3 isolate induced high levels of HI antibodies in the inoculated chickens . In HI tests using postinfection antiserum , the 26 H7 viruses were antigenically homogeneous with HI titers mostly within 2-fold of each other ( 1:160 , 1:320 , or 1:640 ) . To further evaluate the pathogenic potential of the viruses , two representative H7N3 isolates ( A/Shorebird/DE/22/06 and A/Laughing gull/DE/42/06 ) were inoculated intravenously with 108 . 5 and 108 . 75 EID50 , respectively , into ten chickens each to determine their intravenous pathogenicity index ( IVPI ) . Virus was detected from the cloaca of all the birds on day 3 postinfection . Although neither of the H7N3 viruses tested were HP , one bird did die 7 d after inoculation with A/Laughing gull/DE/42/06 , and four of the birds were reported as lethargic with an IVPI of 0 . 28 . Tissue samples from the dead bird had modest levels of virus in the pancreas ( 102 . 5 EID50/ml ) and the highest levels in the kidney ( 105 EID50/ml ) . No virus was detected in samples from the brain , lung , heart , liver , or spleen of the dead bird . Repassaging of the virus from the kidney into ten chickens intravenously did not kill any of the birds . To determine whether the virus isolated from the dead chicken was capable of replicating in the absence of exogenous trypsin , a characteristic of HP viruses , we performed a plaque assay in MDCK cells with and without trypsin . Plaques were produced only in the presence of trypsin , indicating that the virus reisolated from the chicken could not be characterized as possessing this trait of HP influenza viruses . Phylogenetic analysis of the HA of the H7N3 viruses from gulls and shorebirds at Delaware Bay in 2006 shows they reside on a neighboring branch to that of the HP H7N3 viruses from chickens in British Columbia in 2004 ( blue text , Figure 3 ) . Although the viruses were related to those found in British Columbia , the connecting peptide of the HA of the shorebird viruses had the typical sequence of a nonpathogenic H7 virus . Phylogenetic analysis of the internal genes and the NA gene showed that all genes reside in the American clade , and the PB2 , PB1 , NP , NA , M , and NS genes are located on a branch adjacent to the HP H7N3 viruses from British Columbia ( Figure S1 ) . A BLAST search was performed to determine the similarity of the shorebird and gull H7N3 viruses with the HP strains isolated in British Columbia . Nucleotide identity was high—93 . 8% to 98 . 9%—for all genes except PA , which had an identity of 87 . 8% . Even though these H7N3 viruses are classified as nonpathogenic , they clearly have some potential for replication in domestic chickens and have the unusual characteristic of killing chicken embryos . The detection of the HA of the H16 Swedish black-headed gull influenza virus among the rarely identified H16 viruses from North America ( Figure 2 ) raised the question of the frequency of influenza virus transfer between continents . To investigate this , we examined the 6 , 767 genome segments of influenza viruses from wild birds , described previously by Obenauer et al . [26] ( Table 3; Figures S2–S9 ) . These viruses were generally separable into two large polyphyletic clades , one made up of viruses from the Americas , the other composed of viruses from Eurasia ( and including Australia and Africa ) . In addition , we found a small number of viral segments isolated in either the Americas or Eurasia that clustered with strains from the opposing hemisphere ( Table S1 ) . Any remote segment or paraphyletic clade of segments , which clustered in the above manner , was termed as an outsider event . When the 6 , 767 influenza A gene segments were examined , 3 , 040 were found have been isolated in the Americas , while 3 , 727 were isolated in Eurasia ( including Australia and Africa ) ( Table 3 ) . Among the 3 , 040 American segments , there were 32 total outsider events , encompassing 101 total segments ( Table S1 ) , of which 30 were from Anseriformes ( ducks ) and 71 were from Ciconiiformes ( which for our phylogenetic analyses included Charadriiforms [shorebirds and gulls] ) isolated in Eurasia . Among the 3 , 727 Eurasian clade segments , there were 24 outsider events , encompassing 35 total segments , of which 20 were from Anseriformes and 15 from Ciconiiformes isolated in the Americas . Among the internal segments ( PB1 , PB2 , PA , NP , M , and NS ) no outsiders were present in all six segments ( Table S1 ) , indicating that , among the viruses examined , no entire viral genome from one clade was detected in the other clade . However , the 24 outsider events among the internal gene segments were detected in a non-linked fashion . Outsider events involving American isolated strains occurred most often with HA and PB1 genes ( eight and seven events , respectively ) . Fewer events involving American strains were found with the remaining segments , ranging from five events ( PB2 ) to zero events ( NP ) . In contrast to the American outsider events found in Eurasian clades , the detection of outsider Eurasian events in American clades was lower . Eurasian virus outsider genes were only detected for HA ( ten events ) , NA ( eight events ) , PA ( three events ) and NP ( three events ) . The overall more frequent detection of the HA and NA outsider events probably reflects the ability of these genes to exchange by reassortment and to be maintained in the population , presumably after the rare occurrence of an exchange of entire viruses between the continents . Thus , there does not appear to be a restriction on compatibility of the HA and NA genes between the two superfamilies . When the internal genes were compared , PB2 , PB1 , M , and NS showed marked differences in the number of outsider segments detected in American clades compared to the number of outsider segments found in Eurasian clades . None of these segments were identified as outsiders in the American clades; however , the Eurasian clades had seven PB1 American wild bird outsiders while PB2 had five , M had three , and NS had four . When the two types of aquatic birds were compared , the Ciconiiformes ( shorebirds and gulls ) more frequently contained outsider gene segments than the Anseriformes ( ducks ) . Among all the Ciconiiformes ( n = 728 ) , 9 . 8% of American virus gene segments were detected in Eurasian clades , and 2 . 1% of Eurasian virus gene segments resided in American clades . The frequency of outsider segments among all Anseriformes ( n = 3 , 436 ) , however , was similar across both hemispheres: 0 . 87% of the outsider segments were isolated in the Americas , whereas 0 . 58% were isolated in Eurasia . These figures represent the relative frequency of detection of outsider gene segments in either the Eurasian or American clades , but they do not give a measure of the frequency of exchange of the entire virus ( all eight segments ) between continents . In our studies , we failed to detect any complete genome exchanges among the 248 viruses examined ( Table S2 ) . Thus , the frequency of exchange was less than 0 . 6% .
Virologic surveillance of apparently healthy birds has established that the waterfowl of the world are the natural reservoirs of all known influenza A viruses . The Anseriformes ( waterfowl ) and Charadriiformes ( shorebirds and gulls ) are the major reservoirs in which the 16 HA and nine NA subtypes are perpetuated . However , a wide range of birds can support limited replication but do not perpetuate influenza A viruses . In this study , we characterized the H16 influenza viruses from the Americas that were first described in Europe in 2004 [27] and found an unusually high frequency of H7N3 influenza viruses from shorebirds and gulls at Delaware Bay in 2006 . Despite virologic and serologic surveillance in migratory waterfowl , no evidence of the Asian HP H5N1 influenza virus was found . Continued prospective surveillance at two sites—one in Alberta for ducks and another at Delaware Bay for shorebirds and gulls—established the continued circulation of HA subtypes H1 through H13 and H16 and all nine NA subtypes . However , the rare H14 and H15 subtypes were not detected . One frequently mentioned possibility is that wild migratory birds from Eurasia will carry the Asian HP H5N1 to the Americas . There are several articles describing the detection of influenza viruses belonging to Eurasian phylogenetic clades in the Americas and vice versa [19–24] . Each of these studies has been based on analysis of single genes ( e . g . , the HA , M , etc . ) . To date , we know of no study that has investigated the spread between the continents of entire influenza genomes containing all eight segments . In this study , we examined 6 , 767 individual gene segments and 248 entire influenza genomes . Phylogenetic analysis supported the contention that the influenza viruses are clearly separable into distinct Eurasian and American superfamilies [6 , 16–18 , 20] . We measured the frequencies of gene segments belonging to one superfamily occurring among influenza virus isolates in the other superfamily , defined here as outsider events , and found low frequencies of inter-hemispheric gene transfer . The rates varied according to gene segment , ranging from 0 . 25% for M ( three outsider events seen among 1 , 187 sequences ) to 1 . 77% for HA ( 18 events among 1 , 014 sequences ) . Our analysis of entire influenza genomes containing all eight gene segments revealed no detectable whole-genome transfers between superfamilies among the 248 viruses examined . After such an event occurs , there are relatively frequent reassortment events involving the surface glycoprotein genes ( HA and NA ) , implying that there is no restriction in compatibility between the internal genes of one superfamily and the surface glycoproteins of the other superfamily . The internal genes are exchanged at a lower frequency and maintained , but there are no obvious patterns implying linkage of genes . Based on these studies , it is more likely that the Asian H5N1 viruses will be imported into the Americas with birds moved legally or illegally by humans . The available evidence is that at least one of the introductions of Asian HP H5N1 into Nigeria was with poultry that was imported either alive or frozen [31] . The recent introduction of HP H5N2 of American lineage into Japan indicates another mechanism for moving influenza virus between continents . The close genetic similarity between the HP H5N2 and the vaccine strain of H5N2 used in poultry in Central America suggest that this virus was transferred with smuggled vaccine [32] . Japan has a strict ban on the use of agricultural vaccines , yet the HP H5N2 of American lineage was introduced into Japan and caused lethal disease in poultry . Whether this originated from improperly inactivated vaccine or from contaminated vaccine is unknown . The earliest H16 in the repository at St . Jude Children's Research Hospital is A/Kittiwake/Alaska/295/75 ( H16N3 ) , indicating that this virus has been in the Americas for a considerable period . It is notable that an H16 virus from a black-headed gull in Sweden is included in the American phylogenetic lineage , indicating transfer of the HA between continents . To date , the H16 viruses have been isolated only from gulls and shorebirds but not from ducks . The H16 viruses have shown limited antigenic drift from 1975 to 2006 and , to date , have not been associated with disease in any species . The detection of a cluster of H7N3 influenza viruses in shorebirds in May 2006 was not found in previous years and raises questions about the propensity of these H7N3 viruses to replicate well in gallinaceous poultry and kill chicken embryos . In experiments using four H7N3 influenza viruses isolated from shorebirds and wild ducks between 1977 and 2000 , little or no replication of virus was observed in experimentally infected chickens . In contrast , most of the H7N3 viruses used in our study replicated to high titers after oral inoculation in chickens , and 20 of the 24 viruses killed chicken embryos . Phylogenetic analyses and sequence homologies of the genome of this cluster of H7N3 isolates showed that these viruses are closely related to the H7N3 influenza strain isolated from poultry in British Columbia in 2004 . Taken together , the data indicate the potential of these H7N3 viruses to replicate in domestic chickens and underline the importance of biosecurity in commercial poultry raising and the need to keep such influenza viruses out of commercial poultry so they do not have the opportunity to evolve into HP strains . The continued reemergence of HP H5N1 influenza viruses from the hypothetical epicenter in Guangdong Province , China [33 , 34] , over the past ten years raises the question of whether the HP Asian H5N1 viruses are being perpetuated in wild birds . Surveillance in migratory waterfowl does show that many species will support the replication of the HP H5N1 viruses [35]; however , to date there is no evidence that the Asian HP H5N1 viruses are being perpetuated in migratory waterfowl . Although the spread of HP H5N1 viruses from Qinghai Lake , China , to central Asia and Europe has been attributed to migratory birds [15] , there is no evidence that the HP H5N1 viruses are being perpetuated in migratory waterfowl . Although the number of samples collected per year from migratory ducks in Alberta and shorebirds and gulls at Delaware Bay are rather modest , they do support the findings of the recent extensive surveillance done by the Canadian Cooperative Wildlife Health Center [36] and by the United States Geological Survey [37] . Surveillance studies in more than 100 , 000 wild birds have provided support for the contentions that ( 1 ) waterfowl are a major reservoir of influenza viruses; ( 2 ) low-pathogenic H5 influenza viruses are present in these wild birds; and ( 3 ) the HP H7 viruses previously found in Canada were not detected . Similarly , studies done over the 30 years of influenza virus surveillance in wild birds [3] in the Americas found no evidence for the perpetuation of HP H5 or H7 influenza viruses in migratory birds . The available evidence suggests that the perpetuation of HP H5N1 Asian influenza viruses occurs through domestic waterfowl [33 , 38] . In the cooler months of the year , the virus load in the infected but apparently healthy domestic waterfowl spills over into both migratory waterfowl and gallinaceous poultry flocks and reignites the spread of HP H5N1 viruses . There is still a paucity of surveillance data from migratory birds in Asia , but the extensive surveillance of migratory waterfowl in Europe [6] does not support the hypothesis that wild migratory birds are perpetuating the Asian HP H5N1 virus . Thus , the above studies on the low frequency of exchange of entire influenza virus genomes of nonpathogenic influenza viruses between Eurasia and the Americas , plus the absence of evidence of HP H5N1 virus perpetuation in the influenza viruses found in migratory waterfowl , make the probability of introduction of HP H5N1 into the Americas by migratory birds an unlikely event .
During 2001–2006 , influenza surveillance in wild ducks that began in Alberta in 1976 was continued . During the same years , surveillance studies that were initiated in 1985 in shorebirds migrating through Delaware Bay were continued annually . Collection sites , collection of specimens , virus isolation , and characterization of the isolates have been described previously [3 , 4] . The A/Black-headed gull/Sweden/5/99 H16N3 virus used in this study was kindly provided by A . D . M . E . Osterhaus , Erasmus Medical Center , Rotterdam , The Netherlands . Healthy adult wild ducks , captured in Alberta by the Canadian Wildlife Service for banding and tracking migration , were bled in the field to obtain 3–4 ml of blood from each duck and were released . Blood was clotted in serologic tubes overnight at room temperature to separate the serum , and the serum was transferred by pipette to cryovials and stored in liquid nitrogen for shipment . In 2004 , 326 serum samples were collected; in 2005 , 149 serum samples were collected; and in 2006 , 226 serum samples were collected . All sera were treated with receptor destroying enzyme overnight at 37 °C , inactivated at 56 °C for 30 min , and diluted 1:10 before antigenic testing . Sera were screened by HI test as previously described by Palmer [39] for antibodies to four subtypes of influenza antigen: H5N1 , H6N2 , H7N1 , and H9N2 . The representative viruses for each respective subtype were as follows: rg-A/Hong Kong/213/2003 ( H5N1 ) , A/Chicken/CA/1255T/2002 ( H6N2 ) , rg-A/Canada/RV444/2004 ( H7N1 ) , and A/Duck/Hong Kong/Y280/1997 ( H9N2 ) . H7N3 and H16N3 phylogenies . Gene sequences used for construction of phylogenetic trees were either obtained from the Los Alamos Influenza Sequence Database or from samples used in this study . The sequences were then aligned , and the ends were trimmed to equal lengths using BioEdit sequence alignment editor software , version 7 . 0 . 5 [40] . The neighbor-joining algorithm was applied using PHYLIP software , version 3 . 66 [41] and 100 bootstrap replicates . The H16N3 tree was rooted using an HA segment isolated from a distantly related H11 strain as an outgroup , whereas the H7N3 tree was rooted using an HA segment isolated from an equine host as its outgroup . The final tree outfile was visualized using TreeView Win32 software [42] . Outsider analyses . Two sets of phylogenetic trees were constructed using previously described methods [26] to infer the relationships among individual avian flu gene segments . The initial set analyzed a total of 6 , 767 segments ( 560 PB1 , 586 PB2 , 585 PA , 1014 HA , 685 NP , 883 NA , 1187 M , and 1267 NS ) that were deposited in GenBank through April 2006 ( Figures S2–S9 ) . The second set of trees included 248 complete flu genomes , deposited into GenBank through February 2007 ( Figures S10–S17 ) . Any segment that was subsequently determined to be a duplicate sequence ( as determined by strain name [Table S3] ) , isolated from domesticated fowl ( Galliformes ) or obtained from a non-avian source ( e . g . , blow fly or human ) was excluded from further analyses . An outsider event was determined to have occurred if a strain isolated in one hemisphere was found to be clustered with a majority of strains from the opposing hemisphere . In the event that several closely related strains formed a clade within the strains from the opposing hemisphere , a single outsider event was determined to have occurred ( Table S4 ) . The 136 segments involved in the 56 outsider events , found in this study , represent 100 viral strains ( Table S1 ) . Only 68 of those strains represent cases where only a single segment underwent viral transfer; all of the segments , however , were considered for outsider analyses . Two large polyphyletic clades , containing either isolates from the Western Hemisphere or isolates from the Eastern Hemisphere , were generally identified for each gene segment . The exception to this was with HA and NA genes . Within both of these genes there are distinct amino acid sequences , which result in phylogenetically distinct subtypes . Since there were several cases in which certain subtypes of these genes were infrequently sequenced , it was decided that individual subtypes needed to be examined to avoid terming these infrequent sequences as outsider events ( Table S5 ) . In addition to grouping outsider events by gene segment , strains were additionally grouped by bird order . Eight known orders of wild birds were identified based on virus name , with unidentifiable birds grouped as “undetermined . ” Because the orders Anseriformes and Ciconiiformes represent the source of the samples collected for virus isolation in this study , we selected those orders for more detailed analysis .
Accession numbers for gene sequences produced by this study and deposited in GenBank ( http://www . ncbi . nlm . nih . gov/Genbank/index . html ) are EU030965–EU030988 . The accession numbers of gene sequences that were not generated by this study , but were used for the analysis of the complete genome of the H7N3 viruses and the HA gene of the H16N3 viruses , can be found in Table S6 .
|
Influenza surveillance in wild migratory birds has been done at two sites in North America: 1 ) in Alberta , Canada , for the past 31 years , and 2 ) along Delaware Bay , United States , for the past 22 years . These studies support the concept that wild migratory birds are the reservoirs of all influenza A viruses and that the influenza viruses in the world can be divided into two distinct superfamilies , one in Eurasia and the other in the Americas . From time to time these viruses spread to domestic poultry and to humans and cause pandemics of disease . Many investigators have expanded these studies particularly in Europe , Asia , and the Americas . The emergence of highly pathogenic H5N1 in Asia a decade ago and the continuing evolution and spread of these H5N1 viruses to the whole of Eurasia is a continuing problem for veterinary and human public health . The available evidence from Eurasia is that migratory birds can be infected and may be involved in local spread of the highly pathogenic H5N1 virus . The question addressed in the present study is why the highly pathogenic H5N1 influenza virus has not yet reached the Americas despite the overlap in migratory bird pathways , particularly in Alaska . Genomic analysis of influenza viruses from our repository failed to provide evidence of influenza viruses with their whole genome originating from Eurasia . However , we found occasional influenza viruses from North America with single or multiple genes that originated in Eurasia . Our interpretation is that while influenza viruses do exchange between the two hemispheres , this is a rare occurrence . Regardless , enhanced surveillance should be continued in the Americas in case this rare event occurs .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] |
[
"influenza",
"virology"
] |
2007
|
Influenza in Migratory Birds and Evidence of Limited Intercontinental Virus Exchange
|
The basidiomycete Ustilago maydis causes smut disease in maize . Colonization of the host plant is initiated by direct penetration of cuticle and cell wall of maize epidermis cells . The invading hyphae are surrounded by the plant plasma membrane and proliferate within the plant tissue . We identified a novel secreted protein , termed Pep1 , that is essential for penetration . Disruption mutants of pep1 are not affected in saprophytic growth and develop normal infection structures . However , Δpep1 mutants arrest during penetration of the epidermal cell and elicit a strong plant defense response . Using Affymetrix maize arrays , we identified 116 plant genes which are differentially regulated in Δpep1 compared to wild type infections . Most of these genes are related to plant defense . By in vivo immunolocalization , live-cell imaging and plasmolysis approaches , we detected Pep1 in the apoplastic space as well as its accumulation at sites of cell-to-cell passages . Site-directed mutagenesis identified two of the four cysteine residues in Pep1 as essential for function , suggesting that the formation of disulfide bridges is crucial for proper protein folding . The barley covered smut fungus Ustilago hordei contains an ortholog of pep1 which is needed for penetration of barley and which is able to complement the U . maydis Δpep1 mutant . Based on these results , we conclude that Pep1 has a conserved function essential for establishing compatibility that is not restricted to the U . maydis / maize interaction .
The initial step of pathogenic development for both necrotrophic and biotrophic fungal pathogens is the successful penetration of the plant surface . Penetration can occur directly via specialized infection structures , called appressoria , which promote the localized secretion of plant cell wall degrading enzymes or build up turgor and allow penetration through mechanical force . Alternatively , fungal pathogens may use natural openings like stomata or wounds for entry [1] . The infection strategy does not appear to be linked to the subsequent lifestyle of the fungal pathogen , i . e . necrotrophs like Botrytis cinerea as well as hemibiotrophs such as Colletotrichum ssp . and Magnaporthe grisea directly penetrate the plant surface via appressoria [2]–[4] . Some biotrophs like most rust fungi invade plant tissue via stomata , while other biotrophs like the smut fungi and the powdery mildew fungi form appressoria that allow direct entry into the plant epidermis [5] , [6] . Necrotrophic pathogens kill the invaded cell by secretion of toxic compounds or induction of reactive oxygen species ( ROS ) , and subsequently feed on dead plant material . In biotrophic interactions and during the initial stages of hemibiotrophic interactions the infected plant cell stays alive . In such interactions , the plant plasma membrane is invaginated and encases the infecting hyphae , thereby forming a biotrophic interface . This interface , which can be established by intracellularly growing hyphae or by specialized structures ( haustoria ) , provides nutrients to the pathogen and facilitates exchange of signals maintaining the interaction [5] , [7] . Compatibility in a biotrophic interaction requires the pathogen to overcome basal plant defense responses that are elicited by recognition of conserved pathogen associated molecular patterns ( PAMPs ) and which can lead to pathogen arrest [8] . This initial PAMP-triggered immunity needs to be overcome by successful pathogens that use secreted effectors to interfere with these processes , and use such effectors to trigger susceptibility . Effectors may also be specifically recognized by R proteins , leading to effector triggered immunity which is often associated with cell death [8] . Haploid U . maydis cells mate on the leaf surface and the resulting dikaryon switches to filamentous tip growth . The growing tip cell is separated from the older parts of the hypha by a septum , and the older septated hyphal parts appear empty and are often collapsed [9] . The need of two compatible wild type strains complicates generation of deletion mutants . Therefore , the solopathogenic strain SG200 [10] , which is a haploid strain engineered to carry composite mating type loci is frequently used . This strain forms filaments on the maize surface and causes disease without prior mating . On the leaf surface , SG200 as well as the dikaryon formed after mating of two compatible haploid wild type strains , develops non-melanized appressoria that directly penetrate the host tissue and establish a biotrophic interaction . Gene-for-gene systems , i . e . effectors that are specifically recognized by cognate resistance genes of the plant have not been described in this pathosystem . After penetration U . maydis grows intracellularly and during this stage the hyphae are surrounded by the host plasma membrane [9] , [11] . U . maydis does not develop haustoria [12] and the intracellular hyphae pass from one cell to the next . At later stages fungal hyphae accumulate in mesophyll tissue and are found mostly in apoplastic cavities that arise in the developing tumors [13] . In these tumors plant cells enlarge , undergo mitotic divisions and the hyphal aggregates differentiate into spores . The genome sequence of U . maydis revealed that this organism is poorly equipped with plant cell wall degrading enzymes [10] , which is in line with its biotrophic life style where the infected plant cells stay alive . However , U . maydis codes for a large set of novel secreted effectors [14] and many of the respective genes are arranged in clusters [10] . During biotrophic development , the majority of these clustered effector genes is upregulated [10] . 12 of these gene clusters encoding secreted proteins were deleted and five of the respective mutants were significantly altered in virulence . Deletion of one cluster resulted in increased virulence , while mutants of the four other clusters were attenuated in virulence and showed defects at different stages of pathogenic development [10] . However , none of these clusters was reported to be essential for the initial step of biotrophic development , the penetration of epidermal cells . Based on these studies it became clear that the repertoire of effectors with a function during disease was unlikely to be restricted to effectors whose genes reside in clusters . We have therefore initiated a systematic analysis of effector genes in U . maydis which is solely based on two criteria: the protein should carry a secretion signal and the predicted product should be novel i . e . should not match data base entries . Here we describe one of these novel effectors , Pep1 ( Protein essential during penetration 1 ) .
The pep1 gene ( um01987 ) resides on chromosome 3 of the U . maydis genome . pep1 is not part of a gene cluster , i . e . upstream we find a putative oxidoreductase ( um01988 ) and downstream a sterol carrier ( um01986 ) , two proteins not predicted to be secreted . The Pep1 protein comprises 178 aa and is expected to be cleaved behind a putative N-terminal secretion signal ( Figure 1A ) . Pep1 lacks known sequence motifs associated with enzymatic function and also lacks paralogs in the U . maydis genome as well as homologs in known published genome sequences . To study the function of pep1 , gene deletions were generated in the solopathogenic strain SG200 [10] . To elucidate whether Pep1 is needed for growth of U . maydis , SG200 and SG200Δpep1 strains were grown under conditions of nutrient deprivation , cell wall stress or oxidative stress . In addition , filamentation was tested on charcoal containing plates ( Figure S1 ) . Under none of the tested conditions we could detect differences between these four strains , illustrating that pep1 is not affecting growth under these conditions ( Figure S1 ) . To show that Pep1 is secreted we generated strain SG200Δpep1oma:pep1-GFP in which pep1-GFP is expressed from a strong constitutive promoter [15] . Using GFP specific antibodies , the full-length fusion protein was detected in the supernatant while supernatants of SG200 did not show a signal ( Figure S4A ) . Next , SG200Δpep1 was assayed for pathogenicity . The deletion of pep1 resulted in complete loss of tumor formation ( Figure 1B and Table S3 ) . To demonstrate that the mutant phenotype resulted from disruption of pep1 , the pep1 gene was introduced in single copy into the ip locus [16] , [17] of strain SG200Δpep1 . The resulting strain SG200Δpep1-pep1 was fully pathogenic and showed disease ratings similar to SG200 ( Figure 1B ) , indicating successful complementation . To examine at which stage of pathogenic development SG200Δpep1 is defective , we followed appressorium formation on inoculated maize leaves . For proper quantification of appressorium formation , GFP fluorescence of the AM1 marker , which is specifically upregulated in the hyphal tip cell forming an appressorium , was monitored [18] . 24 hpi SG200 and SG200Δpep1 strains had switched to filamentous growth and about 20% of SG200 filaments ( 19 . 73%±5 . 21; n = 1039 ) and a comparable percentage of SG200Δpep1 cells ( 19 . 76%±2 . 48; n = 1643 ) had developed appressoria . This demonstrates that the differentiation of appressoria does not require pep1 . In addition , 48 hpi we observed that a small percentage of SG200Δpep1 cells had engaged in multiple penetration attempts ( see below , Figure 3C ) , which was never observed after infections with SG200 . Furthermore , the multiple penetration attempts seen in the pep1 mutant suggest a defect in invasion of host tissue . To analyze this presumed invasion defect in detail , we used confocal microscopy to visualize the fungus in infected leaf tissue . For a better visualization of the infection process we infected the maize line ZmPIN1a-YFP that expresses a YFP-tagged version of the PIN1a protein which locates to the plant plasma membrane [19] . Fungal hyphae growing on and inside the plant tissue were detected by cytoplasmic expression of RFP under control of the otef promoter in the respective strains . At 24 hpi SG200rfp hyphae were already detected in epidermal cells and were encased by the plant plasma membrane . Since the fungal cytoplasm moved into the intracellularly growing hyphae , hyphal sections on the leaf surface did not contain cytoplasm any more ( Figure 2A , B ) . 24 hpi , hyphae of SG200Δpep1rfp could not be detected inside epidermal cells; instead , mutant hyphae were arrested immediately after penetration of the epidermal cell wall . The plant plasma membrane was found to be invaginated around mutant hyphal tips; however , no progression of mutant hyphae into the lumen of the epidermal cell was observed ( Figure 2C , D ) . To test whether the Δpep1 phenotype is also evident when haploid wild type strains are used instead of the solopathogenic SG200 strain , the pep1 gene was deleted in the two compatible U . maydis wild type strains FB1 and FB2 [20] . Maize plants infected with a mixture of FB1 and FB2 as well as a mixture of the deletion strains FB1Δpep1and FB2Δpep1 were analyzed by confocal microscopy 24 hpi . To visualize hyphae , they were stained by WGA-AF488 , plant structures were stained with propidium iodide . Similar to SG200Δpep1 , the FB1Δpep1/FB2Δpep1 dikaryon formed appressoria but penetration of epidermal cells was blocked after a short peg had entered the host cell ( Figure S3 ) . However , in rare cases , thin hyphae were found to grow into the epidermis cells and these plant cells then collapsed ( Figure S3C , D ) . Together , this shows that the deletion of pep1 results in a complete block of pathogenic development at the stage of host penetration . Leaf areas infected with U . maydis SG200 showed visible symptoms such as chlorosis , anthocyanin accumulation and small , primary tumors 4 dpi . In rare cases , small necrotic spots representing small clusters of dead cells developed ( Figure 3A; [21] ) . In contrast , leaves infected with SG200Δpep1 did not show chlorosis but displayed large necrotic areas 4 dpi ( Figure 1A , Figure 3A ) . Already 48 hpi SG200Δpep1 infected plants reacted with strong cell wall autofluorescence and formation of large papillae ( Figure 3B ) . In addition , accumulation of H2O2 could be shown by staining with diamino-benzidine ( DAB ) at sites where SG200Δpep1 attempted to penetrate while it was absent around appressoria of SG200 ( Figure 3C; [22] ) . To obtain a more comprehensive picture of the plant responses induced by the Δpep1 mutant , we performed microarray analyses of infected leaf tissue . In a previous study , the transcriptional responses of maize after infection with U . maydis strain SG200 have been described [21] . Using identical experimental conditions , we now compared expression profiles of SG200 infected leaves to SG200Δpep1 infected tissue 24 hpi using the Affymetrix maize genome array . At this stage , SG200 starts to establish the biotrophic interaction which goes along with a down-regulation of various defense-related genes [21] . In SG200 infected plants 24 hpi 116 genes were differentially regulated compared to mock-infections [21] . In contrast , in SG200Δpep1 infected plants 220 maize genes were found to be differentially regulated compared to mock-infected control tissue at the same time point ( Table S1 ) . In line with this , the expression of 110 maize genes was found to be significantly different ( fold change ≥2 ) in SG200Δpep1 infected tissue compared to SG200 infected tissue ( Figure 3D , Table S2 ) . In particular , defense related genes like PR6b ( Zm . 791 . 1 . S1_s_at ) , an endochitinase ( Zm . 16805 . 8 . S1_at ) and terpene synthase 6 ( Zm . 14496 . 1 . A1_at ) were strongly induced by SG200Δpep1 while in infections with SG200 the expression of these genes was already attenuated at this time point [21] . Interestingly , several genes associated with jasmonate biosynthesis like the lipoxygenase LOX1 ( Zm . 3303 . 1 . A1_at ) as well as several serine protease inhibitors that are typically activated by jasmonic acid ( JA ) [23] lack transcriptional induction in response to SG200Δpep1 . Induction of JA signaling is a typical feature of compatible biotrophic interactions [24] , i . e . its absence is therefore likely to indicate that the pep1 mutant is incapable of establishing a biotrophic interaction . To follow expression and localization of Pep1 during different developmental stages , the coding region of gfp was fused to the C-terminus of Pep1 . By homologous recombination , wild type pep1 was replaced by pep1:gfp resulting in strain SG200pep1:gfp . In addition , a strain was generated in which pep:gfp was introduced into the ip locus of SG200Δpep1 ( SG200Δpep1-pep1:gfpIP ) . As shown in Figure 1B , these strains were indistinguishable from SG200 with respect to causing disease , indicating that the C-terminal fusion of GFP to Pep1 did not impair its function . To follow expression of pep1:gfp during growth , strain SG200pep1:gfp was modified to additionally express cytoplasmic RFP under control of the otef promoter . In SG200pep1:gfpR , no GFP fluorescence could be detected during growth in liquid culture , whereas RFP was detected in the cytoplasm of all cells ( Figure 4A ) . When SG200pep1:gfpR was inoculated to maize leaves , Pep1-GFP expression appeared for the first time in penetrating hyphae ( Figure 4B ) . During intracellular growth , Pep1-GFP accumulated in a slightly uneven pattern around growing hyphae ( Figure 4C ) , at hyphal tips and particularly strong at hyphal tips during cell to cell passages ( Figure 4D , E ) . In addition , some intracellular fluorescence was detected which is likely to reflect Pep1 during processing through the ER ( Figure 4C , D ) . During tumor formation , i . e . 5–8 dpi , when U . maydis grows mainly intercellularly , Pep1-GFP could not be detected any more ( not shown ) . In addition , expression of pep1 was monitored by quantitative RT-PCR . In accordance to the microscopic observations , pep1 was not detected in axenic culture while the gene was expressed at the penetration stage 18 hpi ( Figure S2 ) . During biotrophic growth , high expression levels were detected at all timepoints tested from 2–8 dpi ( Figure S2 ) . Due to autofluorescence of maize cell walls especially at penetration sites and in tumor tissue [12] , [21] interference with the secreted GFP signal cannot be excluded . To overcome this problem , pep1 under control of its own promoter was fused to the rfp derivate mcherry [25] and introduced into the ip locus of strain SG200Δpep1 . Maize infections with the resulting strain SG200Δpep1-pep1M showed that the Pep1-mCherry fusion-protein was fully functional ( not shown ) . SG200Δpep1-pep1M was used to infect maize lines ZmPIN1a-YFP and ZmTIP1-YFP expressing either PIN1-YFP as plasma membrane marker or TIP1-YFP , an aquaporin localizing to the tonoplast membrane ( http://maize . tigr . org/cellgenomics/index . shtml ) . The Pep1-mCherry fusion protein was detected around intracellular hyphae , where it partially co-localized with the PIN1-YFP signal ( Figure 5A ) . At cell to cell passages of hyphal cells , Pep1-mCherry was observed to spread between the plasma membranes of adjacent cells ( Figure 5A , B ) . As we could not discriminate between Pep1-mCherry being localized in the plant cell wall or in the apoplastic space , plasmolysis of infected tissue was induced to enlarge the space around intracellularly growing hyphae . After plasmolysis Pep1-mCherry showed an even distribution in the now enlarged apoplastic space . mCherry fluorescence was not observed in cells which were not colonized by U . maydis ( Figure 5C , D ) . In addition to life cell imaging , strain SG200Δpep1-pep1:HA was generated and used for in situ immunolocalization of Pep1 . Similar to what has been observed with Pep1 fused to fluorescence tags , the protein was detected on the surface of intracellularly growing hyphae ( Figure 6A , B ) and had a patchy distribution . The strongest accumulation of Pep1-HA was observed at sites where fungal hyphae traversed from one plant cell to the next , consistent to what has been observed with fluorescently tagged Pep1 . Pep-HA could be isolated from infected maize leaves by immunoprecipitation with HA-specific antibodies and was found to be of the expected size ( Figure S4B ) . To isolate Pep1-mCherry from infected tissue , strain SG200Δpep1-pep1:MHA was generated in which Pep1-mCherry carries an additional C-terminal HA tag . Western blot analysis of the immunoprecipitated protein revealed a signal at the expected size of the full length fusion protein . In addition , two smaller fragments were detected ( Figure S4B , C ) . Since SG200Δpep1 is blocked already upon penetration of the leaf epidermis , the mutant could not provide information concerning a role of Pep1 at later stages of the interaction between U . maydis and its host . To address this , we infected maize plants with U . maydis expressing pep1-gfp under control of the otef promoter ( strain SG200Δpep1otefpep1:gfp ) . The artificial otef promoter exhibits strong , constitutive expression in haploid sporidia , penetrating filaments and during the early biotrophic phase of U . maydis but is shut down during the late biotrophic stage of U . maydis ( G . D . , unpublished observation ) . SG200Δpep1otefpep1:gfp was able to penetrate and grow intracellularly , demonstrating that expression of pep1 under the otef promoter rescued the penetration defect of the pep1 mutant ( Figure 7A , B ) . However , tumor formation was only partially rescued; visible symptoms caused by this strain were mainly anthocyanin production , chlorosis as well as necrosis and only very small tumors were observed ( Table S3 ) . Microscopic analysis of SG200Δpep1otefpep1:gfp infected leaves 7 dpi revealed an accumulation of fungal hyphae inside plant cells . Such hyphae displayed multiple appressorium-like structures indicating unsuccessful penetration attempts ( Figure 7C , D ) . From these results we conclude , that pep1 is not only needed for primary penetration of the leaf epidermis , but plays an essential role for cell-to-cell passage during the intracellular phase of biotrophic growth . After 454 sequencing of the genome of the barley covered smut fungus Ustilago hordei ( J . Schirawski and R . Kahmann , unpublished ) we identified an ortholog of pep1 that shows 61% identity to U . maydis Pep1 . Both proteins have an N-terminal secretion signal as well as four cysteine residues whose spacing is conserved ( Figure 8A ) . Calculation of the ratio of synonymous to non-synonymous substitutions ( ds/dn ) ( http://www . hiv . lanl . gov; [26] ) between Pep1 of both organisms revealed a ds/dn ratio of 4 . 67 , indicating a preference for amino acid conservation . This is particularly true for the central part of the protein that contains the conserved cysteine residues ( Figure 8B ) . To investigate whether Pep1 is also required for penetration in U . hordei , pep1 was deleted in the compatible U . hordei strains 4875-5 ( Mat1 ) and 8a ( Mat2 ) . Four days post infection of barley seedlings , growth of wild type and mutant strains was analyzed by confocal microscopy . After penetration , the U . hordei wild type strains displayed directed growth towards the vascular bundles ( Figure 8C ) . The U . hordei Δpep1 strains also managed to enter epidermal cells ( Figure 8D , E ) , but proliferation inside the plant tissue was never observed . Instead , the attacked epidermis cells underwent cell death which could be visualized by propidium iodide staining of disintegrated cells ( Figure 8D , E ) . To test whether U . hordei pep1 can substitute for U . maydis pep1 , the coding region of uh-pep1 was expressed in SG200Δpep1 under control of the um-pep1 promoter . The resulting strain was fully pathogenic towards maize ( Figure 8F ) , which illustrates that the two proteins are exchangeable . Pep1 does not contain conserved motifs which would allow a prediction of its mode of action . However , especially the C-terminus of U . maydis Pep1 is enriched in glycine residues . To test a putative function of this region , a truncated allele of pep1 ( pep1Δ141–178 ) was generated . This truncated pep1 allele was inserted in single copy in SG200Δpep1 and shown to restore wild type pathogenicity ( Figure 9A ) . Pep1 contains four conserved cysteine residues in the central part of the protein which might be involved in formation of disulfide bridges . Mutant alleles of pep1 were generated in which each cysteine residue of Pep1 was exchanged to serine . Mutant alleles containing substitutions in one cysteine residue ( pep1CS59; pep1CS75 ) , the first two cysteins ( pep1CS59 , CS75 ) and all four cysteine residues ( pep1CS59 , 75 , 94 , 112 ) were expressed in SG200Δpep1 . When single cysteine residues ( C59 or C75 ) were substituted , pathogenicity of the respective strain was reduced ( Figure 9A ) . The reduction was much more pronounced when C59 was mutated compared to the allele containing the C75 substitution . However , in both cases some tumors developed , indicating residual Pep1 activity ( Figure 9A ) . Substitution of both C59 and C75 led to a complete loss of pathogenicity similar to Pep1 in which all four cysteins were replaced by serine ( Pep1CS59 , 75 , 94 , 112 ) ( Figure 9A ) . To disclose the reason for this essential role of the cysteine residues , a pep1CS59 , 75:gfp fusion was introduced in U . maydis strain SG200 . The resulting strain SG200pep1:gfpCS59 , 75 which carries the endogenous pep1 gene and in addition pep1:gfpCS59 , 75 was used for maize infections . Microscopic analysis 2 dpi showed that the mutated Pep1 protein was expressed , but was found exclusively inside fungal hyphae ( Figure 9B–D ) . This could indicate the mutant Pep1CS59 , 75-GFP being destabilized and therefore degraded immediately after secretion . However , when comparing secreted Pep1-GFP to Pep1CS59 , 75-GFP , the mutant protein was significantly enriched inside fungal cells . In addition , accumulation of the protein at the hyphal tip was absent in case of Pep1CS59 , 75-GFP ( Figure 9D , E ) . We take this to indicate that the cysteine residues are necessary for secretion of Pep1 .
We have shown that Pep1 , a novel secreted effector protein of Ustilago maydis , is essential for successful invasion of maize plants . Expression of Pep1 was not observed under axenic culture conditions and the first stage where the protein could be detected coincided with penetration . The deletion of pep1 did not impair saprophytic development and also the overexpression of pep1 did not cause any alterations in growth , morphology or stress resistance . However , when pep1 was deleted U . maydis was unable to invade plant cells and failed to establish a compatible interaction with the host plant . In SG200Δpep1 , infection-related development like filamentation and appressorium formation were unaffected . Since Δpep1 mutant hyphae were found to invaginate the plant plasma membrane after appressorium formation , this must indicate that lysis of the plant cell wall itself is still possible when Pep1 is absent . This was even more evident when plants were infected with a mixture of compatible FB1Δpep1 and FB2Δpep1 strains . In this case the dikaryon formed short penetration pegs and this was associated with the collapse of the invaded cell . Similarly , the dikaryon of U . hordei Δpep1 strains initially penetrated the epidermal cell but was arrested , in the penetrated cell that underwent cell death . The finding , that the U . hordei pep1 can fully complement U . maydis pep1 mutants shows complete functional conservation of Pep1 in the both pathosystems . The slight difference in arrest point between U . maydis and U . hordei mutants is likely to be caused by different responses or cell wall composition of the two host plants . This is also supported by the observation that U . maydis is arrested in the first epidermal cell when non-host barley plants are infected ( G . D . , unpublished ) . It is obvious , that both U . maydis and U . hordei Δpep1-mutants are not defective in the ability to penetrate plant cell walls but fail to establish a biotrophic interaction immediately after entry into the host plant . Colonization of epidermal cells by biotrophic fungi requires the establishment of a biotrophic interface which mediates nutrient uptake and provides the contact zone where suppression of defense responses by the fungus takes place [7] . In infections with U . maydis strain SG200 early plant defense responses are induced and these are downregulated upon penetration [21] . In the absence of pep1 this downregulation was not observed , i . e . of the 37 defense related genes which were significantly repressed in the interaction with SG200 24 hpi [21] , 23 genes were found to be highly induced 24 hpi in SG200Δpep1 infected maize tissue . Another major difference concerned genes associated with JA signaling . These were strongly upregulated after infection with SG200 but not in response to SG200Δpep1 [21] . Similarly , two Bowman-Birk type trypsin inhibitors were highly induced after infection with SG200 but induction was absent after infection with SG200Δpep1 . For Bowman-Birk type trypsin inhibitor genes in rice it has been demonstrated that they are transcriptionally induced by JA but repressed by salicylic acid ( SA ) [27] . This suggests that the typical transcriptional response to biotrophic pathogens that coincides with elevated JA levels and a repression of SA signaling [24] is not established after infections with SG200Δpep1 . Moreover , production of ROS , papilla formation and the transcriptional induction of PR genes observed in response to SG200Δpep1 are typical for non-host responses in incompatible plant-pathogen interactions [28] . To understand the function of secreted effector proteins it is necessary to establish where they localize . Most extensive work on localization and function has been done on bacterial effectors which are translocated into the host cell via the type III secretion systems [29] . Remarkable advances have also been made in the oomycete field where many effectors carry a RXLR-EER motif that mediates translocation of effectors into the plant cell while a second group of effectors that lack this motif function in the apoplast [30]–[33] . The described secreted fungal effectors follow similar principles , i . e . they either have an apoplastic function or act inside the plant cell . However , the group of fungal effectors which are translocated to plant cells lack common motifs . Among these are M . grisea AVR-Pita , Uromyces fabae RTP1 and the flax rust effectors AvrM , AvrL657 , AvrP123; AvrP4 [34]–[37] . From these proteins only RTP1 was directly detected inside host cells by immuno-localization [36] . Transfer of the other fungal effectors was inferred from their ability to trigger cell death when expressed in the cognate resistant line or their interaction with a cytoplasmic resistance gene in yeast two-hybrid assays [7] . Apoplastic fungal effectors like Cladosporium fulvum effectors Avr2 , Avr4 and Ecp6 have been directly isolated from apoplastic fluid of infected tomato plants and several oomycete effectors were detected in isolated apoplastic fluid after antibodies had been raised [38]–[41] . Pep1 secretion from intracellularly growing hyphae could be shown by generating biologically active GFP and mCherry fusions and this did not require overexpression . Secretion of Pep1 and accumulations at sites where hyphae passage from cell to cell was confirmed by immunolocalisation of HA-tagged Pep1 protein . However , it was impossible to determine in which plant compartment Pep1 resides because of the tight encasement of the intracellular hyphae by plant plasma membrane . This problem could be solved by inducing plasmolysis , which allowed to detect Pep1:mCherry now in the drastically enlarged apoplastic space around intracellular hyphae . By immunoprecipitation full length HA-tagged Pep1 could be isolated from infected plant tissue . This contrasts the situation in tomato where it was not possible to recover affinity-tagged secreted effectors from infected plant after overexpressing the genes via a PVX system [42] . In addition , immunoprecipitations of a mCherry-HA tagged Pep1 allowed to demonstrate that a significant amount of full-lengh fusion protein could be isolated from infected plant tissue . However , some material being significantly smaller than mCherryHA ( which is therefore unlikely to show fluorescence ) was detected . Another fragment of about 35 kD is indicative of processing/degradation within the Pep1 part of the fusion protein . This was not observed when immunoprecipitating Pep1-HA and therefore we consider it likely that this form was generated during protein extraction . Since this clearly shows that no cleavage of full length mCherryHA from Pep1 occurs inside the plant tissue , we conclude that the observed fragments , even if they were present in the infected tissue , should not affect the Pep1-localization shown by fluorescence microscopy . Collectively , the presented data suggest an apoplastic localization of Pep1 . The elicitation of plant defense responses typically results in the massive accumulation of PR proteins in the apoplast [43] . Many of these PR proteins have enzymatic functions and ß-1 , 3 glucanases or proteases can directly harm the pathogen or degrade secreted effectors with the result of disabling the pathogen . For several fungal and oomycete effectors it has been demonstrated that they target such PR proteins: The C . fulvum effector protein Avr2 has been shown to inhibit the apoplastic tomato proteases RCR3 and PIP1 [40] , [44]; and Phytophtora infestans secretes several inhibitors for apoplastic proteases of tomato [38] , [45] , [46] . A different function has been shown for Avr4 , which prevents hydrolysis of fungal cell walls by plant chitinases [47] . While the role of individual protease inhibitors for disease progression has not been analyzed in Phytophtora , silencing of avr2 and avr4 leads to decreased virulence of C . fulvum on tomato [48] , [49] . Similarly the C . fulvum effector Ecp6 ( whose function is unknown ) is required for full virulence [39] . In contrast to these effectors which are virulence factors , Pep1 is essential for compatibility . When absent , U . maydis and U . hordei fail to establish a biotrophic interface . pep1 mutants are recognized by their respective host plants and elicit defense responses that are so strong that a host now acts as if it was a non-host . This , however , does not suffice as an explanation for host specificity . In this case we would have expected that all smuts that express pep1 should cause disease on the same host plants ( which is not the case ) . Therefore , we propose that pep1 affects compatibility on an early level that precedes the action of host specificity factors . Which is the molecular function of Pep1 ? At present , we can only speculate about its mode of action . Pep1 of U . maydis , which is predicted to comprise 152 aa after signal peptide cleavage , is unrelated to proteins or functional domains of described database entries . This makes it unlikely that Pep1 has an enzymatic function . A glycine-rich domain of 37 aa at the C-terminus was deleted without affecting biological activity . This domain is considerably less conserved between U . maydis Pep1 and U . hordei Pep1 than the central domain . Given the apoplastic localization and the importance of the four cysteine residues for secretion of Pep1 we consider a compact structure of Pep1 that requires disulfide bridge formation most likely . Fungal and Oomycete plant pathogens have been shown to secrete a broad range of putative enzyme inhibitors to counteract plant hydrolases and many of these are cysteine-rich and attain their compact structure through disulfide bridge formation [33] . Among these are small cysteine-rich apoplastic proteins like Avr2 , the EPI and EPIC proteins of Phytophtora that all target specific pathogenesis related plant proteases [38] , [44] , [46] , [49] . Another small effector of P . sojae specifically targets ß-1 , 3-glucanases of soybean [50] . Due to selective pressure , both , the genes encoding the plant enzymes and the genes encoding the fungal/oomycete inhibitors exist in large gene families . These features were proposed to provide robustness to the systems but at the same time limit the effects of individual genes due to redundancy [33] . With respect to Pep1 these criteria do not apply , i . e . paralogous genes for pep1 are neither found in U . maydis nor in U . hordei . We have not analyzed allelic variation , however , the degree of sequence conservation and the preference of synonymous nucleotide substitutions over non-synonymous substitutions in the central domain is remarkably high . This likely indicates that this domain adopts a defined structure that cannot be altered by mutation without affecting the function of the protein . And finally , the phenotype of pep1 deletion is dramatic , reinforcing the absence of redundant functions . Thus , if Pep1 is an enzyme inhibitor , we would predict that it should have little or no specificity , i . e . interacts with many enzyme isoforms . Fungal effectors like the C . fulvum protease inhibitor Avr2 which specifically interacts with two plant proteases shows strong diversifying selection , and this is likely the consequence of preventing recognition [44] . This contrasts the situation in Pep1 where we find a high conservation of the central domain which is essential for Pep1 function . Alternatively , Pep1 could act as a kind of chaperone protecting/activating other secreted effectors or facilitate the establishment of the fungal/host interface by binding toxic compounds or interfere with plant signaling . Solving the molecular structure of Pep1 and identification of interacting molecules will help to disclose its function and the processes it interferes with . As two-hybrid screens were unsuccessful , presumably due to incorrect folding of the protein ( G . D . , unpublished ) , biochemical approaches are now under way . The understanding of how Pep1 affects plant defense responses is likely to provide fundamental new insights into the initial steps that are required for the establishment of a compatible , biotrophic interaction between fungi and their host plants .
U . maydis SG200 [10] and its derivatives ( Table 1 ) were grown at 28°C in YEPSL ( 0 . 4% yeast extract , 0 . 4% peptone , 2% sucrose ) and used in plant infections as described [22] . Disease symptoms were scored 12 dpi as described previously [10] . Symptoms caused by SG200Δpep1 mutants were classified into the additional category “chlorosis/necrosis” . For growth assays , U . maydis strains were grown for 48 hours on plates containing CM agar supplemented with 1% glucose and various stress-inducing compounds whose concentrations are indicated ( Figure S1 ) . To induce filamentous growth , strains were cultured on PD agar containing 1% activated charcoal . U . hordei strains 4875-5 and 8a as well as their derivatives ( Table 1 ) were grown under the same experimental conditions as U . maydis . For infection of barley plants ( Golden Promise ) , cultures of the compatible strains were grown until an OD600 of 1 . 0 in YEPSL , and mixed prior to needle infection of barley plants 10 days post sawing . Barley plants of the variety Golden Promise were obtained from the IFZ ( Giessen , Germany ) . Maize lines of the variety Early Golden Bantam were obtained form Olds Seeds ( Madison ) . Maize lines ZmPIN1a-YFP and ZmTIP1-YFP were provided from Cold Spring Harbor Laboratory . All U . maydis strains generated in this study are derived from the solopathogenic strain SG200 and the wild type isolates FB1 and FB2 ( Table 1; [10] , [20] ) . For the deletion of pep1 ( Gene bank accession: XP_758134 ) a PCR-based approach using hygromycin as resistance marker [51] was used . 1 kb of each flanking region of pep1 were amplified by PCR using primers 5′-TTGGTGGACAGTCACGAGCATTC-3′ and 5′-TTCGGCCATCTAGGCCAC TCTGCTCGCCAGCATATCAC-3′ for the left border and primers 5′-CACGGCCTGAGTGGCCCAACTGCTTTCTGCCCTTTG-3′ and 5′-TTTCA GGGCAGCTCAGAGTG-3′ for the right border . PCR products were digested with SfiI and ligated to the hph cassette of pBS-hhn [51] . For integrations into the ip locus of U . maydis , plasmids derived from p123 were used [52] . For cytoplasmic rfp expression under control of the otef promoter , p123-rfp [53] was introduced into the ip locus of strains SG200 , SG200Δpep1 and SG200pep1:gfp , respectively . To substitute pep1 by pep1:gfp , 1 kb of U . maydis genomic sequence containing the coding region of pep1 was amplified by PCR as left border using primers 5′-GCAAGCCTAGCAATCTTCGATAGC-3′ and 5′-CACGGCCGCGTTGGCCCCGGTGGCGATCGAGCGCATGCCAAACATGCTACCGATTCC-3′ , digested with SfiI and ligated to the gfp:hph cassette of plasmid pUMa317 [54] . As right border , 1 kb including the terminator region of pep1 was amplified by primers 5′-CACGGCCTGAGTGGCCGCTGCGACGTCGTTGATGATGAC-3′ and 5′-CTCCACTCAAGACTCACAGACT-3′ , digested with SfiI and ligated to the gfp:hph cassette of plasmid pUMa317 . For complementation of SG200Δpep1 , the pep1 gene with its complete promoter region was amplified using primers 5′-GCAAGCTTACGACGGATGCGCTATCGTCAC-3′ and 5′-TAGCGGCCGCCTGG CGAGCAGAGTCATCATCAAC-3′ and ligated into the HindIII and NotI sites of vector p123 resulting in p123-pep1 . To complement SG200Δpep1 with pep1 pep1Δ141–178 , the truncated pep1 coding region with its complete promoter region was amplified using primers 5′-GCAAGCTTACGACGGATGCGCTATCGTCAC-3′ and 5′-TTGCGGCCGCTTGGCTTGAACCGCATCGTAAGC-3′ and ligated into the HindIII and NotI sites of vector p123 which resulted in plasmid p123- pep1Δ141–178 . To introduce pep1:gfp into the ip locus , plasmid p123-pep1:gfp was constructed by amplifying the pep1 gene using primers 5′-GCAAGCTTACGACGGATGCGCTA TCGTCAC-3′ and 5′-CACCCATGGCGGTGGCGATCGAGCGCATGCCAAACA TGCTACCGATTCC-3′ , and ligating the PCR product via HindIII and NcoI into p123 . To express pep1:gfp under control of the otef promoter , the coding region of pep1 was amplified using primers 5′-ATGGATCCGATGATG ACCACACTGGTGCAAAC-3′ and 5′-CACCCATGGCGGTGGCGATCGAGC GCATGCCAAACATGCTACCGATTCC-3′ . The PCR product was digested with BamHI and NcoI and ligated to the respective sites in p123 resulting in plasmid p123-otefpep1:gfp . The C-terminal HA-tag was introduced by amplification of the pep1 with primer 5′-GCAAGCTTACGACGGATGCGCTATCGTCAC-3′ and primer 5′-TAGCGGCCGCTCAGGCATAGTCGGGGACGTCGTAGGGATAGCCGCCCGACATGCCAAACATGCTACCGATTC-3′ which contains the HA-tag encoding sequence . This PCR product was digested with HindIII and NotI and ligated into p123 resulting in plasmid p123-pep1HA . To fuse pep1 with mcherry , plasmid p123-mcherry was constructed by excision of the gfp coding region from p123 using NcoI and NotI and substitution by mcherry derived from plasmid pCRII-mcherry ( kindly provided by G . Steinberg ) . Similarly , for mcherry::HA constructs , mcherry was amplified by primer 5′-CTCCATGGTGAGCAAGGGC-3′ and primer 5′-CTGCGGCCGCTTAAGCGTAATCTGGAACATCGTATGGGTACTTGTAC AGCTCGTCCATGCCGC-3′ that contains the HA sequence and introduced into the NcoI and NotI sites of p123 and subsequently fused to pep1 as described for p123-pep1:gfp . To express U . hordei pep1 in SG200Δpep1 , the coding region of uhpep1 was amplified with primers 5′-TTGATATCAACGATGAAGCTCAC ACTCAACACCG-3′ and 5′-TTGCGGCCGCTCAGAGCCCAACCATCTTACC-3′ genomic DNA of U . hordei strain 4875-5 . The PCR product was digested with EcoRV and NotI and ligated with EcoRV / NotI digested PCR product of primers 5′-ACCGCTGCGACGTCGTTGATGATG-3′ and 5′-GTCGAGAGTCCTCAG GATGGTTC-3′ that facilitate an inverse amplification of p123-pep1 without the U . maydis pep1 coding region . Standard molecular techniques were used [55] . Transformation of U . maydis and isolation of genomic DNA was performed as described previously [56] . All generated constructs were sequenced prior to U . maydis transformation . Isolated U . maydis transformants were tested for single integration events in the desired loci by southern analysis . To substitute cysteine residues in pep1 by serine , single point mutations were introduced in plasmid p123-pep1 using the “Quick Change Multi” site directed mutagenesis kit ( Stratagene , La Jolla , USA ) . Introduced mutations were confirmed by sequence analysis . For the Affymetrix microarray experiments , maize plants ( Early Golden Bantam ) grown in a phytochamber were infected with SG200Δpep1 as described previously and samples of infected tissue were colleted 24 hpi , 1 h before the end of the light period and directly frozen in liquid nitrogen [21] . Samples were collected in three independently conducted experiments by sampling 30 plants per experiment . For RNA isolation , material from the 30 plants was pooled , ground in liquid nitrogen and RNA was extracted with Trizol ( Invitrogen , Karlsruhe , Germany ) and purified using an RNeasy kit ( Qiagen , Hilden , Germany ) . Affymetrix Gene chipR maize genome arrays were done in three biological replicates , using standard Affymetrix protocols ( Midi_Euk2V3 protocol on GeneChip Fluidics Station 400; scanning on Affymetrix GSC3000 ) . Expression data were submitted to GeneExpressionOmnibus ( http://www . ncbi . nlm . nih . gov/geo/ ) ( Accession Number: GSE12892 ) . Data analysis was performed using Affymetrix GCOS1 1 . 4 , bioconductor ( http://www . bioconductor . org/ ) and dChip1 . 3 ( http://biosun1 . harvard . edu/complab/dchip/ ) , as described ( Doehlemann et al . , 2008b ) . We considered changes >2-fold with a difference between expression values >100 and a corrected p-value<0 . 001 as significant . Expression of pep1 was analyzed by qRT-PCR . RNA samples were isolated with Trizol as described above . To isolate U . maydis cells during the penetration stage 18 hpi from the maize leave surface , infected leaves were coated by liquid latex . The latex was dried and then peeled from the leaves . Peeled latex , containing the fungal structures extracted from the leaf surface was then used for RNA-isolation as described above . For cDNA synthesis , the SuperScript III first-strand synthesis SuperMix assay ( Invitrogen , Karlsruhe , Germany ) was employed , using 1 µg of total RNA . qRT-PCR was performed on a Bio-Rad iCycler using the Platinum SYBR Green qPCR SuperMix-UDG ( Invitrogen , Karlsruhe , Germany ) . Cycling conditions were 2 min 95°C , followed by 45 cycles of 30 sec 95°C / 30 sec 61°C / 30 sec 72°C . Control gene primers for amplification of the U . maydis peptidylprolyl isomerise ( ppi ) were rt-ppi-for: 5′-ACATCGTCAAGGCTATCG-3′ and rt-ppi-rev: 5′- AAAGAACACCGGACTTGG-3′ . To amplify a pep1 PCR-fragment , primers rt-pep1-for: 5′- CACTGACGACGACACCT-3′ and rt-pep1-rev: 5′- TGCTACCGATTCCTCCT-3′ were used . Fungal hyphae were stained with WGA-AF 488 ( Molecular Probes , Karlsruhe , Germany ) . Plant membranes were visualized using Propidium Iodide ( Sigma ) : Samples were incubated in staining solution ( 1 µg/ml Propidium Idodide , 10 µg/ml WGA-AF 488; 0 . 02% Tween20 ) for 30 min and washed in 1× PBS ( pH 7 . 4 ) . Visualization of H2O2 by DAB was performed as described [22] . Confocal images were recorded on a TCS-SP5 confocal microscope ( Leica , Bensheim , Germany ) ; using WGA-AF 488: excitation at 488 nm and detection at 500–540 nm . Autofluorescence of cell wall material was excited at 405 nm and detected at 415–460 nm . For mCherry fluorescence of hyphae in maize tissue , an excitation of 561 nm and detection at 580–630 nm was used . GFP fluorescence was excited with a 488 nm laser , emission was detected at 495–530 nm . YFP fluorescence of tagged plant proteins was excited at 495 nm and detected at 510–550 nm . For immunoprecipitation of Pep1-HA and Pep1-mCherry-HA from infected maize tissue , infected areas of 60 plants were excised 3 dpi after infection with the respective U . maydis strains and directly frozen in liquid nitrogen . Frozen leaves were ground in liquid nitrogen , mixed with extraction buffer and centrifuged for 30 min at 28100g . All samples were adjusted to a protein concentration of 2 . 4 mg/ml in a volume of 7 , 5 ml and mixed with 10 µl HA-matrix ( Roche ) for 16 h at 4°C on a shaker . Elution was performed according to the HA-Kit protocol ( Pierce ) . Overnight cultures of U . maydis strains SG200 and SG200Δpep1oma:pep1-gfp were harvest by centrifugation , washed once and were resuspended in 50 ml NM media containing 0 , 5% glucose to an OD600 nm of 0 , 20 and grown at 28°C to an OD600 nm of 0 . 80 . Cells were harvest by centrifugation , the supernatant was collected and percipitatetd by TCA . Then the pellets were washed seven times with 80% icecold acetone and resuspendet in 30 µl SDS loading buffer . All protein samples were separated by SDS-PAGE and transferred to a nitrocellulose membrane . After electroblotting , filters were saturated with 5% non-fat dry milk in TBS ( 20 mM Tris-HCl , 137 mM NaCl , pH 7 . 6 ) , 0 . 1% Tween for 1 hr at room temperature ( RT ) . For detection of Pep1-GFP , a monoclonal GFP specific antibodies ( Clontech , Mountain View , USA ) was used ( 1∶10000 ) . To detect HA-tagged proteins , a monoclonal mouse-anti-HA antibody ( Sigma-Aldrich ) ( dilution 1∶7500 ) was used . As secondary antibody an anti-mouse peroxidase conjugate ( 1∶10000 ) ( Sigma-Aldrich ) was used . For chemiluminscence detection , ECL Plus Western Blot detection reagent ( GE Healthcare ) was used . For in situ detection of Pep1-HA , maize leaves were harvested three days after infection with SG200pep1HA . Infected tissue was treated as described previously [57] . For detection of the HA-tag , a monoclonal mouse-anti-HA antibody ( Sigma-Aldrich , dilution 1∶7500 ) was used . As secondary antibody , anti-mouse conjugated with AF488 ( Molecular Probes ) was used in a 1∶5000 dilution . Confocal microscopy of the samples was done as described above . Control samples were maize leaves infected with SG200 and these were treated identical to SG200pep1HA infected tissue to verify Pep1-HA detection . In another control , SG200 infected leaves were used for detection of maize tubulin ( mouse-anti-tubulin; Sigma-Aldrich , dilution 1∶7500 ) . In both control samples , plant structures showed the same background , but no fluorescence of fungal hyphae was detected ( Figure S5 ) .
|
For many fungi that infect plants , successful invasion is coupled to a series of differentiation steps that are necessary to breach the plant cuticle . Such fungi form specialized infection structures which allow direct penetration of the plant cuticle . The smut fungus Ustilago maydis establishes a biotrophic interaction with its host plant maize in which the infected host cells stay alive . During biotrophic growth , the intracellularly growing hyphae are encased by the host plasma membrane . We show here that a small effector protein , which is secreted by fungal hyphae during penetration , is absolutely essential for fungal entry into plant cells . When this effector is absent , hyphal cells penetrate the plant cell wall and invagination of the plant plasma membrane is observed , but any further fungal development is arrested . This arrest coincides with the induction of massive plant defense responses . Thus , this effector , which is conserved in related fungal species , plays an essential role in suppression of plant defense responses and is critical for establishing compatibility . This is the first example where a single effector protein assumes such a crucial role for infection-related development in a plant pathogenic fungus .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"microbiology/plant-biotic",
"interactions",
"plant",
"biology/plant-biotic",
"interactions",
"genetics",
"and",
"genomics/gene",
"expression"
] |
2009
|
Pep1, a Secreted Effector Protein of Ustilago maydis, Is Required for Successful Invasion of Plant Cells
|
Computational neutralization fingerprinting , NFP , is an efficient and accurate method for predicting the epitope specificities of polyclonal antibody responses to HIV-1 infection . Here , we present next-generation NFP algorithms that substantially improve prediction accuracy for individual donors and enable serologic analysis for entire cohorts . Specifically , we developed algorithms for: ( a ) selection of optimized virus neutralization panels for NFP analysis , ( b ) estimation of NFP prediction confidence for each serum sample , and ( c ) identification of sera with potentially novel epitope specificities . At the individual donor level , the next-generation NFP algorithms particularly improved the ability to detect multiple epitope specificities in a sample , as confirmed both for computationally simulated polyclonal sera and for samples from HIV-infected donors . Specifically , the next-generation NFP algorithms detected multiple specificities in twice as many samples of simulated sera . Further , unlike the first-generation NFP , the new algorithms were able to detect both of the previously confirmed antibody specificities , VRC01-like and PG9-like , in donor CHAVI 0219 . At the cohort level , analysis of ~150 broadly neutralizing HIV-infected donor samples suggested a potential connection between clade of infection and types of elicited epitope specificities . Most notably , while 10E8-like antibodies were observed in infections from different clades , an enrichment of such antibodies was predicted for clade B samples . Ultimately , such large-scale analyses of antibody responses to HIV-1 infection can help guide the design of epitope-specific vaccines that are tailored to take into account the prevalence of infecting clades within a specific geographic region . Overall , the next-generation NFP technology will be an important tool for the analysis of broadly neutralizing polyclonal antibody responses against HIV-1 .
The HIV-1 Env glycoprotein , the sole target of antibody responses on the surface of the virus , exhibits extreme levels of sequence diversity [1–3] , possibly explaining why antibodies capable of broad and potent neutralization of the virus have been found to target only a small set of conserved Env sites of vulnerability [1 , 4] . A large number of broadly neutralizing HIV-1 antibodies ( bNAbs ) have been isolated within the last decade [5–20] and have been shown to be useful in preclinical studies of therapy and prevention [21–30] . Yet , no HIV-1 vaccine capable of eliciting such bNAbs is currently available . With the limited success of traditional vaccinology approaches , significant effort has been devoted to “rational” vaccine design based on understanding and manipulating the interactions between bNAbs and HIV-1 [31–34] . The identification and characterization of antibodies from infected or vaccinated individuals provides insights into the specifics of the antibody response against the virus [35–47] and can help generate templates for antibody-specific vaccine design [48 , 49] . A challenge to the field is that neutralizing antibody responses to HIV-1 infection or vaccination are complex and difficult to deconvolute , often comprised of diverse bNAb lineages targeting a variety of epitopes on Env [4 , 12 , 18 , 50] . Mapping the epitope specificities of polyclonal HIV-1 antibody responses therefore requires substantial effort . Standard epitope mapping methods include experimental techniques such as binding competition with monoclonal antibodies , neutralization or binding of Env variants containing epitope-specific knockout mutations , and neutralization blocking by epitope-specific antigens [9 , 50–57] . These methods often fail to yield definitive answers , particularly when more than one specificity is targeted by the serum , or when the true epitope target has not yet been defined . In addition to these experimental methods , mapping of antibody responses can be achieved through computational analysis of the neutralization of diverse HIV-1 strains by serum or plasma [4 , 58] . Previously , we developed and validated the NFP ( neutralization fingerprinting ) algorithm for delineating antibody specificities in polyclonal sera [4] . The NFP algorithm uses a reference set of monoclonal antibody neutralization fingerprints ( the potency pattern with which an antibody neutralizes a set of HIV-1 strains ) in order to estimate the relative contribution of different types of known antibody specificities to the neutralization by a given polyclonal serum [4] ( S1A Fig ) . Since serum neutralization data is typically obtained in the very first steps of serum characterization , algorithms like NFP can be substantially more efficient , and may also present advantages and improved ability for detecting antibody specificities in polyclonal sera , compared to standard methods [4] . NFP has been successfully used for mapping the antibody specificities in previously uncharacterized sera , for selecting donors for the isolation of broad and potent HIV-1 neutralizing antibodies , and for analysis of the dominant antibody specificities found during different stages of infection [1 , 4 , 12 , 14 , 35 , 59] . Here , we present next-generation serum mapping algorithms that address some of the major challenges for the NFP approach , substantially improving the accuracy of the computational predictions and enabling the exploration of new biological questions ( Figs 1A and S1B ) . Specifically , the next-generation NFP approach involves the following new algorithm developments ( S1B Fig ) : First , we developed a method for simulating HIV-1 neutralization by polyclonal antibodies in order to circumvent the limited availability of serum neutralization data from donors with well-defined specificities . Using the simulated neutralization as a quantitative benchmark for algorithm performance , we further developed algorithms for optimizing the selection of HIV-1 strains for virus neutralization assays , as well as for identifying sera with potentially novel antibody specificities and for estimating the confidence in the computational predictions for a given serum . The significance of these new NFP algorithm capabilities was confirmed through analysis of both simulated and real donor sera . Finally , extending from the analysis at the individual donor level , we developed algorithms that specifically enable polyclonal antibody analysis for entire cohorts . To confirm the utility of the next-generation NFP algorithms , we performed cohort-level analysis of a diverse set of HIV-infected donors . Our analysis suggested a potential connection between plasma clade and types of elicited antibody specificities , and therefore has direct implications for the design of epitope-specific HIV-1 vaccines . Overall , the next-generation NFP technology provides a valuable resource for analysis of broadly neutralizing antibody responses to HIV-1 infection .
The ability to compare and evaluate the performance of computational algorithms requires the development of quantitative benchmarks . In the case of NFP , success is determined by the ability of the algorithms to delineate component antibody specificities in polyclonal sera . Real donor sera , however , are not suitable as quantitative benchmarks because the actual antibody composition of a given polyclonal serum cannot be fully defined with current technologies , even for sera with extensively characterized antibody specificities . An alternative solution would be to utilize mixtures ( or combinations ) of monoclonal antibodies that can , to an approximation , serve as a proxy for polyclonal serum analysis . First , we analyzed published data for combinations of 2 , 3 , and 4 bNAbs with different epitope specificities [60] , and applied the NFP algorithm to delineate the component antibody specificities from the bNAb combination data ( S2A Fig ) . For all but one of the 21 combinations , including the two 4-bNAb combinations , the top-scoring NFP predictions matched the actual component antibody specificities , suggesting that such monoclonal combination data could be used as a benchmark for the NFP algorithm optimization . Since experimental data for bNAb combinations is limited , we developed a method for large-scale simulation of polyclonal neutralization data ( Figs 1B and 1C and S2B and S2C ) . Specifically , simulated sera were generated as pairwise bNAb combinations that mixed equal amounts of an antibody representative from two of the ten epitope-specific antibody groups ( see Materials and Methods ) . The simulation method showed excellent agreement with experimental data for a set of 11 pairwise bNAb combinations , with Spearman correlations between simulated and experimentally determined neutralization data ranging from 0 . 88 to 0 . 98 ( median of 0 . 97 ) ( Figs 1B and S2B ) . Given these results , we proceeded to generate large-scale sets of simulated neutralization data for different pairwise bNAb combinations , to be used as a quantitative benchmark for NFP algorithm evaluation ( Figs 1C and S2C ) . While many polyclonal responses against HIV-1 can be attributed to known antibody specificities , new broadly neutralizing epitopes continue to be discovered [11 , 12 , 19] and there are samples whose neutralizing activity cannot be explained by known specificities [35 , 70] . The first-generation of NFP algorithms , however , are only capable of predicting specificities that are already part of the reference set of known antibodies ( S1A Fig ) . To address this challenge , we sought to develop algorithms for predicting the existence of dominant potentially novel bNAb specificities in a given sample . To that end , we used virus panel f61 to generate simulated sera with neutralization patterns that had virtually no correlation with any of the neutralization fingerprints from the reference bNAb set ( see Materials and Methods ) , and applied the NFP algorithm to deconvolute these sera . We used normalized residual scores as a measure of the fit between the neutralization fingerprints of the reference bNAbs and the neutralization pattern of a given serum ( see Materials and Methods ) . The residual scores for the sera with unknown specificities were significantly different from the scores for the test set of sera with known specificities ( p<0 . 0001 , Mann-Whitney test ) , with only minor overlap between the two distributions ( Fig 5A ) . In fact , ~95 . 5% of the sera with known specificities had residual scores of at most -0 . 1 ( median: -0 . 7704; range: -1 . 617 to 0 . 2605 ) , whereas more than 98% of the sera with unknown specificities had residual scores greater than -0 . 1 ( median: 0 . 2172; range: -0 . 2719 to 0 . 5527 ) . These results confirmed that high NFP residual scores can indicate the existence of potentially novel specificities in a given serum . Analysis of the predicted signals for the simulated sera with unknown specificities revealed that b12-like signals were by far most common ( ~40% ) ( Fig 5B ) . This was in agreement with the observation that the highest false positive rate for the sera with known specificities was for b12-like signals ( Fig 3B ) . To determine whether some specificities are more likely to be predicted as particular other specificities , we analyzed the correspondence between false negatives and false positives in the set of simulated sera ( S4 Fig ) . The results indicated that , for most of the other specificities , b12-like signals were the most common false positives for sera with false negatives . To determine whether the assignment to b12-like signals may preferentially occur for limited-breadth sera that may share a large number of neutralization-resistant values with the b12 fingerprints , we analyzed the variation in observed signals for increasing serum neutralization breadth for the sera with unknown specificities ( S5 Fig ) . The prediction frequency of b12-like signals was generally similar for all levels of serum neutralization breadth , and similar consistencies were observed for the other nine specificities , with b12-like signals being the most common for all levels of serum neutralization breadth ( S5 Fig ) . Taken together , at least in the case of virus panel f61 , these results may indicate that the b12-like neutralization fingerprint can play a role as a sink for neutralization signals that may not be b12-related but that are difficult for the NFP algorithm to deconvolute . Even if a given polyclonal serum is dominated only by already known antibody specificities , the NFP algorithm may in some cases still result in inaccurate predictions , in the form of , e . g , false positives or false negatives ( Fig 3 ) . It would therefore be helpful to develop additional algorithms that can estimate the confidence in the NFP predictions for a given serum . To that end , we focused our analysis on the set of simulated sera with known specificities , in order to identify potential markers of NFP prediction confidence . The residual scores discussed above appeared to be a good indicator of prediction confidence , with high residual scores typically associated with high serum delineation error ( Fig 5C ) . In addition , the median of the delineation scores for a given serum was also found to be associated with the magnitude of the serum delineation error ( Fig 5D ) . Finally , we observed that the frequency with which random neutralization fingerprint signals were observed for a given serum was also generally associated with how accurate the computational predictions are ( Fig 5E ) . The correlation between serum delineation error with each of the three measures was significant ( p<0 . 0001 , Spearman ) . Moreover , each of the three measures was able to contribute to the identification of less reliable predictions not identified by the other two measures ( Fig 5F ) . To interrogate the usefulness of these scores for real data , we analyzed the results from panel f61 for donor and rhesus macaque sera ( Fig 4E ) , and observed that only the predictions for donors CAP256 , C38 , and CAP255 were associated with lower confidence ( S7 Fig ) , in agreement with the existence of a relatively strong unconfirmed signal in these predictions ( Fig 4E ) . The scores for the other five donors were associated with high confidence ( S7 Fig ) , reflecting the agreement with the dominant signals for experimentally confirmed specificities for these donors ( Fig 4E ) . Taken together , these results suggest that residual scores , median of delineation scores , and frequency of random signals can all be useful as criteria for determining the confidence in the NFP predictions . Knowledge about the common specificities found in HIV-1 infected individuals and participants in vaccine trials can provide valuable insights into the interactions between virus and host at the cohort , or even population , level . While it is possible to use experimental mapping to delineate the antibody specificities observed in large cohorts [70] , the efficiency of the neutralization-based NFP approach makes it particularly well-suited to such large-scale , cohort-level , analysis . To assess the applicability of NFP to cohort-level analysis , we developed a simulation procedure in which multiple cohorts of simulated sera were generated , and the ability of NFP to predict the overall prevalence of the different antibody specificities within each cohort was evaluated ( see Materials and Methods ) . Initially , we focused on simulated cohorts in which half of the simulated sera had dominant unknown antibody specificities , while the remaining sera in the cohort was divided in 1:1 ratio of sera with one vs . two dominant known specificities ( Fig 6A ) . Since f61 was not a subset of previously published large-panel serum-virus neutralization data [63] , we applied our virus panel search algorithm and identified an optimized 50-strain panel for use in the cohort-level analysis ( see Materials and Methods ) . We observed that using the original , first-generation , NFP algorithms with the optimized virus panel resulted in slightly better overall accuracy compared to the published 21-strain panel ( Fig 6A , left and middle ) . To improve the results further , we applied the next-generation NFP algorithms , which use the same calculation of delineation scores as the first-generation NFP algorithms , and then incorporate filters based on residual scores , median of delineation scores , and frequency of random signals as described above . In addition , we applied a procedure specifically designed to incorporate expected algorithm accuracy into the cohort-level analysis ( see Materials and Methods ) . This next-generation NFP analysis on average resulted in excellent agreement between predicted and actual prevalence of the different antibody specificities within the samples of simulated sera ( Fig 6A , right ) . Notably , while the selection of an optimized virus panel resulted only in a slight improvement of the accuracy of the cohort-level analysis ( p = 0 . 2754 , Wilcoxon ) , the use of the next-generation ( as opposed to the first-generation ) NFP algorithms led to a dramatic improvement in the results ( p = 0 . 002 , Wilcoxon ) . Since the actual relative occurrence of donor sera with unknown specificities vs . sera with known specificities is unclear and may vary depending on the specific cohort ( e . g . , duration of infection , dominant HIV-1 subtypes , etc . ) , we next assessed different distributions of the simulated serum samples , ranging from 0% to 50% occurrence of sera with unknown specificities , with the remaining sera divided between sera with one vs . two dominant specificities ( Fig 6B ) . The first-generation NFP algorithms with the optimized virus panel outperformed the published 21-strain panel for all of the different sample distributions ( Fig 6B , left and middle ) . In both cases , the worst performance was observed for samples with a high percentage of sera with unknown specificities combined with sera with a single dominant known specificity . As expected , the best performance was associated with samples consisting only of sera with one dominant known specificity ( and no sera with unknown or two known dominant specificities ) ( Fig 6B , left and middle ) . In contrast , the next-generation NFP algorithms on average showed excellent accuracy for all of the different sample distributions ( Fig 6B , right ) . Taken together , these results suggested that the next-generation NFP algorithms are appropriate for cohort-level analysis of antibody responses against HIV-1 . We next sought to apply the next-generation NFP algorithms to the analysis of plasma samples from a large collection of HIV-infected donors from diverse clades of infection . To that end , we compiled published neutralization data for ~200 donors [63] , which was then reduced to a final set of 143 donors with neutralization breadth of at least 30% ( Figs 7 and S8 ) . Approximately one fifth of the sera were predicted to have dominant potentially novel specificities ( associated with high residual scores ) , while approximately one third of the sera were predicted to have one or more dominant known specificities based on good confidence scores ( Fig 7A and 7B , and Materials and Methods ) . The predictions for the remaining approximately one half of all sera were deemed inconclusive , since at least one of the respective confidence scores was not within the defined threshold ( Figs 7A and S8C , and Materials and Methods ) . For the sera with predicted known specificities , all of the different types of known antibody specificities were observed , albeit with different frequencies ( Fig 7B ) . Moreover , the specificity predictions appeared to be associated , to an extent , with the clade of the donor’s infection ( Fig 7B ) . In the case of clade C plasmas , PGT128-like specificities were the most common , followed by PG9-like and 10E8-like specificities , although the small number of clade-C samples may be a limiting factor in this analysis . In contrast , 10E8-like signals were by far the most common in clade-B samples ( Fig 7B ) . In fact , although 10E8-like specificities were predicted in both clade-B and non-clade-B samples , an enrichment of 10E8-like signals was observed in clade-B samples ( Fig 7C ) . Overall , for the collection of samples analyzed here , 10E8-like specificities were the most common , followed by PGT128-like and PG9-like specificities ( Fig 7D ) . Taken together , these results show the potential of using the NFP approach for large-scale analysis of antibody responses against HIV-1 .
Computational analysis of virus neutralization data is now an established method for mapping the antibody specificities found in polyclonal responses against HIV-1 . The success of NFP and other computational approaches for analyzing the interactions between HIV-1 and the immune system [4 , 41 , 58 , 71] underlines the power and potential of computation to develop transformative technologies with significant biological impact . One of the major challenges for such approaches , however , has been the limited amounts of available data . In the case of NFP , while matrices of several hundred virus strains against several hundred monoclonal or polyclonal antibodies are pushing the boundaries of what is feasible experimentally , advanced computational techniques such as machine learning would benefit immensely from the existence of substantially larger training datasets . To address this issue , we developed a framework for simulating polyclonal antibody responses . While antibody responses to HIV-1 are complex and polyclonal [72] , there is limited to no evidence to suggest that within any given serum there are a large number of different antibody specificities that each achieve broad neutralization [12 , 14 , 18 , 66] . We therefore elected to utilize simulations of pairwise monoclonal antibody combinations in the current study , although the simulation framework here can be extended to accommodate higher-order polyclonal simulations . Advances in the modeling of neutralization data ( e . g . , [73] or the addition of modeling for antibody synergism or antagonism ) can lead to further improvements in the simulation framework , and consequently , the accuracy of the predictive algorithms . Overall , the ability to build general computational platforms that simulate or mimic experimental assays of interest ( such as antibody-virus neutralization data ) can provide an efficient means to generate otherwise infeasible large-scale datasets , opening the door for new technology developments and novel biological insights . Here , we developed next-generation technology that addresses a number of key challenges , significantly advancing the utility of the neutralization-based NFP approach and enabling the investigation of important new biological questions , for analysis both at the individual donor level and for entire cohorts . These two types of analyses , however , have notably different goals . At the individual donor level , the primary goal is the successful identification of antibody neutralization signals in a given polyclonal response , typically with the goal of performing subsequent antibody isolation and characterization . Improving the accuracy and confidence in the computational predictions is thus essential for this type of analysis . To that end , we developed methods for the identification of sera with potentially novel specificities , as well as for assessing the confidence in the computational predictions for a given sample . In addition , we developed an approach for selecting viral strains for neutralization experiments that improve the ability of the algorithm to predict component antibody specificities in polyclonal responses . Virus panel size and selection was found to be significant for the accuracy of the NFP predictions—for example , the delineation for donor CAP256 ( also called 100256 ) , was much cleaner ( a single positive PG9-like signal ) when using a 50-strain panel vs . either the 20-strain f61 or the published 21-strain panel ( Figs 4E and 7B ) . However , while our results indicated that increasing the size of the virus panels , at least up to a certain level , may contribute significantly to the accuracy of the computational predictions , practical considerations such as limited sample availability or feasibility of large-scale neutralization experiments may in many cases still necessitate the use of virus panels of size 20 or less , which can still exhibit excellent levels of prediction accuracy . In addition , the large 132-strain panel was found to be less optimal for NFP analysis compared to many smaller-size panels identified in our search . This observation could potentially be due to favoring broader vs . weaker antibodies when using large panels , or to an increased unnecessary redundancy in large panels . Since the NFP method is affected , in part , by the ability to discriminate between the neutralization fingerprints of antibodies with different epitopes vs . antibodies with similar epitopes , it is not surprising that ad hoc selection of virus strains may not lead to optimal prediction accuracy . Overall , the results from the virus panel analysis performed here emphasize the importance of virus panel selection for neutralization analysis . Generally , the virus panel selection method here can be adapted to the particular problem at hand , such as selecting a subset of strains from already existing neutralization data , or performing clade-specific strain selection . In addition , as novel anti-HIV antibodies continue to be discovered and thereby improve our understanding of the major neutralizing epitopes , this method provides a framework for assessing virus panels to be used with alternate sets of antibodies as templates for serum specificities . At the cohort or population level , the primary goal is to accurately predict the overall prevalence of different types of antibody specificities in the analyzed samples . To achieve this , in addition to the technology developed for improving individual-level analysis , we introduced new algorithms specifically designed for handling collections of samples . A potential additional application of population-level analyses would be for quantifying the advantages of using the NFP method for serum delineation . For example , by using the predicted frequency of occurrence of the different bNAb specificities ( Fig 7D ) , the probability of choosing the correct signals for individual donor sera can be computed: e . g . , 0 . 016 for identifying a single positive ( VRC01-like ) signal for donor 45 and 0 . 006 for identifying two signals ( VRC01-like and PG9-like ) for CHAVI 0219 . Such quantification analyses can further underline the utility of using the NFP approach for delineating polyclonal antibody responses to HIV-1 infection . The large-scale HIV-infected cohort analysis provides initial evidence that NFP can be adapted to cohort-level analysis , although some interesting trends , such as potentially clade-specific variation of the predicted antibody responses , were also observed with the data in the current study . In particular , the high frequency of 10E8-like signals was an interesting observation ( Fig 7B and 7D ) . While NFP analysis of additional HIV infection cohorts and further experimental mapping will help clarify the overall prevalence of 10E8-like antibodies , the fact that false positive 10E8-like signals were not common in our simulations of sera consisting of known bNAbs ( S8B Fig ) and the various filters applied to remove sera with potentially inaccurate predictions , act in support of the bNAb prevalence observations for the samples shown in Fig 7B . Expanding the population-level analysis to a larger set of samples , with increased diversity that more closely reflects the actual global distribution of infections , will be informative for deciphering potential relationships between the phenotype of the neutralizing antibody response and the genetic characteristics of the infecting virus . Ultimately , such analyses can help guide the design of epitope-specific HIV-1 vaccines that are tailored to take into account the prevalence of infecting clades within a specific geographic region .
Single round of replication Env-pseudoviruses were prepared , titered and used to infect TZM-bl target cells ( obtained from George Shaw , University of Pennsylvania ) as described previously [74] . Neutralization data of monoclonals was determined using a panel of ~200 geographically and genetically diverse Env-pseudoviruses representing the major subtypes and circulating recombinant forms . The data were calculated as a reduction in luminescence units compared with control wells , and reported as half-maximum inhibitory concentration ( IC50 ) in micrograms per milliliter for monoclonal antibodies , or reciprocal dilution ( ID50 ) for serum samples . Since the antibody-virus neutralization matrix had missing data ( i . e . , no available data for a subset of the antibody-virus pairs ) , the initial set of ~200 HIV-1 strains was filtered to the largest common subset of strains with data for antibodies from all known major broadly neutralizing epitopes . That subset was further filtered to exclude strains for which data showed they were easy to neutralize by polyclonal samples or by weakly neutralizing monoclonal antibodies , as well as strains that had been noted to have yielded inconsistent data in previous experiments with serum . This filtering process resulted in a panel of 132 strains that were used in the analysis . The reference set of bNAb specificities was constructed to include antibodies from the known major sites of vulnerability . Specifically , antibodies were grouped into ten epitope-specific clusters , with each cluster represented by one or more antibodies . An antibody neutralization breadth of at least 30% was used as a cutoff for inclusion in the reference antibody dataset . Further , only one antibody from a given lineage was included in the reference set , and multiple repeats of the same antibody-virus neutralization data were averaged . The final reference set consisted of 22 antibodies with well-characterized epitopes , as follows: CD4 binding site—VRC01-like: VRC01 , VRC27 , VRC-CH31 , VRC-PG20 , VRC-PG04 , VRC23 , 12A12 , 3BNC117; b12-like: b12; HJ16-like: HJ16 . V1V2 apex—PG9-like: PG9 , PGT145 , CH01 . Glycan-V3—PGT128-like: PGT121 , PGT128 , PGT135 . Membrane-proximal external region - 2F5-like: 2F5; 10E8-like:10E8 , 4E10 . The gp120-gp41 interface - 35O22-like: 35O22; 8ANC195-like: 8ANC195; PGT151-like: PGT151 . These ten antibody groups differed compared to [4] , reflecting the discovery of new broadly neutralizing antibody epitope specificities since the original NFP publication . For each antibody , the neutralization fingerprint was defined as the IC50 values for that antibody against the set of HIV-1 strains . Neutralization data against selected viral strains was generated for the following serum samples: NIAID 45 and Z258 ( from Mark Connors , NIAID ) , C38 ( from VRC , NIAID ) , Rhesus CE8J ( from Malcolm Martin , NIAID ) , CAP256 , CAP255 , and CAP206 ( from Lynn Morris , CAPRISA ) , and CHAVI 0219 ( from Barton Haynes , Duke-CHAVI-ID ) . The prediction of antibody specificities from polyclonal neutralization was performed as described previously [4] ( S1A Fig ) . Briefly , polyclonal neutralization on a set of HIV-1 strains was represented as a combination of the neutralization fingerprints of the ten epitope-specific clusters from the antibody reference set . The ten epitope-specific clusters were formed by performing antibody clustering based on the neutralization fingerprints , as described previously [4] . A least-squares fitting procedure was then used to estimate the relative contribution to polyclonal neutralization of each of the ten epitope-specific antibody specificities . Effectively , for a given serum , this procedure results in a list of ten coefficients ( one for each antibody specificity ) , each ranging from 0 to 1 , that sum to 1; output coefficients ( referred to as delineation scores for each antibody specificity ) of more than 0 . 25 were considered as positive signals , therefore allowing at most four positive signals in any given serum . A new Octave implementation of the algorithm was used , and additional functionality and scripts were implemented in Mathematica and Java . The next-generation NFP technology uses the same polyclonal specificity prediction module as the original NFP [4] . However , the next-generation NFP incorporates a number of new algorithms compared to the original NFP: algorithms for predicting the presence of potentially novel specificities; algorithms for reporting prediction confidence scores; algorithms for population-level analysis; and algorithms for selection of improved virus panels for neutralization experiments . Simulated sera were generated as pairwise bNAb combinations that mixed equal amounts of an antibody representative from two of the ten epitope-specific antibody groups . Two sets of 4 , 500 simulated sera were generated: ( i ) one set of sera was used as a training set for the virus panel search procedure , while ( ii ) the other set was used as a test set for evaluating the performance of candidate virus panels identified during the search . The training set included pairwise combinations between antibodies from the reference set . The test set included an expanded set of 56 antibodies that consisted of clonal variants of antibodies from the reference set , as well as additional antibodies targeting epitopes that generally overlapped with , and showed similar neutralization fingerprints to , one of the epitope-specific clusters from the reference set ( S9 Fig ) . Of note , the neutralization-based antibody clustering was not found to be dependent on the antibody neutralization breadth: antibodies with substantially different neutralization breadth were successfully clustered close to the appropriate groups ( for example , CAP256-VRC26 . 03 with 38% and PG9 with 79% ) ; similar observations were made for antibody development intermediates with limited breadth [75] . Due to the limited number of bNAbs for which there exist large-scale antibody-virus neutralization data , especially for classes where only one bNAb representative has been characterized ( such as 35O22 ) , it is not possible to avoid the inclusion of antibodies in the test set of sera that are similar to or the same as the antibodies in the training set . However , because of the fingerprint transformation methods for serum simulations ( described below ) , the training and test sets of sera would be diverse even if using the same set of bNAbs . In addition , the expansion of the antibody set for the test sera allows for further improvement in serum diversity . Analysis of the serum delineation error for the top 5 , 000 20-strain panels against test sera that either include bNAb clonal relatives ( the 56-antibody set described above ) or include only a single representative from each clonal group showed a significant correlation ( p<0 . 0001 , Spearman; S10 Fig ) . Overall , while the test set of sera are not optimal due to the use of antibodies that overlap the training set , these test sera provide a useful quantitative measure for comparison of virus panel performance . For each pairwise combination , an antibody was selected from two of the ten epitope-specific clusters , and the neutralization fingerprint for the resulting combination was obtained as follows . First , the neutralization fingerprint for each of the two antibodies was transformed using a simulation of variability for the neutralization experiments . Specifically , a set of actual experimental variability values observed for multiple repeats of the same neutralization experiments ( for example , for multiple repeats of the neutralization of strain YU2 by antibody VRC01 ) was compiled; then for each n ( a , v ) ( the measured neutralization IC50 value for antibody a against viral strain v ) , a scaling factor k1 was obtained from a distribution formed by the set of experimental variability values ( using the EmpiricalDistribution function of Mathematica ) . Next , the potency of the resulting adjusted fingerprints for each of the two antibodies were scaled by a random factor k2 between 1/10 and 10 , where k2 was constant for all n ( a , v ) for antibody a in the selected antibody pair . These two transformations ( to account for experimental variability and to incorporate a notion of potency variation between antibodies with similar epitopes ) aim at increasing the diversity of the simulated fingerprints . Finally , the two transformed fingerprints were combined by taking the minimum of the two neutralization values for each of the HIV-1 strains , in order to obtain the neutralization fingerprint for the simulated serum . Specifically , given two selected antibodies a1 and a2 and a specific viral strain v , the neutralization value for the simulated serum for the combination of a1 and a2 is computed as: min [t ( n ( a1 , v ) ) , t ( n ( a2 , v ) ) ] , where t ( x ) = k1 k2 x is the transformation of the original antibody-virus neutralization data x , as described above . Taking the minimum of the two neutralization values for an antibody pair within a combination is a modeling simplification of monoclonal antibody interactions in polyclonal sera . In reality , antibodies in serum would have complex interactions that could be antagonistic ( such as for antibodies that bind to different , incompatible , Env conformations ) , synergistic , or in some cases–independent . Ideally , a serum simulation model would be able to accurately account for such interactions , but current knowledge about the determinants of these interactions is limited , and such model improvements will be the aim of future work on simulating polyclonal serum neutralization . Nevertheless , despite the modeling simplifications , the excellent agreement between simulated and actual neutralization fingerprints ( Figs 1B and S2B ) suggests that the current model is appropriate for modeling serum simulations . Interestingly , for two antibody combinations in which the two antibodies in each combination had similar or overlapping epitopes ( VRC01+HJ16 and 2F5+10E8 , Fig 4A ) , the experimentally determined and the computationally predicted neutralization values were in excellent agreement ( Spearman correlations of 0 . 96 and 0 . 98 , respectively ) , further underlining the feasibility of the proposed serum simulation procedure . The generated simulated sera represented diverse levels of neutralization breadth ( S2C and S3B Figs ) . Sera with potentially novel ( or unknown ) specificities were computed by identifying random neutralization fingerprints that , for a given virus panel , had low ( <0 . 2 ) Spearman correlation with any of the antibody neutralization fingerprints from the reference set; a total of 10 , 000 such simulated sera were used in the analysis . While it is possible to specifically optimize virus panels for detection of novel specificities , we used virus panel f61 for this analysis because of its improved performance with the known bNAb specificities , which would likely be commonly found in serum analysis . For a given virus panel , reference antibody set , and simulated sera , the average serum delineation error was computed as follows . For each simulated serum , the NFP algorithm was applied to generate predictions about the neutralization prevalence of each of the antibody specificities from the reference set: we refer to these predictions as the NFP delineation scores . The serum delineation error for a given simulated serum was then computed as the RMSD between the NFP delineation scores and the actual antibody composition in that serum . The average serum delineation error was computed as the average of the RMSD’s for the entire set of sera . The serum delineation error was used as a measure to determine the prediction accuracy of a given virus panel . Several methods were used for searching for optimized virus panels of size 20 . Random search . Virus panels were generated by randomly selecting subsets of size 20 from the 132 strains . Sequence diversity . HIV-1 Env sequence analysis for the 132 strains was performed either for all residues between 1–683 or only for a subset of the residues that were found to be part of antibody epitopes from antibody-antigen complex structures . In each case , the selected Env sub-sequences were grouped into 20 clusters based on sequence similarity [4] . The A* search algorithm [65] was then applied by constraining enumeration of candidate panels to include a single strain from each of the 20 clusters . Given these constraints , the A* algorithm generated a set of virus panels with optimized sequence diversity . Monte Carlo-based optimization search . Multiple starting points were each initialized to a random selection of 20 strains from the full set of 132 strains , and an optimization search was performed for each starting point . For each starting point , each step of the search procedure consisted of randomly replacing one of the strains in the current 20-strain panel , and retaining the new panel if a success criterion was achieved , or otherwise discarding the replacement and keeping the previous panel . The success criterion was determined by comparing the prediction accuracy ( serum delineation error ) over the set of training sera for the new vs . previous panel: the new panel was retained either if its prediction accuracy was better , or with a probability dependent on the difference in prediction accuracy between the two panels . For each search method , a selected set of top virus panels was evaluated on the independent set of test sera that was not used during the search procedure . When selecting a specific virus panel for neutralization experiments , additional criteria could be utilized , such as ability to use neutralization fingerprints to successfully cluster antibodies consistent with epitope similarity ( S3A Fig ) , clade distribution of selected strains , rate of false positive predictions , minimum antibody neutralization breadth over the given panel , etc . In particular , an antibody clustering procedure analogous to the procedure described in [4] can be used for filtering out candidate virus panels that do not correctly cluster the neutralization fingerprints of antibodies according to epitope similarity . Specifically , for a given virus panel , neutralization-based antibody clustering is performed using the Agglomerate function in Mathematica with default parameters ( squared Euclidean distance and single linkage ) . The resulting hierarchical clustering is then compared to the ten pre-defined epitope-specific antibody groups by using the ClusterSplit function in Mathematica , and virus panels that misclassify antibodies ( e . g . , S3A Fig ) are discarded from further consideration . In the case of the 20-strain virus panel f61 , that panel was selected based on a combination of criteria , including a good ( though not the best ) serum delineation accuracy among the panels identified in the 20-strain panel optimization search , ability to correctly cluster the neutralization fingerprints of reference bNAbs according to epitope specificity ( S3A Fig ) , high minimum bNAb neutralization breadth ( to allow for a non-negligible serum neutralization signal from each respective antibody group ) , and sequence diversity ( for example , presence of at least two strains each of clades A , B , and C ) , among others . For a given pair of monoclonal antibodies a1 and a2 , the strain-potency mismatch value was defined as the absolute value of the difference between the number of strains for which a1 is more potent and the number of strains for which a2 is more potent . Pearson correlation and p-value were computed for strain-potency dominance vs . NFP serum delineation error . For a given simulated serum , the residuals from the fit of the serum neutralization fingerprint and the combination of reference-antibody fingerprints resulting from the NFP delineation were computed . For a given virus panel , the residual score for a given serum was obtained as a sum of two t-scores based on the computed residuals , divided by a factor of 1000: ( a ) a t-score computed over the set of sera with known specificities , and ( b ) a t-score computed over the set of sera with potentially novel ( or unknown ) specificities . In addition to residual scores , two other scores were defined as means to estimate the confidence in the computational predictions for a given serum . Median of delineation scores: computed as the median of the NFP delineation scores for the ten specificities from the reference antibody set . Frequency of random signals: For a given serum , the frequency of random neutralization signals was computed as follows . The reference set of ten epitope-specific antibody specificities was expanded to include an eleventh specificity . For each serum , NFP delineation was performed by incorporating , separately , each of the ( 10 , 000 ) random neutralization fingerprints for simulated sera with unknown specificities as the eleventh reference specificity . Then , for each serum , the frequency of random signals was computed as the fraction of cases in which the eleventh ( random ) specificity had a positive signal , within the 10 , 000 NFP runs for the given serum . Sets of sera , either with unknown antibody specificities or with 1 or 2 dominant known antibody specificities ( with the assumption that sera with more than 2 dominant broadly neutralizing specificities would not be common [12 , 14 , 18 , 66] ) , were generated using the algorithms described above . From these , multiple simulations , with different fractions of sera with unknown specificities vs . sera with 1 or 2 known antibody specificities , were performed . In each simulation , a sample size of 200 sera was used . Sera with unknown specificities were allowed to represent 0 to 50% , at increments of 10% , of the 200 sera , while the remaining set of sera was distributed between sera with 1 vs . 2 specificities at a ratio ranging from 0:100% to 100:0% , at steps of 10% . For a fixed distribution of serum specificities , 1 , 000 samples of 200 sera each were generated . NFP analysis was performed for each of the 1 , 000 samples , and the prediction accuracy for the prevalence of each of the ten antibody specificities was computed as follows . For each of the 1 , 000 samples , the ratio of the NFP-predicted prevalence of a given antibody specificity and the actual prevalence in the respective set of simulated sera ( after filtering of sera based on confidence scores , if using the next-generation NFP ) was reported as the fold difference between actual and predicted prevalence for a given antibody specificity . The fold differences were presented either as detailed box plots ( Fig 6A ) or their medians were normalized on a 0–1 scale ( setting fold differences >1 to 1/fold difference ) and averaged over all ten antibody specificities ( Fig 6B ) . For the next-generation NFP algorithms , an adjustment procedure was developed that incorporated the expected prediction accuracy for each of the ten antibody specificities for a given virus panel . The expected prediction accuracy ei for a given antibody specificity i was determined by the true/false positive rate as determined on the test set of pairwise simulated sera ( e . g . , S8B Fig ) , which was independent from the simulated sera described in this section . Specifically , ei was defined as the ratio ti/ ( ni+ti ) , where ti is the number of sera with ( a ) two predicted positive signals and ( b ) a positive signal for specificity i; and ni is the number of false negative sera for specificity i . In effect , ei allows us to estimate the false negative rate from the observed frequency of positive signals for specificity i in sera that are predicted to have two positive signals . To apply the adjustment procedure in prospective cohort-level analysis , the frequency fi of a given antibody specificity i , for sera with two positive signals , was computed as xi/ei , where xi is the number of sera with positive signals for specificity i . This frequency was then combined with the ( unadjusted ) frequencies for the remaining sera ( for which the number of positive signals was different from two ) , in order to obtain the overall NFP-predicted frequency for a given antibody specificity . HIV-1 neutralization for a set of 205 donor plasma was obtained from published data , which also included information about the clade of infection [63 , 76] . The 132 strains from the monoclonal antibody dataset described above were compared to the strains from the plasma dataset , and a common panel of 69 strains was selected ( S8A Fig ) . Plasma samples with missing data for the 69 common strains were discarded from the analysis , resulting in a set of 189 samples . A virus panel search for optimized NFP accuracy was performed , and a panel of size 50 with reduced false positive rates was selected for the plasma analysis ( S8A and S8B Fig ) . We specifically selected a panel with reduced false positives , in order to facilitate the estimation of specificities at a population level given the individual predictions ( see section above ) . Samples with less than 30% neutralization breadth were discarded from further analysis ( since the information content for sera with low neutralization breadth does not permit accurate delineation of the polyclonal signal ) , resulting in a final set of 143 plasma for analysis by the next-generation NFP algorithms . A residual score cutoff of greater than 0 . 085 ( greater than the max for all simulated sera with known dominant specificities , for the selected 50-strain panel ) was used for predicting plasma with potentially novel specificities . Cutoffs of less than or equal to -0 . 1 for residual scores , 0 . 06 for the frequency of random signals , and 0 . 065 for the median of the delineation scores were used for dividing the remaining plasma samples into samples predicted to have dominant known specificities and samples with inconclusive delineation . GraphPad Prism , Octave , and Mathematica scripts were used for performing statistical analysis . Data deposited in the Dryad repository: http://dx . doi . org/10 . 5061/dryad . c73q1 [77]
|
HIV-1 remains a significant global health threat , with no effective vaccine against the virus currently available . Since traditional vaccine design efforts have had limited success , much effort in recent years has focused on gaining a better understanding of the ways select individuals are able to effectively neutralize the virus upon natural infection , and to utilize that knowledge for the design of optimized vaccine candidates . Primary emphasis has been placed on characterizing the antibody arm of the immune system , and specifically on antibodies capable of neutralizing the majority of circulating HIV-1 strains . Various experimental techniques can be applied to map the epitope targets of these antibodies , but more recently , the development of computational methods has provided an efficient and accurate alternative for understanding the complex antibody responses to HIV-1 in a given individual . Here , we present the next generation of this computational technology , and show that these new methods have significantly improved accuracy and confidence , and that they enable the interrogation of biologically important questions that can lead to new insights for the design of an effective vaccine against HIV-1 .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"immune",
"physiology",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"applied",
"mathematics",
"immunology",
"microbiology",
"retroviruses",
"viruses",
"immunodeficiency",
"viruses",
"algorithms",
"simulation",
"and",
"modeling",
"optimization",
"mathematics",
"rna",
"viruses",
"antibodies",
"antibody",
"response",
"research",
"and",
"analysis",
"methods",
"monoclonal",
"antibodies",
"immune",
"system",
"proteins",
"antibody",
"specificity",
"proteins",
"medical",
"microbiology",
"hiv",
"microbial",
"pathogens",
"hiv-1",
"immune",
"response",
"biochemistry",
"physiology",
"viral",
"pathogens",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"lentivirus",
"organisms"
] |
2017
|
Mapping Polyclonal HIV-1 Antibody Responses via Next-Generation Neutralization Fingerprinting
|
In Caenorhabditis elegans and Drosophila melanogaster , the aging of the soma is influenced by the germline . When germline-stem cells are removed , aging slows and lifespan is increased . The mechanism by which somatic tissues respond to loss of the germline is not well-understood . Surprisingly , we have found that a predicted transcription elongation factor , TCER-1 , plays a key role in this process . TCER-1 is required for loss of the germ cells to increase C . elegans' lifespan , and it acts as a regulatory switch in the pathway . When the germ cells are removed , the levels of TCER-1 rise in somatic tissues . This increase is sufficient to trigger key downstream events , as overexpression of tcer-1 extends the lifespan of normal animals that have an intact reproductive system . Our findings suggest that TCER-1 extends lifespan by promoting the expression of a set of genes regulated by the conserved , life-extending transcription factor DAF-16/FOXO . Interestingly , TCER-1 is not required for DAF-16/FOXO to extend lifespan in animals with reduced insulin/IGF-1 signaling . Thus , TCER-1 specifically links the activity of a broadly deployed transcription factor , DAF-16/FOXO , to longevity signals from reproductive tissues .
When the germline of C . elegans is removed , either by laser ablation or by mutations that block germline-stem cell proliferation , the animals live approximately 60% longer than normal [1] , [2] . This longevity is not a simple consequence of sterility , as removing the entire gonad ( the germ cells as well as the somatic reproductive tissues ) does not extend lifespan [1] . These and other findings suggest that both the germline cells and the somatic reproductive tissues influence lifespan [3] , [4] . A link between reproductive tissues and aging may be widespread in nature , as removing germline-stem cells during adulthood increases the lifespan of flies [5] , as well as worms , and transplanting ovaries of young mice into older animals can increase lifespan as well [6] . The mechanism by which loss of the germline extends lifespan is not well understood . At least two transcription factors , the FOXO-family transcription factor DAF-16/FOXO [7]–[9] and the nuclear hormone receptor DAF-12 [10] , are required for germ-cell loss to extend lifespan in C . elegans [1] . When germ cells are removed , DAF-16/FOXO accumulates in nuclei , primarily in the animal's intestine [11] . The intestine , which functions as the worm's entire endoderm , including its fat tissue , appears to play a key role in this pathway , as expressing DAF-16/FOXO only in the intestine completely restores the lifespan extension of germline-defective daf-16 ( − ) animals [12] . DAF-16/FOXO nuclear localization requires an ankyrin-repeat , intestinal protein called KRI-1 [13] . In addition , full DAF-16/FOXO nuclear localization requires DAF-12/NHR [10] , [13] , as well as DAF-9 , a cytochrome P450 protein that synthesizes a DAF-12/NHR ligand [13]–[15] . DAF-16/FOXO is particularly interesting because of its evolutionarily-conserved role in another lifespan regulatory pathway , the insulin/IGF-1 endocrine pathway [16]–[18] . In wild-type worms , the insulin/IGF-1 receptor DAF-2 activates a series of conserved kinases that ultimately phosphorylate and inactivate DAF-16/FOXO [19]–[21] . When insulin/IGF-1 signaling is inhibited , DAF-16/FOXO accumulates in nuclei , where it regulates the transcription of downstream antioxidant , chaperone , innate-immunity , and metabolic genes that more directly affect lifespan [8] , [9] , [11] , [22]–[25] . Insulin/IGF-1 signaling and FOXO proteins influence lifespan in worms , flies and mice [16] , [26] , [27] . The pathway appears to affect human longevity as well , as variants in the human Foxo3a gene have been linked to longevity in several human populations [28]–[32] , and functionally significant IGF-1 receptor mutations are overrepresented in populations of centenarians [33] . In an effort to better understand how reproductive cues trigger lifespan extension , we carried out a genetic screen for RNAi clones that prevent germline loss from extending lifespan . Surprisingly , in our screen we identified a gene , tcer-1 , that encodes the homolog of the human transcription elongation factor TCERG1 . Loss of tcer-1 in C . elegans sharply curtails the lifespan extension produced by removal of the germline , but has little or no effect on wild-type lifespan . Thus , TCER-1 is not likely to be a general component of the transcriptional machinery; instead , it has a much more specific function in the animal . We find that TCER-1 plays a key regulatory role in transducing signals from the reproductive system to the somatic tissues . When the germline is removed , TCER-1 levels rise in the intestine . This up-regulation is functionally significant because overexpressing tcer-1 in normal , fertile animals bypasses the requirement for germ-cell loss and extends lifespan . Our findings indicate that the role of tcer-1 in the reproductive longevity pathway is to promote the transcription of a subset of DAF-16/FOXO-target genes that are up-regulated upon germline removal . Thus , TCER-1 appears to act in association with DAF-16/FOXO to extend lifespan . Interestingly , tcer-1 is not invariably required for DAF-16/FOXO transcriptional activity , as it is not needed for DAF-16/FOXO to up-regulate its target genes or to extend lifespan in daf-2 insulin/IGF-1 receptor mutants . At least some of the genes that are up-regulated by DAF-16/FOXO in a tcer-1-dependent fashion in long-lived , germline-defective animals do not require tcer-1 for their expression in long-lived insulin/IGF-1 mutants . Thus , this transcription elongation factor appears to link germline loss to a precise DAF-16/FOXO-dependent transcriptional program .
We identified tcer-1 in a Chromosome II RNAi screen for genes required to extend the lifespan of germline-depleted glp-1 mutants ( see Materials and Methods ) . When grown at the non-permissive temperature ( 25°C ) , temperature-sensitive glp-1 mutants are sterile because the germline-stem cells fail to proliferate [34] . These mutants recapitulate the lifespan extension of animals whose germline-precursor cells have been eliminated by laser ablation [2] . Since lifespan can be shortened by conditions that simply compromise the animals' health , we focused on RNAi clones that suppressed the extended lifespan of glp-1 mutants but had little or no effect on the lifespan of wild-type animals . From this screen , we identified a clone of the gene tcer-1 ( ZK1127 . 9 ) , which encodes the C . elegans homolog of a human transcription elongation factor , TCERG1 ( also known as CA150 ) [35]–[37] . tcer-1 RNAi strongly suppressed the extended lifespan of glp-1 mutants , but had no effect on the lifespan of wild-type animals ( Figure 1A and 1B; Table S1A ) . Wild-type animals grown on bacteria expressing tcer-1 dsRNA appeared normal and healthy , had regular developmental rates and displayed normal reproduction . We obtained a tcer-1 mutant , tcer-1 ( tm1452 ) , from the C . elegans National Bioresource Project , Japan . The tm1452 mutation is a 10 bp insertion coupled to a 392 bp deletion that eliminates parts of the WW domains of the protein and is likely to reduce gene function . We constructed a tcer-1; glp-1 double mutant and found that it had a much shorter lifespan than did glp-1 mutants; for example , in one experiment the 74% extension in lifespan produced by the glp-1 mutation was reduced by tcer-1 ( tm1452 ) to 15% ( Figure 1C , Table S1B ) . There was no effect on the lifespan of wild-type worms ( Figure 1C , Table S1C ) . This finding is significant , as it indicates that tcer-1 is unlikely to play a general role in transcription elongation , but instead has a more specific function in the animal . tcer-1 mutants displayed a delay in developmental timing , and a modest reduction in brood size that was not elicited by tcer-1 RNAi treatment ( data not shown ) . The observation that reducing tcer-1 activity had only mild effects on wild type but almost completely prevented germline loss from extending lifespan suggested that tcer-1 might play a regulatory role in this pathway . If so , it seemed possible that its level or location in the animal might change upon germline removal . To investigate this possibility , we constructed transgenic animals expressing a TCER-1::GFP fusion protein under the control of the endogenous tcer-1 promoter ( see Materials and Methods ) . In the wild type , TCER-1::GFP was visible at all stages of embryonic and larval development ( data not shown ) . In adults , we observed strong nuclear localization of TCER-1::GFP in intestinal cells , many head and body neurons , muscle and hypodermal cells ( Figure 2A–2C ) . In some intestinal cells , weak expression was also observed in the cytoplasm . To test if this fusion protein was functional , we examined the ability of the construct to rescue the shortened lifespan of tcer-1; glp-1 double mutants . Two independent lines expressing the TCER-1::GFP fusion protein completely rescued the lifespan of tcer-1; glp-1 mutants in one trial ( Figure 2D , Table S1C ) . In a second trial , the rescue was between 77–88% ( Table S1C ) . Thus , TCER-1::GFP is a functional protein that likely reflects the endogenous expression of tcer-1 . Importantly , these observations showed that a functional TCER-1 protein was present in intestinal nuclei in the adult , where DAF-16/FOXO has been shown to act to increase lifespan [11] , [12] . Next , we examined the effect of germline ablation on the pattern of TCER-1::GFP expression . We found that eliminating the germline increased the level of TCER-1 in somatic cells ( Figure 3A–3C ) . This finding raised the possibility that TCER-1 might have an important regulatory role in the animal's response to germline ablation . An independent microarray analysis designed to identify longevity genes whose expression is altered by germline ablation also found tcer-1 amongst the top 10% of germline-regulated genes ( M . McCormick and C . Kenyon , unpublished data ) . Thus , tcer-1 is likely to be transcriptionally up-regulated following germline depletion . TCER-1 levels increased primarily in two tissues , the intestine and neurons , upon germline removal ( Figure 3A and 3B ) . tcer-1 RNAi suppressed longevity without affecting neuronal expression ( Figure S1 ) . Because neurons are relatively resistant to RNAi in C . elegans , this finding suggests that intestinal TCER-1 function , like intestinal DAF-16/FOXO , is likely to be particularly important for promoting longevity . The finding that loss of the germline elevated TCER-1 levels raised the possibility that TCER-1 might play a rate-limiting step in this longevity pathway . To test this , we asked whether overexpressing tcer-1 in wild-type animals might bypass the requirement for germline loss and extend the lifespan of fertile animals . We found that tcer-1 overexpression produced a modest but consistent increase in the lifespan of wild-type worms ( average mean lifespan extension ∼15%; Figure 3D and 3E; Table S2A and S2B ) . We also found that , in keeping with its role in the reproductive pathway , tcer-1 overexpression did not produce any further increase in the extended lifespans of worms whose germ cells had been eliminated either by the glp-1 mutation ( Figure 3F; Table S2C ) or by laser ablation ( Figure S2; Table S2C ) . To understand why tcer-1 was required for loss of the germ cells to extend lifespan , we asked whether tcer-1 might impact DAF-16/FOXO function . DAF-16/FOXO undergoes nuclear accumulation primarily in intestinal cells following germline removal and we first explored the possibility that tcer-1 was required for this step . However , we found that in tcer-1 ( tm1452 ) worms lacking a germline , DAF-16/FOXO accumulated normally in intestinal nuclei ( Figure S3 ) . Curiously , we noticed a small increase in the levels of DAF-16::GFP , largely in the cytoplasm of intestinal cells , in these animals ( Figure S3; see Discussion ) . Next , we examined the effect of tcer-1 ( tm1452 ) on DAF-16/FOXO's transcriptional output following germline loss . Many putative ( direct or indirect ) DAF-16/FOXO target genes have been identified [12] , [22]–[25] , but only a few , sod-3 , dod-8 , gpd-2 and nnt-1 , are known to be up-regulated by germline removal [3] . To increase the repertoire of such genes , we obtained available transgenic worm strains expressing GFP-tagged transcriptional reporters for additional putative DAF-16/FOXO targets [38] ( see Table S3 ) . Germline precursors of each of these strains were laser ablated , and the animals were examined as adults for changes in expression . We found several genes whose expression was sharply elevated upon germline removal , including K07B1 . 4 , T21D12 . 9 , F52H3 . 5 , aat-1 and pssy-1 ( see Table S3 ) . We then asked whether the up-regulation of these new genes , as well as the known daf-16-dependent germline-regulated genes , required tcer-1 . We found that this was the case for six of the nine genes we examined . The up-regulation of gpd-2 , dod-8 , nnt-1 , K07B1 . 4 , and T21D12 . 9 was strongly attenuated in tcer-1; glp-1 mutants ( Figure 4A–4T ) and pssy-1 expression was moderately reduced ( data not shown ) . In contrast , aat-1 expression remained unaltered in two trials and showed a small increase in expression in one ( data not shown ) . These findings suggest that loss of TCER-1 may eliminate at least a portion of DAF-16/FOXO's transcriptional response to germline depletion . Unexpectedly , sod-3 and F52H3 . 5 levels were further increased upon tcer-1 inactivation ( Figure 4U–4X and Figure S4 , respectively; see Discussion ) . As tcer-1 reduction of function reduced the expression of many DAF-16/FOXO-target genes , we asked if TCER-1 overexpression would increase the expression of the same DAF-16/FOXO-target genes that were downregulated in tcer-1 ( − ) ; glp-1 ( − ) mutants . Using quantitative RT-PCR ( Q-PCR ) , we found that several such genes ( dod-8 , gpd-2 , nnt-1 , K07B1 . 4 and pssy-1 ) showed a modest but statistically significant increase in expression when tcer-1 was overexpressed ( Figure 5A ) . In addition , the lifespan extension obtained by tcer-1 overexpression required the presence of wild-type daf-16 ( Figure 5B; Table S4 ) . Together these findings suggest that tcer-1 overexpression may extend the lifespan of wild-type , fertile animals by stimulating expression of DAF16/FOXO-target genes that are induced in response to germline loss . Since daf-16 is also required for the doubling of lifespan produced by reducing insulin/IGF-1 signaling [16] , [39] , we examined the role of tcer-1 in this longevity pathway too . We found that neither tcer-1 ( tm1452 ) nor tcer-1 RNAi suppressed the extended lifespans of the insulin/IGF-1-receptor mutants daf-2 ( e1370 ) or daf-2 ( e1368 ) ( Figure 6A and 6B; Table S5 ) . Therefore , TCER-1 is not part of the insulin/IGF-1 pathway . Consistent with this , we found that TCER-1 was not required in daf-2 ( e1370 ) mutants for the expression of any of the DAF-16/FOXO targets that were tcer-1-dependent in the reproductive pathway . The tcer-1 ( tm1452 ) mutation either had no effect , or in some cases produced a small increase in expression ( Figure 6C and 6D ) . Removing the germ-cell precursors of daf-2 mutants further doubles their already long lifespan [1] , [4] . Similarly , we found that overexpressing tcer-1 in daf-2 ( e1370 ) mutants extended their long lifespan by another 6–12% ( Figure 6E; Table S2C ) . In contrast , we did not observe a further extension of lifespan when tcer-1 was overexpressed in germline-depleted worms ( Figure 3F , Figure S2; Table S2C ) . These observations support the model that tcer-1 functions in the germline longevity pathway but not the insulin/IGF-1 longevity pathway . Not all longevity pathways in C . elegans are daf-16 dependent . For example , the longevity response to caloric restriction caused by the feeding-defective mutant eat-2 ( ad1116 ) [40] is daf-16 independent , as is the longevity response to inhibition of mitochondrial respiration caused by isp-1 ( qm150 ) [41]–[43] . We found that tcer-1 was not required for either of these mutations to extend lifespan ( Figure 6F and 6G; Table S5 ) . These findings further reinforced the interpretation that TCER-1 is not a general longevity factor but instead extends lifespan specifically in response to reproductive signals . The ability of TCER-1 overexpression to activate DAF-16/FOXO-dependent gene expression and extend lifespan suggested that the up-regulation of TCER-1 is a key switch-point by which loss of the germ cells triggers a longevity response . This up-regulation could also be part of the mechanism by which TCER-1 activity is specifically directed towards this pathway and not towards the insulin/IGF-1 pathway . To test this possibility , we examined TCER-1::GFP levels in daf-2 ( e1370 ) mutants . We found that daf-2 mutants did not have elevated TCER-1::GFP levels . In contrast , TCER-1 levels were significantly reduced ( Figure 6H , compare with Figure 3C ) . TCER-1 levels were also reduced in long-lived eat-2 ( ad1116 ) mutants ( Figure S5 ) , and we found that tcer-1 overexpression further extended the lifespan of eat-2 ( ad1116 ) mutants ( Figure S5; Table S2C ) . These data suggested that the specificity of TCER-1 for the reproductive pathway might be achieved , at least in part , by the regulation of its level in the animal . How is TCER-1 up-regulation controlled by the reproductive system ? To begin to address this question , we asked whether daf-16 itself , or genes that are known to influence DAF-16/FOXO nuclear localization in response to germline loss [13] , were involved . We found that the daf-16 ( mu86 ) null mutation did not alter TCER-1::GFP levels in wild-type or in germline-deficient animals ( Figure 7A ) . Likewise , neither daf-9 nor daf-12 mutations , both of which partially inhibit DAF-16/FOXO nuclear localization in germline-defective animals , affected TCER-1 up-regulation ( data not shown ) . However , in kri-1 ( ok1251 ) mutants , nuclear TCER-1::GFP levels were severely diminished in the intestine and were not increased following germ-cell ablation ( Figure 7B–7F ) . kri-1 ( ok1251 ) did not alter TCER-1::GFP levels in any other tissues ( Figure 7E and 7F ) , consistent with KRI-1's presence only in the gastrointestinal tract [13] . Thus , DAF-16/FOXO nuclear localization and TCER-1 up-regulation share a requirement for KRI-1 in their response to germ-cell loss . As expected , kri-1 ( ok1251 ) also abolished the lifespan extension evoked by tcer-1 overexpression ( Figure 7G; Table S4 ) . KRI-1 is not required for DAF-16/FOXO nuclear localization or lifespan extension in daf-2 mutants [13] . Thus , KRI-1's involvement in TCER-1 regulation explains , at least in part , why TCER-1 function is specific for the reproductive system .
We can suggest hypotheses about the molecular function of TCER-1 from studies of its human homolog , the transcription elongation factor TCERG1 ( also known as CA150 ) [35] , [36] . TCERG1 is known to associate with elongation-competent RNA Polymerase II ( RNAPII ) complexes and alter the elongation efficiency of nascent mRNA transcripts [36] , [49] , [50] . It has also been implicated in splicing [51]–[53] . TCERG1 ( like TCER-1 ) has a modular structure , with FF repeats that associate with the C Terminal Domain ( CTD ) of RNAPII , and WW domains that can be used for association with other regulatory proteins [37] , [50] , [51] , [53] , consistent with its role in regulating gene expression . The known function of TCERG1 suggests the model that TCER-1 influences longevity by promoting transcript elongation of DAF-16/FOXO-target genes . Specifically , we propose that following germline ablation , DAF-16/FOXO accumulates in nuclei and initiates transcription of a set of downstream targets , many of which require TCER-1 for transcript elongation . In germline-ablated tcer-1 mutants , the transcription of such genes is stalled , so lifespan extension is prevented . This hypothesis is supported by recent discoveries of transcription elongation factors that selectively influence transcription initiated by particular transcription initiation ( activation ) factors [54] , [55] . FOXO proteins are known to be regulated by many types of covalent modifications , including phosphorylation on multiple sites , deacetlyation , and in addition by regulated proteolysis [56] , [57] . To our knowledge , this is the first suggestion of DAF-16/FOXO regulation exerted at the level of transcription elongation . However , it is important to note that our experiments , which utilized GFP fusions and RT-PCR ( that report on the efficiency of transcription in general and not the elongation step in particular ) , do not rule out the possibility that tcer-1 acts at a different stage of transcription , such as transcription initiation . Unexpectedly , we noticed that the expression of some DAF-16/FOXO target genes , such as sod-3 and F52H3 . 5 , was further increased in tcer-1 ( − ) ; glp-1 ( − ) mutants ( Figure 4U–4X and Figure S4 ) . We also found that the cytoplasmic levels of DAF-16::GFP were elevated in these animals ( Figure S3 ) . One explanation for these observations is that in germline-depleted tcer-1 mutants , the silence of many DAF-16/FOXO targets triggers a compensatory response that elevates DAF-16/FOXO levels in the cytoplasm ( and possibly to some extent in the nucleus as well ) . For tcer-1-dependent genes ( such as nnt-1 ) , this elevated DAF-16/FOXO cannot compensate for the absence of TCER-1 , and their transcription remains stalled . However , for tcer-1-independent genes ( such as sod-3 ) , elevated DAF-16/FOXO levels increase transcription . Since DAF-16/FOXO increases lifespan in response to germline loss and reduced insulin/IGF-1 signaling , a simple model would be that it activates the same downstream lifespan-extending genes in both situations . However , we found that at least some of the genes that are up-regulated in germline-defective mutants in a tcer-1 dependent manner are not up-regulated in daf-2 mutants ( for example T21D12 . 9; see Table S3 ) . This finding suggests that the differential deployment of tcer-1 activity allows reproductive cues and reduced insulin/IGF-1 levels to trigger different patterns of DAF-16/FOXO-dependent gene expression . Thus , this situation illustrates how the activities of broadly-deployed transcription factors can be tailored to precise transcriptional outputs in response to specific signals . Interestingly , some genes that are up-regulated in germline-defective mutants in a tcer-1-dependent way are also up-regulated in daf-2 mutants , but independently of tcer-1 ( for example , dod-8 ) . Perhaps a different transcription-elongation factor , one that can substitute for TCER-1 , is activated by reduction of insulin/IGF-1 activity . A major question in the field of aging is whether a mode of lifespan extension that is normally triggered by one specific condition , such as caloric restriction , can be triggered instead by intervention at a downstream regulatory step of the pathway . As described above , we found that the longevity pathway activated by loss of the germ cells could potentially operate independently of signals related to reproduction . The finding that germline removal triggers the transcriptional up-regulation of tcer-1 suggested that this regulatory step might function as a germline-pathway specific switch that activates downstream longevity processes . In fact this appears to be the case: when tcer-1 is overexpressed in intact animals that have germ cells , lifespan is extended . This lifespan extension has no detrimental effects on progeny production ( data not shown ) and is accompanied by several hallmarks of the germline-longevity pathway , including a requirement for KRI-1 and DAF-16/FOXO , and the stimulation of a pattern of gene expression similar to that produced by germline removal . The lifespan increment obtained by overexpressing tcer-1 was not as large as that produced by germ-cell loss . Nevertheless , these data , and other recent observations [58] , suggest that at least some of the beneficial longevity effects produced by loss of the germline can be replicated without perturbing reproduction itself . The last few decades have transformed our view of aging from a haphazard , unregulated phenomenon to a process influenced by conserved genetic pathways . The insulin/IGF-1 pathway and DAF-16/FOXO orthologs have been implicated in lifespan regulation in worms , flies , mice , dogs and several human populations [16] , [26]–[33] , [59] , [60] . Studies in flies and mice indicate that pathways linking reproduction and aging maybe widespread in nature too . Given this context , the identification of a regulator , TCER-1 , that selectively responds to reproductive signals to influence the transcriptional response of a conserved longevity-determinant , DAF-16/FOXO , is exciting . It will be interesting to learn whether TCER-1's function in aging has also been conserved during evolution .
Strains were maintained as described earlier [61] . The tcer-1 ( tm1452 ) mutant was provided by the National Bioresource Project ( Japan ) and outcrossed three times to the Kenyon Lab N2 stock . Transgenic strains expressing GFP under control of promoters of DAF-16/FOXO-target genes were obtained from the CGC , or from Prof . David Baillie's laboratory ( Caenorhabditis Gene Expression Consortium ) . The details of these transgenic animals , strains generated using them and other strains used in this study are described in Table S6 and Table S3 . The Ptcer-1::tcer-1::gfp fusion construct was generated as described previously [62] . The complete coding sequence of tcer-1 ( 4 . 1 kb ) and 1 . 6 kb 5′ upstream promoter sequence were amplified using the following primers: Forward- 5′ GCA AGT ATT TGA GCA CTA CTG TCA AGG GC 3′ , Reverse- 5′ AGT CGA CCT GCA GGC ATG CAA GCT TTG CTT TCT GCG ATC CCG CTC 3′ . The gfp construct ( 1 . 9 kb ) was amplified from the vector pPD95 . 75 using the following primers: Forward- 5′ AGC TTG CAT GCC TGC AGG TCG ACT 3′ , Reverse- 5′ AAG GGC CCG TAC GGC CGA CTA GTA GG 3′ . The complete gfp fusion construct ( 7 . 6 kb ) was amplified by pooling the products of the above two PCRs and using the following primers: Forward: 5′ GCC GGT CAT GCT CTT CTT CAA C 3′ , Reverse- 5′ GGA AAC AGT TAT GTT TGG TAT ATT GGG AAT GTA TTC TG 3′ . The co-injection marker Podr-1::rfp was amplified from the plasmid pCF155 . To generate the TCER-1::GFP expressing worms , Ptcer-1::tcer-1::gfp was injected as described earlier [63] at 50 ng/µl or 20 ng/µl , along with the co-injection marker Podr-1::rfp ( 75 ng/µl ) , into N2 worms ( see Table S6 ) . As a control , Podr-1::rfp alone was injected into N2 worms ( 75 ng/µl ) . The resulting transgenic control strains ( CF2144 , CF2145 ) had mean lifespans that were the same as N2 . Lifespan assays in general were conducted as previously described [64] . For glp-1 mutant lifespan assays , eggs were incubated at 20°C for 2–6 hrs , transferred to 25°C to eliminate germ cells , then shifted back to 20°C on day 1 of adulthood for the rest of their lifespan . All other lifespan assays were performed at 20°C . In all cases , the L4 stage was counted as day 0 of adulthood . In all experiments with TCER-1::GFP , worms were examined for Podr-1::rfp co-injection marker under a Leica MZ16F stereomicroscope ( Wetzlar , Germany ) on day 0 and isolated for lifespan experiments . The non-transgene carrying siblings were used as negative controls in the same experiment . glp-1 mutant strains used in lifespan assays were completely sterile . Fertile strains were transferred every other day to fresh plates until progeny production ceased . Animals that crawled off the plate , exploded , bagged , or became contaminated were censored . Stata 10 . 0 and 8 . 2 ( Stata Corporation , Texas , USA ) and ( for some lifespans ) Statview 5 . 0 . 1 ( SAS ) softwares were used to calculate mean life spans and perform statistical analyses . P values were determined using log-rank ( Mantel-Cox ) statistics . We screened a C . elegans chromosome II RNAi library to identify RNAi clones that suppressed the extended lifespan of glp-1 mutants . Details of a similar screen were described earlier [13] . All RNAi experiments were performed as described previously [65] , [66] . In general , RNAi clones were inoculated overnight at 37°C in LB medium containing 10 µg/ml tetracycline and 100 µg/ml carbenicillin , and seeded onto NG-carbenicilin plates supplemented with 0 . 1 M IPTG . For all lifespan experiments , and the RNAi screen , worms were exposed to RNAi clones from hatching . All RNAi clones were confirmed by sequencing ( M13-forward primer ) and upon start of every experiment , by PCR ( T7 primers ) . Clones were obtained from the libraries described previously [65] , [66] . pAD12 ( empty vector ) was used as the negative control , and pAD43 ( daf-16 RNAi ) as the positive control for the screen and for individual lifespans [43] . The tcer-1 RNAi clone also targets a partial duplication of tcer-1 , the gene ZK1127 . 6 , whose predicted protein product contains only FF domains . Transcriptional and full-length translational GFP-tagged reporters for ZK1127 . 6 showed no basal expression or induction on germline ablation suggesting that it is a non-functional duplication of tcer-1 ( ZK1127 . 9 ) . Laser ablations of germ-cell precursors ( Z2 , Z3 ) were performed using a Zeiss Axiophot with a laser attachment ( Photonics Instruments , USA ) as described earlier [67] . Briefly , eggs were transferred to fresh plates and left at 20°C for 1–3 hrs . Hatched L1 larvae were mounted on slides with 2% Agarose pads containing 1 . 5 mM Sodium Azide anesthetic . As intact controls , worms were subjected to all the steps of this process except for ablation . Ablated worms ( and controls ) were removed from the slide and grown at 20°C until required for GFP assays or lifespan analysis . Eggs of worms carrying the Pdaf-16::daf-16::gfp reporter construct and the different DAF-16/FOXO-target gene reporters ( in wild-type , glp-1 mutant and tcer-1; glp-1 mutant backgrounds ) were incubated at 20°C for 2–6 hrs , transferred to 25°C to eliminate germ cells , then shifted back to 20°C on day 1 of adulthood . GFP assays were conducted on day 2 of adulthood , using a Leica MZ16F ( Wetzlar , Germany ) stereomicroscope with standard fluorescence filter sets . All assays were performed blind after initial familiarization of GFP levels in control plates by the experimenter . All fluorescence images were captured using a Retiga EXi Fast1394 CCD digital camera ( QImaging , Burnaby , BC , Canada ) attached to a Zeiss Axioplan 2 compound microscope ( Zeiss Corporation , Jena , Germany ) . Openlab 4 . 0 . 2 software ( Improvision , Coventry , U . K . ) was used for image acquisition . GFP assays were conducted on a Leica MZ16F ( Wetzlar , Germany ) stereomicroscope with fluorescence filter sets or ( in the case of Pnnt-1::gfp ) the Zeiss Axioplan 2 compound microscope mentioned above . Preliminary image processing was performed using Photoshop 10 ( Adobe Creative Suite 3 , USA ) . Total RNA was isolated from synchronized populations of approximately 15 , 000 day 1 daf-2 and tcer-1; daf-2 mutants . For comparing , tcer-1 ( − ) and tcer-1 ( − ) ; tcer-1 OE strains , worms were picked manually for RNA isolation . Eggs were transferred to fresh plates . Day 0 ( L4 ) transgenic animals carrying tcer-1 ( − ) ; tcer-1 OE were isolated using a Leica MZ16F stereomicroscope with standard fluorescence filters . 2 , 000–5 , 000 worms were picked manually for RNA isolation for each biological replicate . A similar number of tcer-1 mutants were also collected manually . Both strains were allowed to grow for 24 hrs and on day 1 of adulthood used for RNA isolation . Total RNA was extracted using TRIzol reagent ( Invitrogen ) and purified using Qiagen RNAeasy Mini Kit . cDNA was generated using Protoscript First Strand cDNA Synthesis Kit ( New England Biolabs ) . SybrGreen real-time Q-PCR reactions were performed on an Applied Biosystems 7300 Real Time PCR System . The primers used in this study are listed in Table S7 .
|
The reproductive status and longevity of animals are strongly interlinked . Increasing age influences the reproductive capacities of most animals . However , little is known about how reproductive status might affect lifespan . Experiments in worms and flies have shown that removing cells that give rise to gametes , the “germ cells” , makes them live longer . We know very little about the genes and molecules that are involved in this process . In this study , we have identified a gene called tcer-1 that promotes the longevity of the roundworm Caenorhabditis elegans when its germ cells are removed . The gene tcer-1 codes for a protein , TCER-1 , that is predicted to function as a “transcription elongation factor” ( it allows the completion of RNA synthesis during the process of gene expression ) . Our experiments imply that when the germ cells of worms are removed , TCER-1 collaborates with a transcription factor called DAF-16/FOXO to express genes that contribute to increased longevity . DAF-16/FOXO can extend lifespan in response to other physiological cues besides loss of germ cells . However , TCER-1 specifically helps this widely used longevity protein to respond to signals that reflect the reproductive status . Counterparts of DAF-16/FOXO are known to control aging in other organisms , including humans , so the identification of TCER-1 may lead to a better understanding of the relationship between reproduction and aging in other species , too .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"developmental",
"biology/germ",
"cells",
"developmental",
"biology/stem",
"cells",
"molecular",
"biology/transcription",
"initiation",
"and",
"activation",
"developmental",
"biology/aging",
"molecular",
"biology/transcription",
"elongation",
"developmental",
"biology/developmental",
"molecular",
"mechanisms"
] |
2009
|
A Transcription Elongation Factor That Links Signals from the Reproductive System to Lifespan Extension in Caenorhabditis elegans
|
The ecological success of social insects is often attributed to an increase in efficiency achieved through division of labor between workers in a colony . Much research has therefore focused on the mechanism by which a division of labor is implemented , i . e . , on how tasks are allocated to workers . However , the important assumption that specialists are indeed more efficient at their work than generalist individuals—the “Jack-of-all-trades is master of none” hypothesis—has rarely been tested . Here , I quantify worker efficiency , measured as work completed per time , in four different tasks in the ant Temnothorax albipennis: honey and protein foraging , collection of nest-building material , and brood transports in a colony emigration . I show that individual efficiency is not predicted by how specialized workers were on the respective task . Worker efficiency is also not consistently predicted by that worker's overall activity or delay to begin the task . Even when only the worker's rank relative to nestmates in the same colony was used , specialization did not predict efficiency in three out of the four tasks , and more specialized workers actually performed worse than others in the fourth task ( collection of sand grains ) . I also show that the above relationships , as well as median individual efficiency , do not change with colony size . My results demonstrate that in an ant species without morphologically differentiated worker castes , workers may nevertheless differ in their ability to perform different tasks . Surprisingly , this variation is not utilized by the colony—worker allocation to tasks is unrelated to their ability to perform them . What , then , are the adaptive benefits of behavioral specialization , and why do workers choose tasks without regard for whether they can perform them well ? We are still far from an understanding of the adaptive benefits of division of labor in social insects .
Social insects are enormously successful ecologically . Ants , social bees , social wasps , and termites may make up 75% of the world's insect biomass , they play a major role in soil turnover and nutrient cycling , and they often surpass vertebrates in their biomass in a habitat [1] . Division of labor is often cited as the primary reason for the ecological success of social insects , particularly ants [1–5] . Division of labor implies that individuals within a colony specialize on particular tasks , such as brood care , foraging , nest building , or defense , and conversely that each task is performed by a particular subset of the workers [6–12] . If division of labor caused ecological success in social insects , it must have conferred benefits to colonies . What exactly are these adaptive benefits of specialization ? According to the famous economist Adam Smith [13] , specialization in human industry had three benefits: ( 1 ) increased individual efficiency through learning , ( 2 ) reduction of switching costs , and ( 3 ) the invention of machines . The first of these may be called the “Jack-of-all-trades is master of none” hypothesis: specialists are individually more efficient at performing their task than generalists . Although this hypothesis underlies many discussions of division of labor [2 , 6 , 14–18] , it has rarely been tested , as pointed out by many authors [8 , 9 , 11 , 17 , 19–27] . Here , I address this issue by measuring individual efficiency of more than 1 , 100 workers of the ant species Temnothorax albipennis in several tasks . This allows me to test whether more specialized individuals are also more efficient . Most previous research on division of labor in insects concentrates on the mechanisms of task allocation ( e . g . , [6 , 11 , 27–32] ) instead of its consequences for individual or colony-level performance . Remarkably few studies have investigated the efficiency of individuals and how it relates to which tasks they perform [9 , 15 , 22 , 23 , 25 , 33–36] . In principle , there are two ways in which individuals who are specialists may achieve higher efficiency in performing “their” task: they may learn to perform a task better with frequent experience; or colonies may produce different specialists that are evolutionarily adapted to particular tasks . Worker polymorphism may be such an evolutionary adaptation: in ants with morphological castes , we know that “majors” ( morphologically specialized ants ) tend to be better at some tasks than the generalist “minors , ” for example , they may be good at cutting leaves or walking fast to transport them [35 , 37]; they may also be good at other transport , defense , or food storage: [25 , 38–41] , but are bad at performing brood care [22] . Polymorphism among workers , however , is rare , only occurring in less than 15% of ant genera [21 , 42] . Worker polymorphism also does not occur in bees or wasps . In bumble bees , workers exhibit size polymorphism , albeit not the variation in shape ( allometry ) characteristic of polymorphic ants . Workers may also differ genetically , leading to variation in behavior [43] . Although this may be unlikely to produce evolutionarily stable colony-level benefits [44] , it is clear that worker differences in task preferences [30 , 31 , 45] , as well as variation in quality of task performance [31 , 46 , 47] , may be linked to genetic variation , and such variation may therefore play a role in specialization . Individual efficiency can also increase through learning , without morphological adaptations . For example , bumble bees and honey bees need to invest time in learning to handle particular flower types efficiently [48 , 49] . Since learning incurs various types of cost ( production and maintenance of neural tissue , energy costs of actually using it , and costs in errors made and time invested [50–55] ) , it may be beneficial to minimize the number of skills that an individual has to learn . This could lead to increased individual efficiency in specialists . Indeed , many bee foragers specialize on particular flower types , possibly to minimize costs of learning handling procedures [48 , 56 , 57] ( although see [52] ) . So , is the specialist worker in a colony the “master” of one task , while the generalist is a “master of none” ? The Jack-of-all-trades is a master of none hypothesis would predict that more specialized workers perform a task with higher efficiency than generalists , whether this is a result of learning or adaptation . This hypothesis is what I test here for the ant Temnothorax albipennis . A set of tasks that are relevant in different contexts ( foraging , emigrations , and nest building ) will be used . Specifically , I will test which of the following specific hypotheses best predicts individual efficiency: ( 1 ) More specialized workers perform a task more efficiently . ( 2 ) Workers that are more active overall ( in different tasks ) perform tasks more efficiently . ( 3 ) Workers that engage in tasks with a short delay ( who may have low response thresholds ) perform tasks more efficiently than those who delay longer . To understand the role that individual efficiency plays in colony division of labor , I will also test the following two hypotheses: ( 1 ) At the colony level , most labor is contributed by highly efficient workers . ( 2 ) At the colony level , most labor is contributed by specialized workers .
A total of 1 , 142 ants from 11 colonies of Temnothorax albipennis were marked individually and filmed performing four tasks: carrying brood items in a colony emigration , foraging for honey solution , foraging for protein ( dead Drosophila flies ) , and collecting sand grains ( hereafter , “stones”—they are about a third the size of a worker ant ) as nest-building material . The four tasks studied have in common that it is possible to measure the amount of work performed per time by individual workers , without manipulation of colony composition , which may upset normal task allocation patterns . Of the colonies , four were large relative to the average colony size [58] in this species ( 147–233 workers ) and seven were small ( 27–100 workers ) . These colonies were also used in other studies , in which colony size effects on individual workload in emigrations and the difference between “elite workers” and “specialized workers” were investigated ( [12] and A . Dornhaus , J . -A . Holley , and N . R . Franks , unpublished data ) . In the study presented here , I focus on the quality of performance of individuals . For each individual ant , it was recorded how often it performed each task , how long it delayed before starting to perform the task ( “delay , ” see Materials and Methods ) , and how long it took the ant to perform two task units . A task unit ( hereafter: “trip” ) is defined as leaving the nest and returning to it while delivering either a brood item to a new nest site , a load of honey solution to a nestmate , a piece of fly ( i . e . , proteinaceous food ) to the nest , or a stone to the wall being built . Thus , the duration per trip reflects how much work a worker accomplished per time , and therefore can be used as a performance measure . I calculated the average duration per trip for the first two trips of each ant that performed a task ( hereafter called “performance” ) . For each ant and each task , a measure of specialization was also calculated: an ant was considered more specialized the more it concentrated its work effort in a single task . I used the proportion of total task performances ( trips ) that were in task X as a measure of specialization on task X . Thus , if all performances were in a single task , the worker's specialization for that task was 100%; if a task was never performed , specialization was 0%; and if all four tasks were performed equally frequently , specialization was 25% . Using these measures and ranking workers within each colony according to their performance , I find no correlation between specialization and performance for three tasks: brood transporting , honey foraging , and fly foraging ( Figure 1 ) . This means that in each colony , the workers that were the most specialized in a task were not necessarily the best performers . In the fourth task , collection of stones , I did find a significant impact of specialization on colony-level rank of performance ( Figure 1; regression for large colonies: p = 0 . 03 , R2 = 0 . 13 , df = 27; small colonies: p = 0 . 01 , R2 = 0 . 15 , df = 36 ) . However , this relationship was not in the direction predicted: workers with a high rank on specialization ( mostly performing stone collection ) had a high rank on duration of two trips , which means they were less efficient ( took longer to perform the same amount of work ) . Note also that these results remain unchanged if workers who performed a task only once are excluded ( large colonies: brood transports , p = 0 . 062; small colonies: brood transports , p = 0 . 38; honey foraging , p = 0 . 64; stone collection , p = 0 . 041; in no case were more than two workers excluded; if p-value is not given here , all workers performed the task at least twice or not at all ) . Instead of specialization , it may be that the overall number of trips across all tasks predicts a worker's performance . To test this , I performed a stepwise regression of performance on three factors: overall activity level ( total number of trips in all tasks performed by that worker ) , task-specific delay ( time from start of experiment to first task performance ) , and specialization . None of these factors consistently predicted performance ( Table 1 ) . Only in one case ( performance of brood transports in small colonies ) was specialization a significant factor , although even here , activity , i . e . , amount of work performed overall rather than in a specific task , was more predictive of performance , and only a small amount of the variation in performance was explained by either of these two factors . Another puzzling result is that I do not find that the contribution a worker makes to overall colony workload is predicted by worker efficiency ( Table 2 ) . This means that in a given colony , most of the work is not necessarily performed by those who are best at it; it is , however , often performed by those most specialized in a task ( Table 2 ) . This agrees with a previous result that work in Temnothorax albipennis is generally performed by specialists , not generalists ( A . Dornhaus , J . -A . Holley , and N . R . Franks , unpublished data ) . Colonies did not differ significantly in how well their workers performed , except in the task of brood transports ( Kruskal-Wallis Test , large colonies , transports: p = 0 . 024 , df = 3 , n = 115 workers; honey foraging: p = 0 . 47 , n = 56 , fly foraging: p = 0 . 38 , n = 25; stone collection: p = 0 . 20 , n = 28; small colonies , transports: p = 0 . 36 , df = 6 , n = 64 workers; honey foraging: p = 0 . 016 , n = 32 , stone collection: p = 0 . 26 , n = 37 ) . Median individual performance for colonies does not seem to depend on colony size ( Figure 2A; brood transports: p = 0 . 15 , R2 = 0 . 13; honey foraging: p = 0 . 89 , R2 < 0 . 001; fly foraging: p = 0 . 58 , R2 < 0 . 001; stone collection: p = 0 . 20 , R2 = 0 . 08 ) ; variation , measured as the interquartile interval , among individuals also seems mostly constant even in different-sized colonies , although it was significantly higher in larger colonies for brood transports ( Figure 2B; brood transports: p = 0 . 048 , R2 = 0 . 30; honey foraging: p = 0 . 06 , R2 = 0 . 33; fly foraging: p = 0 . 44 , R2 < 0 . 001; stone collection: p = 0 . 69 , R2 < 0 . 001 ) . Some individuals were only seen performing one task ( 35% of all workers in large , 32% in small colonies ) , and many were never seen performing any of the four tasks investigated here ( 52% of all workers in large , 57% in small colonies; see also [12] for inclusion of more tasks ) . Thus , the experimental conditions did not simply induce all workers to work at maximum level , which would have potentially obscured a normal relationship between specialization and individual performance . For workers whose performance was measured in at least two different tasks , their colony-specific rank in performance in one task did not correlate with that in the other task ( Figure 3; Regression p = 0 . 44 , R2 < 0 . 001 , n = 62 ) . As stated above , for each worker in each task , two task performances ( trips ) were measured: duration of the first trip correlates significantly with duration of the second trip for transports , but not for the other tasks ( regression on colony-specific ranks , transports: p < 0 . 001 , R2 = 0 . 19 , n = 101; honey foraging: p = 0 . 07 , R2 = 0 . 11 , n = 21; fly foraging: p = 0 . 83 , R2 < 0 . 001 , n = 8; stone collection: p = 0 . 18 , R2 = 0 . 05 , n = 19 ) . In honey foraging and stone collection , there was a trend in the same direction ( higher rank in trip 1 ∼ higher rank in trip 2 ) , but the sample sizes were lower than for transports; it thus cannot be said with certainty whether individuals were consistent in their performance over time .
Among the results presented here , two are particularly surprising: that colonies are not adapted to allocate the most efficient workers to each task , and that efficiency seems unrelated to the level of behavioral specialization of individuals . Many previous studies have simply assumed that if there is specialization , it will correlate with improved performance at a task ( although see [22 , 23 , 59–61] ) . My results indicate that at least in this species , a task is not primarily performed by individuals that are especially adapted to it ( by whatever mechanism ) . This result implies that if social insects are collectively successful , this is not obviously for the reason that they employ specialized workers who perform better individually . It also seems that individual performance ( at least in the four tasks investigated ) is not predicted by overall activity or that ant's delay to engaging in a task . Delay may or may not correspond to a task response threshold , an individual- and task-specific factor that defined the probability of engaging in a task . This factor has been used in many previous models of task allocation . It will be interesting for future studies to investigate whether response thresholds do or do not predict individual efficiency in performing tasks . The performance measure here was the average duration of two individual trips , which corresponds to the number of items brought to the nest per time . Although this is a performance measure that is often used and can be objectively quantified , it is possible that the performance of specialist ants was superior to that of generalists in some other way . Perhaps specialists carried larger loads ( although this seems unlikely in the nest-building case , as the sand grains were sieved to a uniform size ) , or perhaps specialists were able to collect more information or watch for predators while performing tasks . However , the time used per load , as measured here , varied by more than a factor of 40 ( for example , fastest brood transport was 100 s , slowest 4 , 363 s ) . Although it cannot be excluded , it seems unlikely that these differences were compensated by load size or minimization of predation risk . It is tempting to say that the ant species studied here , Temnothorax albipennis , is unusual in its colony organization . Maybe it employs less strict division of labor than other ant species ( although other measurements indicate that this is not the case: A . Dornhaus , J . -A . Holley , and N . R . Franks , unpublished data ) , or maybe because of their long lifespan ( workers can live several years in the lab ) , each individual already has had the opportunity to perfect its performance in each task . Also note that , as stated above , T . albipennis does not have worker polymorphism , so any differences in specialization among workers are the result of behavioral specialization only . However , the level of specialization in most social insect species is not known , and it can be argued that Temnothorax is representative of the majority of ant species: it has the same small colony sizes that are typical for most ants [1 , 62]; it forages by preying on and scavenging other arthropods in the leaf litter , as many other ants do [1 , 21]; it is monomorphic ( no allometry among workers ) as most other ants are [21 , 42]; and the genus Temnothorax is cosmopolitan and does not consist of ecological specialists adapted to particularly restricted habitats . To test whether the present results are widely applicable throughout the social insects , it would be desirable if future research employed a wide variety of study systems . That would enable an assessment of how widespread , across species , individual behavioral specialization is , and how it relates to efficiency . In addition , only four tasks that ants perform on a regular basis were studied here . There are a number of other relevant tasks , most notably brood care and colony defense against predators and parasites . Efficiency assays for these tasks should be developed and used to study the benefits of specialization . Although studying other tasks is important , the tasks studied here were previously thought to be the ones that are most likely to be influenced by learning and thus suitable for specialization [30 , 48 , 63] . Tasks that involve leaving the colony require skills of orientation and the specific learning of landmarks [64 , 65]; even in small laboratory settings , such learning can significantly affect performance [66] . Similarly , tasks that involve collecting prey or building material involve identification and handling skills that cannot be easily genetically preprogrammed , as the precise location and type of prey and building material is likely to vary with microhabitat , even within a population ( e . g . , [34 , 63 , 67–69] ) . Indeed the results presented here do not show that learning is absent in this species . Learning , in the context of task performance , may theoretically occur at three time scales . Short-term learning may increase performance from one trip to the next on the same day , without leading to long-term individual differences . Second , performance at foraging and emigration tasks may differ primarily between completely naive individuals who have never left the nest and individuals who have left at least once , i . e . , participated in at least one emigration or foraging bout ( or perhaps performed the equivalent of the well-studied “orientation flights” in honey bees , e . g . , [51 , 64] ) . Third , amount of experience may directly correlate with performance , such that the more experience an individual gains at a specific task over its lifetime , the better it is able to perform a task . The results in this study show that there is no correlation between quality of performance and specialization—suggesting that differential improvement through learning , as in the third type of learning listed above , does not occur . However , it is quite likely that the first two types of learning do occur . Previous studies have demonstrated learning in colony emigrations and foraging in this species [70–75] . However , in many cases the main improvements were achieved after the first performance of the task ( although see [69] for honey bees ) , suggesting an effect of learning similar to the second type discussed above . In summary , this study finds that there is a large amount of variation in individual quality of task performance , not explained by any of the factors studied . The mechanism creating this variation is unknown , and may be genetic , developmental , or an effect of experience . Thus , learning may well affect task performance , but either it affects all individuals equally , or workers do not preferentially perform the tasks in which they are experienced ( although the latter would contradict previous studies: [71 , 76 , 77] ) . Either way , learning does not seem to lead to superior performance by specialists . It will be important for future research to quantify at what time scales learning occurs , and whether it increases or decreases variance among individuals in the long term . In this study , I quantify quality of task performance for individual workers in several tasks . To show that specialization exists , it is necessary to show that an individual performs more of one task , and less of another , compared to nestmates [10 , 11] . It is not sufficient to measure how much an individual performs a single task: this may merely identify high-activity workers from low-activity workers . To measure the benefits of specialization ( at least in terms of individual efficiency or quality of task performance ) , it is equally necessary to compare performance in multiple tasks; otherwise one may simply identify high-quality workers from low-quality ones , without necessarily demonstrating that specialists are better at their task and worse ( or at least no better ) at another . If this is not the case , then one has demonstrated merely that there is variation in quality of task performance among individuals , and possibly that high-quality individuals tend to be allocated to particular tasks , but not that there are benefits to specialization . What , then , are the benefits of division of labor in species without polymorphic workers ? As mentioned above , there are at least three potential benefits of division of labor . Individually increased efficiency was only one of them . Others were a decrease in the costs associated with switching between tasks . For example , division of labor may lead to increased spatial efficiency , as hypothesized for ants [28] , or reduction of other , possibly cognitive , switching costs [57] . It is also possible that specialization simplifies the process of task allocation ( i . e . , minimizes neural or other costs associated with the task selection process itself ) , or optimizes material flow in multistep tasks , [15] . Any of these processes may create colony-level fitness benefits from division of labor , even without improvement in individual efficiency as measured here . Future studies should attempt to quantify switching costs , spatial constraints , the role of learning , and the time scales at which individuals specialize in social insect colonies . My study also highlights that findings from commonly used model species , such as honey bees or leaf-cutting ants , which have very unusual and specific ecology and morphology , cannot necessarily be extended to other species [20 , 21 , 78] . We have much yet to learn about the benefits and evolution of division of labor .
As mentioned above , all workers in 11 colonies of the ant species Temnothorax albipennis , collected in Dorset , England , were individually marked with paints ( a total of 1 , 142 ants were marked; details on this method , as well as colony collection and housing , can be found in [12] ) . The colonies were housed in artificial nests in the laboratory , made of a cardboard perimeter sandwiched between two glass slides . All colonies were filmed in three different contexts spaced at least a week apart ( emigrations , wall building , and foraging ) . Each colony was filmed for at least 180 min in each context , starting at the time of manipulation as described below . This resulted in 166 h of digital video tape . Each context was initiated as follows: emigration—removal of the top glass slide , exposing the ants ( a new , identical nest was offered in 10 cm distance ) ; foraging—colonies were starved for 2 wk ( no food , but water ad lib was offered ) , and then a small dish with honey solution ( 1:10 honey:water ) and a pile of ten frozen Drosophila flies were placed 10 cm from the nest entrance; building—colonies were housed in a nest that had no front wall , creating a 33-mm-wide gap; on the day after the ants had moved into this nest , and a pile of colored and sieved sand grains was offered 10 cm from the nest . Under the latter conditions , ants use the sand grains to build a wall to narrow the nest entrance to approximately 1–3 mm . No food was offered in the emigration and building contexts , and no building material was offered in the emigration and foraging contexts . Each of the three different contexts thus introduced the need to perform a particular set of tasks , creating an opportunity for each ant to take part in these tasks . By using separate contexts , ants' task choices were thus less affected by competing stimuli for different tasks , but solely by the individual's preferences for performing the task at hand . For example , if some individuals had both a high tendency to transport brood and to participate in nest building , their activity level in either of these tasks was not constrained by that in the other task . If all tasks had been offered at once , such individuals may have spent all their time transporting brood simply because it is the more urgent task . Only a separation of tasks as employed here allows the identification of specialists from highly active generalist individuals . From the video tapes , the time that each ant picked up a brood item in the old nest , a sand grain from the pile , or left the nest in a foraging trip was extracted ( start of trip ) . The time from the start of the experiment ( e . g . , removal of the nest cover in emigration experiments ) to the start of the first trip for each ant was designated its task-specific delay . Then , the time that the same ant returned to the old nest , dropped the sand grain at the nest , returned and performed trophallaxis ( to unload honey solution to a nestmate ) , or returned with a piece of dead fly to the nest was recorded ( end of trip ) . The time difference between start and end of trip give the trip duration . The sum of all trips made in one context by one colony was called the colony-level workload ( e . g . , total number of brood transports made in an emigration ) . An individual's colony-level work contribution was measured as that individual's number of trips divided by the total colony-level workload . All measurements were double-checked by a second person to ensure accurate records of ant identity and timing of task performances .
|
Social insects , including ants , bees , and termites , may make up 75% of the world's insect biomass . This success is often attributed to their complex colony organization . Each individual is thought to specialize in a particular task and thus become an “expert” for this task . Researchers have long assumed that the ecological success of social insects derives from division of labor , just as the increase in productivity achieved in human societies; however , this assumption has not been thoroughly tested . Here , I have measured task performance of specialized and unspecialized ants . In the ant species studied here , it turns out that specialists are no better at their jobs than generalists , and sometimes even perform worse . In addition , most of the work in the colony is not performed by the most efficient workers . So the old adage “The Jack of all trades is a master of none” does not seem to apply to these ants , suggesting that we may have to revise our understanding of the benefits of colony organization .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"evolutionary",
"biology",
"ecology"
] |
2008
|
Specialization Does Not Predict Individual Efficiency in an Ant
|
Toxoplasmosis is an important , widespread , parasitic infection caused by Toxoplasma gondii . The chronic infection in immunocompetent patients , usually considered as asymptomatic , is now suspected to be a risk factor for various neurological disorders , including epilepsy . We aimed to conduct a systematic review and meta-analysis of the available literature to estimate the risk of epilepsy due to toxoplasmosis . A systematic literature search was conducted of several databases and journals to identify studies published in English or French , without date restriction , which looked at toxoplasmosis ( as exposure ) and epilepsy ( as disease ) and met certain other inclusion criteria . The search was based on keywords and suitable combinations in English and French . Fixed and random effects models were used to determine odds ratios , and statistical significance was set at 5 . 0% . Six studies were identified , with an estimated total of 2888 subjects , of whom 1280 had epilepsy ( 477 positive for toxoplasmosis ) and 1608 did not ( 503 positive for toxoplasmosis ) . The common odds ratio ( calculated ) by random effects model was 2 . 25 ( 95% CI 1 . 27–3 . 9 ) , p = 0 . 005 . Despite the limited number of studies , and a lack of high-quality data , toxoplasmosis should continue to be regarded as an epilepsy risk factor . More and better studies are needed to determine the real impact of this parasite on the occurrence of epilepsy .
Epilepsy is a major chronic neurological disorder that affects about 70 million people worldwide [1] . However , its importance goes beyond mere numbers [2] . Most of its burden is felt in low- and middle-income tropical countries , where a number of infections that are important risk factors for epilepsy , are predominantly reported [3–5] . Parasitic infections are important causes of epileptic seizures or epilepsy , among other neurological and mental health conditions [6 , 7] . Infection with Toxoplasma gondii ( T . gondii ) in particular is reported to affect one-third of the world’s population , mainly in the low- and middle-income countries [8 , 9] . Certain currently available data strongly suggest the possibility of a relationship between toxoplasmosis and epilepsy [6] [10] , although results to the contrary have also appeared [11] . Therefore , we conducted a systematic review and meta-analysis of published data to estimate the risk of epilepsy due to toxoplasmosis .
To identify published studies on the association between toxoplasmosis and epilepsy , we conducted a systematic search of the literature published in English or French . The search was conducted on MEDLINE , INGENTACONNECT , REFDOC , SCIENCEDIRECT , GOOGLE , Médecine/Sciences , PLOS ONE and the database of the Institut d’Epidémiologie neurologique et Neurologie Tropicale ( IENT ) : http://www . unilim . fr/IENT/recherche_bvna . phpbase . This database contains medical dissertations , theses , and articles on tropical neurology and parasitology . The keywords for epidemiological aspects were: “epilepsy , Toxoplasma , toxoplasmosis” . Similar keywords were used in French . While searching , OR and/or AND were used for combination terms . We also searched the bibliographies of the articles retained after our database search . Articles were selected based on their titles and then their abstracts . Those retained were read in full , but only those articles that met the inclusion-exclusion criteria were finally included . Inclusion criteria were: epilepsy as a disease and toxoplasmosis as an exposure , presence of a control group , sample sizes suitably estimated , details of techniques used to diagnose epilepsy and toxoplasmosis , and details on selection of participants , including socio-economic level . Data were then entered in a database covering: title , principal author , year of publication , type of study , objectives , methods , results , and any additional comments . The meta-analysis was conducted using EasyMA version 2001 and Medcalc ( Belgium ) version 12 . 6 . 0 . The measure of association between toxoplasmosis and epilepsy was a common odds-ratio ( OR ) , recommended for the meta-analysis of observational studies . Respective ORs of each included study were individually verified . Random effects models were used to determine common odd ratios . Homogeneity of the studies was examined using Cochrane Q , 95% confidence intervals ( CI ) were also derived , and the statistical significance was set at 5% . A scatter plot was drawn from the combined data obtained .
The results of the search are presented in Fig . 1 . Five databases ( Pubmed , ScienceDirect , Refdoc , IENT and Ingentaconnect ) , two journals ( Médecine/Sciences , PlosOne ) and Google search gave a total of 684 articles . Of these , 301 were cited more than once . Of the remaining 383 articles on toxoplasmosis and epilepsy , 372 articles were eliminated on the basis of title as either non-epidemiological or in a language other than English or French . The remaining 11 articles were read as abstracts , and one was excluded because its full text was not accessible . Furthermore , two were literature reviews [12] [13] , one was a meta-analysis [10] and one was an epidemiological study that did not meet our inclusion criteria i . e . no control group [14] . The references of the review articles and meta-analysis did not add any new articles . Lastly , six studies were retained [2] [11] [3 , 15] [16] [17] [18] , from Israel , USA , Turkey , Iran and various countries of sub-Saharan Africa . Interestingly , no studies were from Latin America , Europe , or most of Asia , table 1 . As summarized in table 1 , three studies recruited cases from hospital ( s ) , but only one recruited controls from hospital ( s ) . The remaining studies ( n = 5 ) recruited controls from the community or used volunteers . Four studies matched controls for age and gender . Toxoplasma infection status was determined with the use of IgG ELISA , table 1 . As summarized in table 2 , only one study recruited subjects of all ages , as is often recommended . All but one study [11] reported a “risk relationship” between toxoplasmosis and the development of epilepsy , table 2 . Only one of the included studies provided any urban-rural information [16] . The total number of subjects in all included studies was 2888 , of whom 1280 had epilepsy and 1608 did not ( table 2 ) . As summarized in table 2 , the frequency of patients with epilepsy who were Toxoplasma-positive varied from 14 . 1% to 75 . 0% . Among subjects without epilepsy , the proportion of Toxoplasma positivity varied between 4 . 70% and 56 . 50% . Seroprevalence for T . gondii was therefore higher among those who had epilepsy than those who did not . The ORs of included studies varied from 1 . 17 ( 95% CI 0 . 99–1 . 39 ) to 5 . 35 ( 95% CI 2 . 15–13 . 30 ) but only two of these results were significant ( table 2 ) . Scatter plots of the six studies are presented in Fig . 2 . The common OR ( fixed effects model ) was 1 . 32 ( 95% CI 1 . 13–1 . 55 ) , p<0 . 001 . The heterogeneity was statistically significant , p = 0 . 006 . Therefore , we estimated a common OR by a random effects model at 2 . 25 ( 95% CI 1 . 27–3 . 98 ) , p = 0 . 05 . The test of homogeneity was clearly not significant in this case , p = 0 . 6 .
Good-quality epidemiological research will further help to draw broad conclusions on this important subject . These good-quality studies may at least be population-based with samples representative of all age-groups , and taking cases-controls from the same source population with at least 70% power . These studies should also use standard criteria to define various parameters and taking active epilepsy into account with clear description of procedures followed in the article . Few studies evaluating the risk of epilepsy following toxoplasmosis are available , and none cover Europe and Latin American regions . Most Asian countries also remain unaddressed except Israel , Turkey and Iran . Based on the currently available data , and their obvious limitations , it is still good to consider toxoplasmosis as a possible epilepsy risk factor , at least epidemiologically . However , many questions remain to be addressed in future studies . Combinations of parasites may have additive effects on consequential conditions such as toxoplasmosis and onchocerciasis on active epilepsy [27] .
|
Toxoplasmosis is a common parasitic infection and affects one-third of the global population . The burden this figure represents clearly signifies the public health relevance of toxoplasmosis . Epilepsy , another chronic condition , is often caused by a variety of infections that affect numerous low- and middle-income tropical countries—including toxoplasmosis . Earlier meta-analysis found only three studies and reported a 4 . 8-fold greater risk of epilepsy . This had many limitations , as discussed in this paper , our estimate is rather more robust , based on a higher number of studies , and corrects deficiencies that were present previously .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
Toxoplasmosis and Epilepsy — Systematic Review and Meta Analysis
|
The pathogenic non-cultivable treponemes include three subspecies of Treponema pallidum ( pallidum , pertenue , endemicum ) , T . carateum , T . paraluiscuniculi , and the unclassified Fribourg-Blanc treponeme ( Simian isolate ) . These treponemes are morphologically indistinguishable and antigenically and genetically highly similar , yet cross-immunity is variable or non-existent . Although all of these organisms cause chronic , multistage skin and systemic disease , they have historically been classified by mode of transmission , clinical presentations and host ranges . Whole genome studies underscore the high degree of sequence identity among species , subspecies and strains , pinpointing a limited number of genomic regions for variation . Many of these “hot spots” include members of the tpr gene family , composed of 12 paralogs encoding candidate virulence factors . We hypothesize that the distinct clinical presentations , host specificity , and variable cross-immunity might reside on virulence factors such as the tpr genes . Sequence analysis of 11 tpr loci ( excluding tprK ) from 12 strains demonstrated an impressive heterogeneity , including SNPs , indels , chimeric genes , truncated gene products and large deletions . Comparative analyses of sequences and 3D models of predicted proteins in Subfamily I highlight the striking co-localization of discrete variable regions with predicted surface-exposed loops . A hallmark of Subfamily II is the presence of chimeric genes in the tprG and J loci . Diversity in Subfamily III is limited to tprA and tprL . An impressive sequence variability was found in tpr sequences among the Treponema isolates examined in this study , with most of the variation being consistent within subspecies or species , or between syphilis vs . non-syphilis strains . Variability was seen in the pallidum subspecies , which can be divided into 5 genogroups . These findings support a genetic basis for the classification of these organisms into their respective subspecies and species . Future functional studies will determine whether the identified genetic differences relate to cross-immunity , clinical differences , or host ranges .
Non-cultivable pathogenic treponemes include three subspecies of Treponema pallidum: T . pallidum subsp . pallidum ( T . p . pallidum ) , T . pallidum subsp . pertenue ( T . p . pertenue ) and T . pallidum subsp . endemicum ( T . p . endemicum ) . These subspecies are human pathogens and cause venereal syphilis , yaws and bejel , respectively . Other very closely related species or isolates are Treponema paraluiscuniculi and the Fribourg-Blanc or Simian treponeme . T . paraluiscuniculi causes venereal syphilis in rabbits and is reportedly not infectious for humans [1] , [2] . The unclassified Simian treponeme was isolated from a baboon , causes a yaws-like disease in non-human primates , and is able to cause active infections in humans [3]–[5] . All of these organisms can be propagated in rabbits and cause disease following experimental inoculation of rabbits . Treponema carateum causes the human disease , pinta , but no strains of this organism are available . The infections caused by T . pallidum organisms are characterized by chronic infection with distinct early and late clinical manifestations . Syphilis , usually a sexually transmitted infection , is a highly invasive process and can involve virtually any organ or system including the central nervous system . In pregnant women , early syphilis infection often results in transmission to the fetus . Each year , approximately twelve million new cases of syphilis are estimated to occur globally [6] , [7] . Yaws and bejel affect approximately 3 million people worldwide and are transmitted by non-sexual direct contact , usually during childhood and largely affecting people living in remote villages in developing countries . Yaws and bejel have predominantly skin or mucous membrane and osseous manifestations [8]–[10] , with tissue destruction late in infection . Pinta causes significant skin discoloration in the late stages , but rarely causes tissue destruction . Unlike syphilis , these infections are said not to affect the central nervous or the fetus [9] , although some scientists question this statement [11] . T . paraluiscuniculi infection in rabbits appears to be a chronic , but clinically mild , process characterized by long-lasting crusty lesions of the genitalia , nose , and mouth [12] . Treponemal infections in non-human primates have not been traditionally associated with genital disease; however , a recent study by Knauf et al . [13] reports asymptomatic , moderate or severely destructive genital lesions ( and perhaps sexual transmission ) resembling human syphilis , caused by organisms classified phylogenetically as more closely related to the Fribourg-Blanc and T . pallidum subsp . pertenue isolates . The molecular basis for host specificity and the different clinical manifestations caused by these treponemes is not known . These organisms are morphologically identical [1] , [3] , [14]–[17] with very similar antigenic composition [18]–[23] , stressed by the fact that , to date , infection-induced antibody or cellular immune responses cannot distinguish species , subspecies or strains . Protective immunity is induced only by long-term infection and is subspecies-specific [24] . In cross-immunity experiments [1] in which initial infections in the rabbit model lasted at least 3 months , three scenarios are observed: 1 ) inoculation with a particular strain results in complete protection against re-infection with the homologous strain , 2 ) protection against re-infection with another strain of the same subspecies is variable or non-existent , and 3 ) protection against challenge with other species or subspecies is absent . These cross-immunity observations are in concordance with inoculation studies in humans conducted by Magnuson et al . [25] . Subjects with treated late latent syphilis challenged with the Nichols strain had either of two outcomes: 1 ) those that did not develop either clinical signs or serological evidence of re-infection , indicating immunity; and 2 ) those that had increases in serological titers and/or development of darkfield positive lesions after inoculation , interpreted as active reinfection with the challenge strain . Although there was no evidence for waning immunity in the subjects who were susceptible to reinfection , this is a possible explanation . However , the lack of cross-immunity among highly similar species/subspecies may also reflect differences in a set of immunologically “inconspicuous” epitopes , underlying immunodominant , but not protective , antigens such as Tp47 ( TP0574 ) . These immunodominant antigens may act as decoy systems as described for other bacterial pathogens [26] . Recent comparative analyses of whole genome sequences [27]–[30] ( Giacani et al . , unpublished ) reported <0 . 1% sequence differences among T . p . pallidum strains [29]; <0 . 2% between T . p . pertenue and T . p . pallidum subspecies [29]; and <1 . 2% between T . paraluiscuniculi and the human treponemes [31] , [32] . Sequence diversity is primarily localized to six hot spots [29] , which include regions encoding several members of the tpr gene family . The Tpr proteins represent candidate virulence factors , and have been the focus of intense research for the last decade . As a consequence , the distinct clinical presentations , host specificities , and variable cross-immunity studies suggest that foci of sequence diversity , including the tpr genes , may be the basis for explaining the differences described above for the treponemal infections . Sequence homology divides the tpr family , a group of twelve paralogs , into three subfamilies: Subfamily I ( tpr C , D , F and I ) , Subfamily II ( tprE , G and J ) and Subfamily III ( tprA , B , H , K and L ) . As we progressively gain a better understanding of this gene family , an essential role for many of these genes is more apparent . Several studies show that the Tpr antigens are expressed during infection and are able to elicit marked antibody and cellular immune responses in the infected host [33]–[40] . Of the encoded Tpr antigens , TprA , B , C , D , E , F , I , J and K have been predicted to be outer membrane proteins ( OMP ) [33] , [40] , [41] . Opsonization and/or vaccine studies with these proteins support surface exposure [33] , [37] , [42]–[44] and both antigenic variation ( TprK ) [45] and phase variation ( TprE , G , J ) [46] mechanisms have been identified in tpr members . Yet , the high invasiveness and ability of T . pallidum to persist for decades in the host suggest that this spirochete may rely not only on antigenic and phase variation for survival . To influence infection outcomes , T . pallidum may also employ other strategies including genetic drift , genetic shift and or pathoadaptive point mutations , which can arise either during long term evolution or rapidly during a single infection . An important body of evidence has accumulated showing genetic variation in specific regions of the T . pallidum genome among subspecies and among strains [47]–[56] . The present study demonstrates significant sequence diversity in the tpr gene family , which can have important implications in understanding evolution of these organisms , as well as cross-immunity , strain typing and vaccine design .
No investigations were undertaken using humans or human samples in this study . New Zealand white rabbits were used for strain propagation . Animal care was provided in accordance with the procedures outlined in the Guide for the Care and Use of Laboratory Animals , and all work was conducted under protocols approved by the University of Washington Institutional Animal Care and Use Committee . T . pallidum subspecies , T . paraluiscuniculi , and the Fribourg-Blanc treponeme were propagated in New Zealand white rabbits by intratesticular inoculation as previously described [57] . DNA was extracted for PCR amplification from the following isolates: T . pallidum subsp . pallidum ( Sea 81-4 , Mexico A , Bal 3 ) , T . pallidum subsp . pertenue ( Gauthier , CDC2 , Samoa D ) , T . pallidum subsp . endemicum ( Iraq B , Bosnia A ) , the Fribourg-Blanc treponeme ( Simian isolate ) and T . paraluiscuniculi ( Cuniculi A ) . These strains were selected to represent different species/subspecies , geographical regions of origin , years of isolation , and anatomical sources ( Table 1 ) . The sequences of the tpr genes for the T . p . pallidum Nichols and Street 14 strains were downloaded from their corresponding genome sequences , GenBank accession numbers NC_000919 . 1 and NC_010741 . 1 , respectively [27] , [28] . Although we determined tpr sequences for a number of other strains of T . pallidum subsp . pallidum , only strains defining the 5 identified genogroups of T . p . pallidum are included in this manuscript . To ensure that the correct strain was propagated and extracted , only one strain of treponeme was handled at any time during the propagation and freezing process , and rabbit ear tags as well as labels on tubes were double-checked . Bacteria were extracted from infected rabbit testes in sterile saline , collected in sterile 1 . 7-ml microcentrifuge tubes , taking precautions to prevent cross-contamination between samples , and spun immediately in a microcentrifuge at 1 , 000g for 10 minutes to remove rabbit debris , followed by centrifugation of the supernatant at 12 , 000g for 30 min at 4°C [57] . Pellets were resuspended in 200 µl of 1X lysis buffer ( 10 mM Tris [pH 8 . 0] , 0 . 1 M EDTA , 0 . 5% sodium dodecyl sulfate ) , and DNA was extracted with the Qiagen ( Chatsworth , Calif . ) kit for genomic DNA extraction as described in the manufacturer's instructions , but adding 50 ul of proteinase K ( 100 mg/ml stock solution ) and incubating the sample for 2 h at 65°C . After the final elution step in 200 µl of H2O , DNA was used for analysis by PCR and sequencing . The Nichols T . pallidum genome sequence [27] was used to design primers in the 5′ and 3′ flanking regions of the tpr genes to amplify the corresponding DNA regions from genomic DNA of the 10 treponemal strains . Table S1 lists the primers used for amplification and sequencing . Using genomic DNA as a template , whole ORF amplifications were performed in a 50-µl final volume containing 200 µM deoxynucleoside triphosphates , 1 . 5 mM MgCl2 , and 2 . 5 U of GoTaq DNA polymerase ( Promega , USA ) . For larger amplicons such as the tprG-F or tprJ-I operons [38] , the LongAmp Taq PCR Kit was used as instructed by the manufacturer ( New England Biolabs , USA ) . The products were cloned into the pCRII-TOPO or TOPO-XL ( long amplicons ) cloning vectors ( Invitrogen , USA ) according to the manufacturer's instructions . Plasmid DNA was extracted by using the Qiagen Plasmid Minikit ( Qiagen , USA ) , and two to ten clones for each strain were sequenced with the Applied Biosystems dye terminator sequencing kit ( Perkin-Elmer , USA ) . Consensus sequences were obtained with the CAP sequence assembly program [58] and ORFs from each strain at each locus were aligned using the MAFFT alignment program [59] . GenBank accession numbers are listed in Table S2 . Structural homologs were identified using the 3D jury approach [60] . Structural ( 3D ) models for TprC , TprD and TprI were generated using the TMBpro algorithm [61] . The orientation of the predicted loops in the TMBpro models , surface exposed vs . periplasmic , was determined as previously described by Randall et al . [61] . Signal peptide predictions were performed using the Predisi algorithm [62] .
Subfamily I tprs include the tprC , D , F , and I loci . Initial examination of deduced protein alignments from the Nichols strain showed a significantly high degree of sequence conservation within Subfamily I at the amino and carboxyl termini , with central unique regions [33]; however , discrete heterogeneity was later evident in the amino and carboxyl regions when additional strains were analyzed . The Nichols TprC/D proteins reportedly have porin activity and an OM localization [68] . Although not yet experimentally demonstrated , TprF and I are also predicted to have a cleavable signal peptide and to be surface exposed [33] , [40] , [41] , [44] . The tprC and tprD loci in the reference Nichols genome contain two identical coding sequences [27] . Earlier studies [37] , [56] identified tprC and tprD variants among strains and among the three pallidum subspecies . The present study significantly expands our knowledge of the sequences in the tprC and tprD loci , and a schematic representation of all variants at the C and D loci identified to date is presented in Figure 2 . Among the treponemal strains tested in this study , four alleles are found at the tprD locus: the reference tprD ( Nichols ) , the tprD2 allele ( Bal 3 , Mexico A , Sea 81-4 , Street 14 , Samoa D , Iraq B , Bosnia A , Fribourg-Blanc ) , a predicted truncated tprD2 ( Cuniculi A ) , and the tprD-like variants ( Gauthier , CDC2 ) . We previously referred to the sequence in the tprD locus of Gauthier as tprD3 [56] . However , we have now found a very similar ( but not identical ) sequence in the CDC2 strain , and we have chosen to call these “tprD-like” sequences , which are further described below . As in Nichols , those T . p . pallidum strains that have the Nichols tprD allele in the D locus also contain an identical copy of tprD in the C locus [37] , and none of the non-syphilis treponemes carries tprC/D ORFs identical to the Nichols strain . As previously reported by our group , tprD2 has four unique regions that differentiate it from Nichols tprD and the tprD-like sequences: a 330-bp central region and three smaller regions toward the end of the open reading frame ( Figure 2 ) [56] . The tprC locus of the tprD2-containing Bal 3 , Sea 81-4 , Street 14 and Mexico A T . p . pallidum strains contains tprC-like ORFs , with small sequence changes compared to the Nichols tprC . [37] . Overall , the sequence homology among tprC alleles is >95% . All pertenue , endemicum and the Fribourg-Blanc strains also have tprC-like sequences . As previously reported for T . paraluiscuniculi , Cuniculi A strain [31] , [32] , [36] , the tprC and D loci are occupied by two truncated tprD2 variants . In both tprC and tprD , sequence variation does not occur randomly , but rather is found in discrete variable regions ( DVRs; Supplemental Figures 2 . 1 . and 2 . 2 in Figure S2 ) . In the majority of cases , these base pair changes result in amino acid changes . Pore-forming activities for TprC/D have been recently reported by Anand et al . [68] . 3D predictions of peptides without signal peptides suggest typical β-barrel structures of 22 antiparallel transmembrane regions resulting in 11 loops at each end of the structure ( Figure 3 , top panel ) . Our analysis of 22 TprC/D sequences demonstrated seven DVRs , all of which co-localize with surface-exposed external loops predicted by the 3D models ( Figure 3 , and Supplemental Figure 2 . 1 and 2 . 2 in Figure S2 ) . In addition , these 3D predictions suggest four external loops with conserved sequences , located primarily in the amino-half of the proteins . This sequence variation in predicted surface-exposed peptide loops could have significant implications for cross-immunity . In the Nichols genome , tprF and tprI loci are 1107 and 1827 nucleotides long , respectively . Their sequences are identical except that tprF is a truncated version of tprI due to a 720 nucleotide deletion ( spanning the central and most of the 3′ region ) in tprF , resulting in a shorter ORF , frameshifting and a premature termination [27] . In T . p . pallidum strains , tprF genes are identical in all isolates sequenced to date ( Figure 1 and Supplemental Figures 1 and 2 . 3 in Figure S1 and S2 ) . In contrast to the syphilis strains , the pertenue and Fribourg-Blanc isolates have a full length ( not frameshifted ) duplicated tprI-like gene at the tprF locus . Interestingly , however , the tprF locus is deleted in the endemicum strains Iraq B and Bosnia A , and in T . paraluiscuniculi . tprI loci are virtually identical to each other in T . p . pallidum strains except for the presence of a few synonymous SNPs in the 5′ and central regions reported in Street 14 [28] . In contrast , however , tprF or tprI ORFs are absent in the rabbit pathogen T . paraluiscuniculi . [36] . For a more detailed analysis of the polymorphism observed in the tprF and tprI loci , a sequence alignment was generated including all genuine ( not truncated or replaced ) tprF and tprI loci from 11 strains ( Supplemental Figure 2 . 3 in Figure S2 ) . TprF and TprI in syphilis and non-syphilis organisms display DVR patterns resembling the heterogeneity observed in TprC and TprD above , though to a lesser extent ( Supplemental Figure 2 . 3 in Figure S2 ) . Changes are clustered in 9 DVRs spread throughout the protein sequences . Deduced TprF and TprI proteins are also predicted to be outer membrane proteins [33] , [40] , [41] , [68] . Structural predictions also suggest that TprF and TprI are homologs of transport porins with OM localization , and 3D predictions of TprF/TprI peptides without signal peptides ( Figure 3 , bottom panel ) yield typical β-barrel structures . Similar to TprC/D , all TprF/I DVRs show co-localization with predicted surface exposed loops ( Figure 3 bottom panel , and Supplemental Figure 2 . 3 in Figure S2 ) , again suggesting an important role for these variable regions during infection . The Subfamily II genes include tpr E , G , and J , which code for proteins nearly 800 amino acids in length with highly conserved amino termini , unique central regions and carboxyl ends with small unique gene-specific signatures [33] . tprE shows very limited sequence variation among strains and subspecies , however , the observed changes clearly segregate syphilis from non-syphilis treponemes and the Fribourg-Blanc strain ( Supplemental Figure 1 . 4 in Figure S1 ) . T . paraluiscuniculi has a tprGJ chimera ( predicted truncation ) in the tprE locus [36] . In contrast , the tprG locus is more diverse in its gene sequence , in that five different groups of ORFs can be found ( Figure 1 , Figure 4 and Supplemental Figure 1 . 6 . 1 and 1 . 6 . 2 in Figure S1 ) : 1 ) tprG sequences as described in the Nichols genome ( Bal 3 , Street 14 ) ; 2 ) a truncated tprG due to two single and one 3-nucleotide insertions ( position range 1885–1956 ) , frameshifting , and a premature stop at its 3′ end ( Sea81-4 ) ; 3 ) a tprGJ chimera , in which the 3′ end of tprG has been replaced by the corresponding region of tprJ as evidenced by the presence of a tprJ-specific signature ( TAACGGGAACCCTCTCCCTTCCGGCGGTTCCTCAGGGCACATTGGCCT ) near the 3′ end of tprJ ( Mexico A and all T . p . pertenue strains ) ; 4 ) a tprGI chimera in which the 5′ end of the ORF is homologous to the corresponding region of tprG , and its central and 3′ regions of the gene are homologous to the corresponding regions of tprI ( all T . p . endemicum strains ) ; and 5 ) a truncated tprGI chimera due to a single nucleotide insertion ( T . paraluiscuniculi and the Fribourg-Blanc strain ) . While four T . p . pallidum strains have the reference Nichols tprJ sequence , the T . p . pallidum Sea 81-4 strain and all non-pallidum treponemes studied to date contain a tprGJ chimera in the tprJ locus ( Figure 1 and Figure 4 ) . The rabbit pathogen , however , contains a tprGJ chimera that codes for a truncated protein due to an insertion in its 5′ end [36] . Subfamily III tprs show a reduced degree of homology among family members , compared to Subfamilies I and II , with only small regions of sequence identity scattered throughout the coding sequences [27] , [33] . This is contrasted by a lower level of sequence heterogeneity at each locus among strains , subspecies , and species . tprB shows no variation among all strains ( Supplemental Figure 1 . 2 in Figure S1 ) . Among strains and subspecies , the tprH locus also contains highly homologous sequences , with only a few point mutations , of which 3 SNPs consistently distinguish syphilis vs . non-syphilis organisms ( Supplemental Figure 1 . 7 in Figure S1 ) . In the tprA locus , at positions 706 to 711 , there is a short region containing either three or four CT dinucleotide repeats . Strains containing only three CT repeats carry a gene that codes for a truncated protein due to a frameshift leading to a premature stop ( Nichols , Mexico A , Street 14 and Bal 3 ) . In contrast , strains carrying tprA genes with four CT repeats ( the syphilis Sea 81-4 and all non-syphilis isolates ) have no predicted frameshift and generate a sequence encoding a full length TprA product ( Figure 5 ) . tprL ( tp1031 ) shows major changes among strains and subspecies . Re-analysis of this region in the all of the endemicum and pallidum strains and 8 additional syphilis strains ( Brinck Reid et al . , unpublished ) revealed a larger putative tprL ORF coding for a protein sequence of 602 amino acids , compared to 514 amino acids as previously reported for the Nichols and Street 14 strains [27] , [28] . In this extended ORF ( Figure 6 ) , an alternative start codon ( CTG ) was identified with a typical ribosomal binding site ( RBS , GGAGG ) . Furthermore , beginning at position −31 , a 15 to 17 nucleotide poly-G tract flanked by −10 and −35 σ70 signatures ( TAGACA and TGTTGT ) is evident ( Figure 6 ) . Unlike the TprL product annotated in the Nichols genome sequence , the extended TprL is predicted to have a putative OM localization , with a predicted cleavable signal peptide ( cleavage between positions 25 and 26 , VFS-EQ ) . Compared to T . p . pallidum and T . p . endemicum sequences , our analysis revealed a gene fusion in the T . p . pertenue and Fribourg-Blanc strains caused by a deletion of 278 nucleotides ( Figure 6 ) , encompassing the 5′ end and central regions of the tp1030 ORF and a small fragment of the 5′ end of tprL including its start codon . This deletion creates a hybrid sequence ( tp1030 and tprL , here called tprL1 ) of 1668 bp with the start codon ( ATG ) in the plus strand of tp1030 ( the tp1030 coding sequence is located on the minus strand of the chromosome ) in frame with the rest of tprL ( tp1031 ) . As a consequence , the first 130 nucleotides of this new pertenue tprL1 ( Figure 1 , Figure 6 and Supplemental Figure 1 . 9 in Figure S1 ) are unique , not found in T . p . pallidum or endemicum tprL . The new extended TprL ( in T . p . pallidum and T . p . endemicum and T . paraluiscuniculi ) and the newly predicted TprL1 proteins ( T . p . pertenue and the Fribourg-Blanc treponeme ) are 602 and 556 amino acids long , respectively . Because the first 44 amino acids of TprL1 are encoded by the plus strand , this region is unique to the yaws and simian strains , with no homologous peptide in the pallidum and endemicum proteins ( Supplemental Fig . 1 . 9 Figure S1 ) . This unique peptide sequence is also not found elsewhere in the chromosome . Unlike the newly predicted extended TprL , TprL1 does not have a predicted signal peptide ( Figure 6 ) . This raises the possibility that the pallidum and endemicum subspecies may have an OM-localized TprL , while this would be predicted to be absent in the pertenue subspecies .
The 12 treponemal isolates from the three T . pallidum subspecies ( pallidum , pertenue and endemicum ) , the Fribourg-Blanc treponeme , and T . paraluiscuniculi show pleomorphic genetic changes in the tpr family characterized by SNPs , indels , chimeric sequences , and even absence of entire ORFs . Initial comparisons of the currently available full genome sequences of the Nichols , Chicago C , Sea81-4 and Street 14 syphilis strains revealed a high degree of sequence identity and a remarkable conservation of their genome organization [30] ( and Giacani et al . , unpublished ) . The study by Mikalova et al . [29] confirmed these observations , reporting clustering of sequence divergence in only a handful of distinct genomic regions among syphilis and non- syphilis strains , similar to those identified previously by Weinstock and colleagues [69] . Many of the hot spots of diversity are located in genes encoding members of the Tpr antigen family . The present study , however , provides a detailed description of sequence diversity within this paralog family and uncovers a rich number of sequence modifications among species , subspecies and strains . Importantly , our analyses also indicate some alternative genes or modified loci . It is striking that much of the sequence diversity identified in the tpr genes segregates the strains into the same subspecies and species groups that were originally defined according to their modes of transmission , their natural hosts , and the diseases they cause . This is most effectively seen in the colored blocks in Figure 1 . Given that the tpr loci represent the primary regions comprising the extremely low genomic diversity among the T . pallidum subspecies , it is likely that the proteins encoded by these variant genes play a major role in the differing pathogenesis of syphilis vs . yaws vs . endemic syphilis . Assigning a definitive role for individual proteins or combinations of proteins in determining clinical outcomes , however , awaits the determination of the functions of the Tpr proteins and the ability to genetically manipulate these genes within the organism . To inform studies of possible location and function , computational and immunological studies can provide clues for individual gene products . Several arguments emphasize a key role for TprC and TprD during syphilis infection: 1 ) they are the targets of strong antibody and cellular immune responses [35] , [37] , [40] , [56]; 2 ) immunization with recombinant TprC/D induces partial protection against infectious challenge [37]; 3 ) their surface exposure is supported by opsonophagocytosis assays [68] ( Lukehart et al . , unpublished ) ; 4 ) TprC and D show sequence diversity among strains [37] ( and this study ) ; and 5 ) 3D models predict a typical β-barrel structure with surface-exposed loops that contain each of the regions where sequence diversity is localized ( this study ) . It is highly unlikely that the co-location of sequence diversity and predicted surface-exposed loops is coincidental . A recent study by Anand et al . [68] proposes an alternative model for TprC and TprF , suggesting that the amino terminus of these two proteins is localized in the periplasmic space . However , experimental evidence argues against this model . Recombinant amino terminal TprF/I peptide induces partial protection against homologous challenge in immunization experiments in the rabbit model [37] and elicits opsonizing antibodies upon immunization ( Lukehart et al . , unpublished ) , observations supportive of surface exposure . However , the TprC and D sequence diversity ( localized in the exposed DVR ) identified among subspecies in the present study may contribute to the variable degree of cross-protection observed among T . pallidum strains and subspecies in infection-induced immunity . In this context , it is possible that sequence differences in the DVRs of TprC and D could lead to subspecies- or strain-specific surface-exposed epitopes that are critical to opsonic function or other mechanisms of protection . Studies are ongoing to test this hypothesis . A recognized example of functionally important strain-specific epitopes is loop 5 of the OMP P2 protein of non-typeable Haemophilus influenzae , which is associated with elicitation of bactericidal antibodies and protective immunity [70] . An alternative , or complementary , function of variable surface-exposed loops ( e . g . DVR ) could be that of providing steric hindrance to prevent the immune system from recognizing conserved external loops on the antigen , which are perhaps essential for correct protein structure or function . It is noteworthy that TprC and TprD are each predicted by 3D analysis to contain 4 conserved external loops . During natural human infection and experimental infection of rabbits [37] , [56] , antibodies are made against TprC/D and TprD2 . In addition to TprC and D , the TprD2 variant is also predicted to have surface exposure [37] , [40] , [56] , and is found in both syphilis and non-syphilis treponemes ( Figure 1 ) . The regions unique to TprD2 also contain predicted external loops , thus adding another layer of complexity to the already existing set of predicted loops for TprC and D ( not shown ) . Our structural predictions of TprC/D showing co-localization of external loops with DVRs is strong support for our hypothesis that antigenic differences in surface exposed loops of TprC and D have functional significance in immunity to the T . pallidum subspecies , and may be determinants of cross-immunity among subspecies and strains . Of interest is the observation that the CDC2 strain maintained in Seattle ( originally obtained in 2005 from Rob George and Victoria Pope from the Centers for Disease Control in Atlanta , GA ) contains a tprD-like allele while the corresponding sequence reported by Mikalova et al . [29] contains a tprD2 sequence . Re-sequencing of the tprD locus of this strain using our original frozen stocks confirmed that the CDC2 strain indeed contains a tprD-like allele . Also , we have sequenced the tprD locus of the pertenue CDC1 strain , isolated in a neighboring village in Africa from where the CDC2 strain was obtained , and found that the CDC1 strain also contains a tprD-like gene . It may demand a significant effort to identify the source of discrepancy between our data and that of Mikalova et al . , perhaps requiring the analysis of the two CDC2 lineages over the last several years . In contrast to syphilis treponemes , the tprF and I loci in T . p . pertenue and the Fribourg-Blanc treponemes each contain identical full-length ORFs . Although their coding sequences are identical within each location , tprF and tprI are located in separate tprG-F and tprJ-I operons , respectively , and their expression may be differentially modulated . The number of G residues in a polyG string in their promoters controls phase variation of these operons [46] , and the binding of TpCRP ( Tp0262 ) to the promoters was shown to either increase ( tprJ ) or decrease ( tprG ) transcription of the operon [71] . The implications of a “double dose” of tprI in the non-pallidum strains might be reflected in the total amount of message made in tissue specific locations or in differential expression over time during infection . Preliminary studies of antibody reactivity in rabbits infected with T . p . pertenue Gauthier strain demonstrate high levels of antibody to TprI , consistent with high ( or double ) expression of the protein ( Lukehart et al . , unpublished ) . The strong resemblance of the TprI/F 3D predictions to the TprC/D structural models , and the co-localization of DVRs and external loops suggest analogous roles at the microbe-host interface . T . pallidum tprGI chimeras were identified by Giacani et al . [36] in T . paraluiscuniculi and also present in the whole genome sequences later reported by Strouhal et al . [31] and Smajs et al . [32] , whose unique sequence composition was also recognized by these authors . Our analysis shows that , in all strains of T . p . pertenue , T . p . endemicum and the Fribourg-Blanc treponeme , the G and J loci are occupied by either tprGJ or tprGI chimeric genes . In contrast , the Nichols reference tprG and tprJ genes are frequently found in syphilis isolates , but not in any pertenue , endemicum or the Fribourg-Blanc strains tested to date . Only the T . p . pallidum Mexico A and Seattle 81-4 strains carry the GJ chimeric gene in the tprG and tprJ loci , respectively . Of interest is the presence of three truncated chimeras encoded by the tpr E , G , and J loci in T . paraluiscuniculi . This , in addition to predicted truncations or absences of Subfamily I Tprs ( Figure 1 ) , is perhaps related to the inability of T . paraluiscuniculi to infect humans , although further study is needed to explore this issue more thoroughly . One might wonder whether the tpr chimeras identified in this study are artifactual , due to “jumping” between highly similar sequences during PCR amplification [72]–[74] . In our study , tpr chimeras are unlikely to be artifacts for two reasons: 1 ) independent PCR amplifications of treponemal DNA obtained from different strain harvests rendered identical sequences , and 2 ) published sequences obtained by multiple sequencing approaches also show the same chimeras [31] , [32] , [36] , [75] . With the exception of TprK , little is known about the other members of Subfamily III Tprs ( tprA , tprB , tprH , and tprL ) . TprA , B and L are predicted to be OMPs [40] , [41] , and Tpr B induces antibodies that promote opsonophagocytosis ( Lukehart et al . , unpublished ) . Sequence conservation of tprB and tprH across species , subspecies , and strains suggests a required function for these proteins in the biology of T . pallidum . Nucleotide repeats , whether in regulatory or coding regions , are frequently associated with modulation of gene expression in an ON-OFF manner . The structure of the promoter region of the newly proposed extended tprL ORF is highly reminiscent of modulation of gene expression by single nucleotide repeats in the promoters of porA and opc loci of Neisseria meningitidis [76]–[78] . One could argue that predictions of an extended tprL ORF may lack accuracy because of the assumption of CTG as start codon , an underrepresented start codon in the annotated Nichols T . pallidum genome . However , our predictions are supported by the identification of a typical RBS , as well as −10 and −35 σ70 signatures with intervening homopolymeric G repeats of variable lengths resembling classic bacterial phase variation systems . In tprA , the variable number of CT dinucleotide repeats creates frameshifting and premature termination , dividing strains carrying tprA genes coding for full length product from those encoding predicted truncated products ( Figure 5 and Supplemental Figure 1 . 1 in Figure S1 ) . This is another mechanism for possible phase variation . Our analysis of the tpr gene sequences is based on an approach of targeted PCR amplification , cloning , and sequencing a number of clones to obtain consensus sequences . The tpr ORF sequences appear to be unchanging within a given strain during infection . However , limited information at the population level invites speculation about the possible presence of genetically distinct subpopulations within isolates . Smajs et al . [28] , [79] reported that at least two subpopulations are present within the Nichols strain as defined by a ∼1 Kb deletion in the flanking region of tp0131 . Our approach could have overlooked underrepresented variant organisms within isolates and , if intrastrain variation indeed exists , our findings might then reflect amplification of the most predominant subpopulation . Small mutational changes , even SNPs , in coding or non-coding regions can affect transcription , translation , or folding of the protein themselves , of neighboring genes , or those at more distant sites [80]–[83] . This could explain , for example , some of the differences in transcription observed among treponemal strains [40] . On the other hand , the now standard use of template-based assembly of short stretches of sequence generated by newer sequencing technologies can overlook the existence of hybrid genes or missing ORFs , whereas our individual-ORF sequencing approach can clearly identify these variations . Renewed efforts to address all of the above questions may be effectively resolved using next generation approaches such as deep sequencing of targeted regions , single cell isolation , or whole transcriptome sequencing . How might knowledge of tpr sequence diversity be translated into tools that are relevant to persons who are infected with one of the pathogenic treponemes ? The geographical distribution of yaws and syphilis is not as distinct as decades ago , and travel or migration can serve to transport an infection between urban and rural settings , complicating diagnosis . Because of the re-emergence of yaws over the past 20 years [84] , etiological differentiation of yaws vs . syphilis infections is desirable , and a practical approach for diagnosis is needed . The overall reported genetic variability between syphilis and yaws treponemes ( 0 . 2% ) makes these organisms almost genetically indistinguishable , and existing serological tests fail to differentiate the infections . Several small signatures that differentiate the distinct species/subspecies have already been identified in several genes [47]–[49] , [51] , [52] , [85] . The unique sequence composition of TprL described here in pertenue vs . pallidum strains reveals a possible 90 amino acid sequence unique to non-yaws treponemes , which includes a 25 amino acid predicted signal peptide , as well as a 44 amino acid peptide specific to T . p . pertenue . Given that Giacani et al . [40] showed that the tprL ORF is actively transcribed in both syphilis and yaws treponemes during experimental infection , our findings could facilitate the development of targeted serological screening for differentiating these two infections . Treponemal infections are chronic , yet only a minority of infected persons develops the severe late manifestations of disease . Is it possible that small genetic markers in the infecting could predict clinical outcome ? We previously showed that rabbits infected intravenously with the Sea 81-4 strain had higher levels of cerebrospinal fluid ( CSF ) inflammation , compared to other infecting strains , while animals infected with Bal 7 had more severe skin disease [86] . Our more recent work in humans supports the hypothesis that disease outcome may be related to genetically defined strain types [55] . Subfamily II tprs and the arp genes were first utilized for strain typing purposes by Pillay et al . [50] , although they were not able to correlate strain type with clinical outcome . Using an enhanced strain typing system developed by Marra et . al . [55] , which includes the targets initially described by Pillay et . al . [50] and the tp0548 gene , we demonstrated that patients infected by 14d/f type strains were significantly more likely to have neurosyphilis [55] . Four of the pallidum strains shown to represent different genotypes in this report ( Nichols , Street 14 , Mexico A and Sea 81-4 ) fall into four different molecular types using the enhanced typing system . The correlation supports the possibility that sequence changes in the tpr genes may be related to specific disease manifestations . It is noteworthy that T . paraluiscuniculi causes a very mild infection in its natural host , compared to syphilis , and is unable to infect humans [2] , [12] , [87] . One possible explanation for mild natural infection and the failure to infect other hosts is the dearth of functional Tpr proteins in this organism: there are seven truncated Tpr proteins ( TprC , D , F , I , E , G and J ) in T . paraluiscuniculi . In contrast , all T . pallidum subspecies and the Fribourg-Blanc treponemes , which have fuller Tpr repertoires , can multiply in more than one vertebrate host and can cause infection in humans . The Fribourg-Blanc treponeme , isolated from non-human primates from a yaws-endemic region in Africa [3] , [4] , resembles very closely the tpr repertoire of yaws strains ( 10 out of 12 ORFs are of the same type ) , although it resembles T . p . endemicum at the G locus , implying shared evolutionary pathways , as previously proposed [29] , [65] , [88] , as well as common strategies of interaction between microbes and their host . Although the clinical outcome of infection is likely dependent upon several factors , including individual host immunity , inoculum size , and route of infection , sequence changes in the tpr genes could determine differences in antigenicity or function , resulting in different adaptive strategies and differences in pathogenicity . While the distribution of tpr gene variants among the 12 isolates studied here appears , in most cases , to be clustered by subspecies , some isolates in the T . p . pallidum group share tpr variants that are otherwise restricted to non-syphilis organisms . For example , Sea 81-4 contains four tpr ORFs present in the endemicum subgroups ( Figure 1 ) , and Mexico A contains the tprGJ chimera in the tprG locus . The recent demonstration of syphilis-like genital lesions and purported sexual transmission of a yaws-like treponeme in wild baboons [13] suggests that pathogenicity and mode of transmission may not , however , be completely hard-wired in the genome . The sharing of some tpr variants among individual pallidum strains and the non-pallidum strains confounds the concept of a purely genetic basis for the nature of the disease . These findings again raise the 1960's nature vs . nurture controversy between Hudson and Hackett with regard to the biological or environmental/epidemiological basis for the differing clinical manifestations seen among the treponematoses [89] , [90] . Based upon tpr sequencing , there is genetic heterogeneity ( five genogroups ) within the pallidum subspecies , as well as some overlap among subspecies and species . Rather than having discrete organisms for each treponemal disease , there may in fact be a genetic continuum of the pathogenic Treponema , individual components of which affect pathogenesis in an individual host in concert with social or environmental factors that influence routes of transmission and disease manifestations . Finding the answer to this question will depend upon the ability to genetically manipulate T . pallidum so that the effects of individual genes can be definitively assessed .
|
Pathogenic treponemes include three subspecies of Treponema pallidum ( pallidum , pertenue , endemicum ) , T . carateum , T . paraluiscuniculi , and the unclassified Fribourg-Blanc treponeme . Although they share morphology and have very similar antigenic profiles , they have traditionally been distinguished by mode of transmission , host specificity and the clinical manifestations that they cause . The molecular basis for these disease characteristics is not known . Comparative genomics has revealed that sequences differences among the species and subspecies are found in very localized regions of the chromosome . Many of these regions of sequence variation are found in the tpr genes , which encode a family of twelve candidate virulence factors , many of which are predicted to be outer membrane proteins . Most of the tpr-specific sequence changes are consistent within subspecies or species , supporting the historical classification of these organisms into separate subspecies and species . Functional studies are needed to determine whether any of the tpr gene differences are related to differences in host range , immunity , or clinical manifestations .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"emerging",
"infectious",
"diseases",
"gene",
"identification",
"and",
"analysis",
"genetics",
"microbial",
"pathogens",
"molecular",
"genetics",
"host-pathogen",
"interaction",
"biology",
"microbiology",
"pathogenesis"
] |
2013
|
Fine Analysis of Genetic Diversity of the tpr Gene Family among Treponemal Species, Subspecies and Strains
|
Epstein–Barr virus ( EBV ) is considered a ubiquitous herpesvirus with the ability to cause latent infection in humans worldwide . EBV-association is evidently linked to different types of human malignancies , mainly of epithelial and lymphoid origin . Of interest is the EBV nuclear antigen 3C ( EBNA3C ) which is critical for EBV-mediated immortalization . Recently , EBNA3C was shown to bind the E2F1 transcription regulator . The E2F transcription factors have crucial roles in various cellular functions , including cell cycle , DNA replication , DNA repair , cell mitosis , and cell fate . Specifically , E2F6 , one of the unique E2F family members , is known to be a pRb-independent transcription repressor of E2F-target genes . In our current study , we explore the role of EBNA3C in regulating E2F6 activities . We observed that EBNA3C plays an important role in inducing E2F6 expression in LCLs . Our study also shows that EBNA3C physically interacts with E2F6 at its amino and carboxy terminal domains and they form a protein complex in human cells . In addition , EBNA3C stabilizes the E2F6 protein and is co-localized in the nucleus . We also demonstrated that both EBNA3C and E2F6 contribute to reduction in E2F1 transcriptional activity . Moreover , E2F1 forms a protein complex with EBNA3C and E2F6 , and EBNA3C competes with E2F1 for E2F6 binding . E2F6 is also recruited by EBNA3C to the E2F1 promoter , which is critical for EBNA3C-mediated cell proliferation . These results demonstrate a critical role for E2F family members in EBV-induced malignancies , and provide new insights for targeting E2F transcription factors in EBV-associated cancers as potential therapeutic intervention strategies .
EBV nuclear antigen 3C ( EBNA3C ) is expressed in latent infection and essential for EBV-mediated B-cell transformation , which plays an intricate role in regulation of viral as well as cellular gene transcription [1 , 2] . A large proportion of cellular transcription factors are genes encoded by oncogenes and tumor suppressors [3] . Deregulated expression of transcription factors are critical contributors to the oncogenic process . Furthermore , the transcription factors involved in most of oncogenic signaling pathways ultimately control downstream gene expression which leads to transformation , tumorigenesis , tumor progression and metastasis . Under normal physiological conditions , these tumor-associated genes are tightly regulated by upstream transcription modulators , whereas their abnormal expression regulated by the transcription factors result in deregulated expression of multiple tumor-related genes and subsequent oncogenic progression . Interestingly , EBNA3C similar to well-known master regulators can interact with various cellular transcription regulators , including RBP-Jκ/CBF1 , PU . 1 , Spi-B , Nm23-H1 , IRF4/8 , CtBP , p300 , Prothymosin-α , HDAC1 , SUMO1 , SUMO3 [4 , 5] . Recently , we have shown that the IRF4 transcription factor is stabilized by EBNA3C , which leads to degradation of its family member IRF8 through the ubiquitin-proteasome pathway and subsequent inhibition of its tumor suppressive activity [5] . Also , EBNA3C was found to inhibit E2F1-mediated apoptosis through blocking its transcription activity , and prompting its degradation in a ubiquitin-proteasome dependent pathway [6] . E2F transcription factors , which involve activators and repressors of this family members , play critical roles in the regulation of gene expression and diverse cellular functions , such as cell cycle , DNA replication , DNA repair , cell mitosis , and cell fate . The E2F repressor members which includes E2F6 have been shown to inhibit the transcription of E2F-responsive genes through a pRb-independent pathway [7] . Previous studies demonstrated that E2F6 appears to behave as a dominant-negative repressor to inhibit the transcription of E2F-target genes by competing with other E2F family members , and can also function as an active inhibitor [8 , 9] . E2F6 plays an important role in hypoxia-induced apoptosis by regulating the E2F1/Apaf-1 pathway , and suppresses growth-associated apoptosis through counteracting the pro-apoptotic activity of E2F1 in human hematopoietic progenitor cells [10 , 11] . Enhanced expression of E2F6 can also alter important cell growth parameters . For example , increased E2F6 expression can delay the exit of cells from S phase [12] . However , a detailed mechanism of E2F6-mediated transcription repression needs further investigation . In our study , we demonstrated that EBNA3C enhances E2F6 expression in EBNA3C expressing cells and EBV-transformed LCLs . EBNA3C is associated with E2F6 specifically , and also forms a complex by interacting with E2F6 at its amino and carboxy terminal domains . In addition , EBNA3C was shown to be important for E2F6 stabilization and co-localized with E2F6 in the nucleus . Importantly , EBNA3C and E2F6 can both contribute to the reduction of E2F1 transcriptional activity , and compete with E2F1 for E2F6 binding . Therefore , EBNA3C-mediated deregulation of E2F6 and its impact on other E2F family members is an important contributor to EBV-associated malignancies . Our findings also infer that E2F6 or other E2F transcription factors can be targeted as potential therapeutic intervention strategies in EBV-associated cancers .
To investigate the expression levels of E2F family members after infection of primary cells , human peripheral blood mononuclear cells ( PBMC ) were infected with wild-type BAC-GFP-EBV , and cells were then collected at selected time points ( 0 , 2 , 4 , 7 and 15 days post-infection ) . The mRNAs were isolated and detected by subsequent Real-time PCR . The results demonstrated that the E2F family members have a unique pattern of expression after EBV infection ( Fig 1A ) . Most of E2F family members were down-regulated except E2F1 . Interestingly , the opposite trend was seen for expression of E2F1 and E2F6 , which suggests that E2F1 and E2F6 may have opposite roles during EBV infection ( Fig 1B ) . Our previous studies demonstrated that EBNA3C can act as a modulator for different transcription factors [13] and has the ability to interact with different members of the E2F transcription factor family [5] . To examine the comparative binding ability of E2F6 and other E2F family transcription factors with EBNA3C , HEK-293 and Saos-2 cells were transfected with E2F1 , E2F2 , E2F3a , E2F4 and E2F6 expression plasmids with EBNA3C expression construct . The results indicated that E2F1 , E2F2 , E2F3a , and especially E2F6 can strongly bind with EBNA3C ( Fig 1C and 1D , S1A Fig ) . In the context of viral infection , abnormal expression of E2F6 was observed as an independent contributor to nasopharyngeal carcinogenesis [14] . To exploit the role of EBNA3C in modulating E2F6 expression , E2F6 mRNA levels were detected in BJAB cells with wild-type or ΔEBNA3C BAC-GFP-EBV infection . The results showed that E2F6 could be up-regulated at 1 day post-infection ( dpi ) in both wild-type and mutant virus infection . However , E2F6 expression was stable in wild-type EBV infected cells and increased to higher levels at 7 dpi , while its expression was not up-regulated in ΔEBNA3C BAC-GFP-EBV infected cells ( S2A Fig ) . Meanwhile , the highest level of EBNA3C expression was also observed at 7 dpi ( S2B Fig ) . We believe that this strongly supports our hypothesis that E2F6 expression is specifically up-regulated by EBNA3C . To further prove that E2F6 up-regulation is indeed mediated by EBNA3C , E2F6 protein levels were determined by western blot using E2F6-specific antibody in EBV-negative BJAB , and EBNA3C stable expressing BJAB7 , BJAB10 cells . Our results clearly showed enhanced E2F6 expression in BJAB7 , BJAB10 cells compared to BJAB cells ( Fig 2A ) . We extended our studies in EBV-transformed LCL1 , LCL2 cells and also observed higher E2F6 expression in those cells in comparison with EBV-negative Ramos , BJAB cells ( Fig 2B ) . In addition , the up-regulation of E2F6 expression was also observed from the naturally isolated Burkitt's lymphoma cells Mutu Type III ( latency III ) when compared with Mutu Type I ( latency I ) ( S1B Fig ) . Interestingly , stable knockdown of EBNA3C in EBV-transformed LCL1 cells resulted in a substantial reduction in E2F6 protein levels ( Fig 2C ) . Furthermore , increasing amounts of an EBNA3C expression plasmid were transfected into EBV-negative Ramos and HEK-293 cells . Subsequent western blot analysis revealed that increased EBNA3C expression results in a gradual increase of E2F6 protein levels ( Fig 2D and 2E ) , suggesting a specific role of EBNA3C in upregulating E2F6 protein expression . In order to map the binding domain of EBNA3C with E2F6 , we performed GST pull-down assays using full length EBNA3C and its different truncated mutants . The indicated proteins were produced by in vitro transcription/translation assays and co-incubated with purified GST-E2F6 protein generated from bacteria . The interacting domains were identified by western blot analysis . Our results demonstrated that the N-terminal domain ( residues 100–200 ) and C-terminal domain ( residues 900–992 ) of EBNA3C can directly interact with E2F6 ( Fig 3A and 3E ) . Additionally , we also performed GST pull-down assays in BJAB , BJAB7 , BJB10 , LCL1 , LCL2 cells and observed strong association of E2F6 with EBNA3C ( Fig 3B , 3C and 3D ) . These results further corroborates our above results that EBNA3C physically associates in a complex with E2F6 in vitro . To determine the endogenous complexes which would demonstrate association of EBNA3C with E2F6 in the B-cell background , we performed co-immunoprecipitation assays in different B-cell lines , including BJAB , BJAB7 , BJAB10 , LCL1 and LCL2 cells . These results clearly showed that EBNA3C interacts specifically with E2F6 in EBNA3C stably expressing cells as well as EBV-transformed B-cells ( Fig 4 ) . Using specific EBNA3C antibodies , we showed that by co-immunoprecipitation assays E2F6 was brought down in this complex ( Fig 4A ) . The reverse co-immunoprecipitation assay showed that with E2F6 specific antibodies EBNA3C was precipitated with the complex ( Fig 4B ) . Similarly , co-immunoprecipitation experiments performed in EBNA3C stable expressing B-cell lines demonstrated a protein complex of EBNA3C and E2F6 ( Fig 4C and 4D ) . To examine whether the association of EBNA3C and E2F6 takes place in the same cellular compartments , we initially performed immunofluorescence experiments by transfecting GFP-EBNA3C and Flag-E2F6 plasmids into U2OS or Saos-2 cells . Our studies showed that EBNA3C signals were punctate foci with exclusion of nuclei as expected , and E2F6 signal exhibited an intense staining pattern of prominent foci in the nucleus . Importantly , there were a number of places where co-localization staining of EBNA3C with E2F6 was evident ( Fig 5A and 5B ) . To further validate the interaction between these two proteins under more relevant physiological conditions , BJAB , BJAB10 and LCL1 cells were used in the following immunofluorescence assays . Our results clearly demonstrated that endogenous EBNA3C and E2F6 strongly co-localized in BJAB10 and LCL1 cells compared with BJAB cells , supporting the above evidence that they formed a molecular complex in nuclear compartments in EBNA3C stable expressing as well as EBV-transformed B-cells ( Fig 5C ) . Additionally , the co-localization assay was extended using different mutants of EBNA3C and the results demonstrated that N-terminal or C-terminal of EBNA3C could co-localize with E2F6 in nucleus , while the middle domain ( residues 366–620 ) of EBNA3C showed negligible co-localization ( S4 Fig ) . Our previous results indicated that E2F6 protein levels were significantly enhanced by EBNA3C , further suggesting an important role of EBNA3C for the maintenance of E2F6 protein expression . To further validate the observations that EBNA3C-induced E2F6 protein expression levels were due to enhanced E2F6 protein stability , we executed protein stability assays by expressing E2F6 with or without EBNA3C in MEF and Saos-2 cells . These studies showed that E2F6 was stabilized on expression of EBNA3C ( Fig 6A and 6B ) . We further extended this experiment by treating BJAB , BJAB7 , LCL1 , and sh-ctrl ( stable control knockdown ) , sh-EBNA3C ( stable EBNA3C knockdown ) LCL1 cells with cycloheximide ( CHX ) . The results showed that the stability of E2F6 was significantly enhanced in the presence of EBNA3C in comparison with a loss of greater than 75% of the E2F6 signal by 12 hours post-cycloheximide treatment in EBNA3C-negative cells ( Fig 6C and 6D , S1C Fig ) . Previous studies suggested that E2F6 can act as a repressor by inhibiting the transcription of known E2F-responsive genes [9] . To determine whether EBNA3C can play an important role in regulating E2F1 promoter activity involving E2F6 , we performed promoter luciferase assays by expressing increasing amounts of E2F6 , EBNA3C or co-expressing E2F6 with increasing amounts of EBNA3C in the presence of wild-type E2F1 promoter in HEK-293 and Saos-2 cells . These results showed that E2F6 and EBNA3C reduced E2F1 transcriptional activity in a dose-dependent manner in these two cell lines ( Fig 7A–7F ) . Both E2F6 and EBNA3C were effective at independently suppressing the promoter activity . However , when both were added a more dramatic reduction was seen compared to each independent activity ( Fig 7 , compare A/B , D/E with C and F , respectively ) . The inhibition of EBNA3C-mediated E2F1 promoter activity was reversed by down-regulating E2F6 expression with sh-E2F6 ( S3 Fig ) . Interestingly , we also showed that E2F1 was able to increase its own promoter activity but again showed a reduction in promoter activity when E2F6 and EBNA3C were co-expressed ( Fig 7G and 7H ) . Our previous binding assays above demonstrated that E2F6 interacts with EBNA3C in the same region as E2F1 . Therefore , it is possible that E2F1 can form a molecular complex with E2F6 and EBNA3C . To validate this hypothesis , co-immunoprecipitation experiments were performed in B-cells . Our results showed that E2F1 can form a protein complex with E2F6 and EBNA3C in EBV-transformed , EBNA3C stable expressing cells ( Fig 8A and 8B ) . Furthermore , competitive protein-binding assays were performed in HEK-293 and Saos-2 cells by transfecting increasing amounts of EBNA3C expression plasmids with Myc-E2F6 and Flag-E2F1 . The proteins of interest were examined by co-immunoprecipitation using anti-Myc antibody . Our results showed that increasing amounts of EBNA3C can lead to a dose-responsive reduction of binding between E2F1 with E2F6 ( Fig 8C and 8D ) . It has been shown earlier that E2F6 regulated E2F1 expression through the trans-repression of the E2F1 promoter [10] . To explore the effect of EBNA3C on E2F6 recruitment to the E2F1 promoter , a ChIP assay was executed using E2F6 antibody in different cell lines ( HEK-293 and BJAB , BJAB7 , LCL2 cells ) . The transfected or endogenous E2F1 promoter DNA were detected by real-time PCR in HEK-293 cells . Our results demonstrated that both the transfected and endogenous E2F1 promoter DNA were significantly immunoprecipitated when E2F6 and EBNA3C were co-expressed compared with the control groups ( Fig 9A and 9B ) . This suggests that accumulation of E2F6 at the E2F1 promoter is greatly enhanced by EBNA3C . Moreover , similar results were also observed in BJAB7 and LCL2 cells ( Fig 9C ) . These results showed that the presence of EBNA3C obviously increased the binding of E2F6 on E2F1 promoter . Our studies above showed that EBNA3C promotes E2F6-mediated E2F1 inhibition . This then leads to cell proliferation . To further explore whether E2F6 was a key factor in this pathway , stable knockdown of E2F6 in BJAB and BJAB7 cells were generated by lentivirus transduction and selection ( Fig 10A and 10C ) , and E2F6 expression levels were monitored by western blot analysis ( Fig 10B and 10D ) . The following CFSE assay was designed to detect cell proliferation in stable E2F6 knockdown B-cell lines . The results demonstrated that E2F6 knockdown in BJAB cells did not affect cell proliferation . However , in the corresponding BJAB7 cells , which stably expressed EBNA3C , cellular proliferation was significantly inhibited ( Fig 10E ) . To further assess these results in context of EBV , E2F6 knock-down LCL1 stable cell lines were generated , and the cell proliferation was also investigated ( S5 Fig ) . Compared with control cells , E2F6 knock-down indeed inhibits the ability of cellular proliferation in LCLs . This suggested that E2F6 was a critical factor for EBNA3C-mediated cell proliferation , and that E2F6 was recruited by EBNA3C as a major contributor to increased cell proliferation in EBV-transformed LCLs .
E2Fs are a designated large family of transcription factors having one or more conserved DNA binding domains . They bind with the target promoters to regulate their expressions [15] . E2Fs can cooperate to generate a coordinated network of carefully regulated cell cycle and cell proliferation events during cancer development . The E2F family of transcription factors can be categorized into several subgroups based on their potential functional activities . E2F1–E2F3 can be regarded as activating E2Fs [16 , 17] . They are required for transactivation of the target genes which are involved in transition through the G1/S phase of the cell-cycle , as well as proper cell-cycle progression [16 , 18 , 19] . E2F4 and E2F5 are considered to have predominantly repressive activities and bind the pRb protein family members [20] . E2F6 , E2F7 and E2F8 , are considered to be transcription repressors and regulate downstream gene expression through an Rb-independent mechanism [21 , 22] . Although E2F family members have been studied for many years , we still know little about how they collaborate to control cellular proliferation or tumorigenesis ( S6 Fig ) . A recent study showed that the dynamic interactions of NF-κB and E2F1/E2F4 can control cellular proliferation [23] . The well-known tumor suppressor Rb directly interacts with the E2F transcription factors and can be recruited to E2F-responsive promoters to control diverse signaling pathways and lead to aberrant cell proliferation or human malignances [24 , 25] . Studies explaining the Rb/E2F pathway have revealed that it is critical for controlling cell proliferation , and plays central roles in cancer development [26] . Our previous study showed that E2F1 could also be down-regulated by EBV latent protein EBNA3C in Rb-independent pathway [27] , but it is still unknown whether this regulation could be mediated by other E2F family members . Our current studies demonstrated that EBNA3C-mediated E2F6/E2F1 interaction in the context of EBV infection , and further provides a deeper understanding of the dynamic regulation of E2Fs in EBV-induced malignancies . E2F6 represents a third subclass of the E2F family which is likely to show distinct biological characteristics when compared to the other E2Fs . Previous studies suggested that E2F6 appears to function as a dominant negative inhibitor of known E2F-responsive genes through competition with other E2F family members , and also act as an active repressor [8 , 9] . E2F6 has important roles in regulating hypoxia-induced apoptosis via its modulation of E2F1/Apaf-1 pathway , and repressing growth-associated apoptosis by counteracting the pro-apoptotic activity of E2F1 in human hematopoietic progenitor cells [10 , 11] . In the context of viral infection , abnormal expression of E2F6 was found as an important contributor to nasopharyngeal carcinogenesis [14] . In the current study , we showed that EBNA3C plays an important role in induction of E2F6 expression in LCLs . Our study also demonstrated the physical association between EBNA3C and E2F6 , and that EBNA3C interacts with E2F6 through its amino and carboxy-terminal domains . Our endogenous co-immunoprecipitation assays clearly showed that EBNA3C and E2F6 can specifically associate in a protein complex with each other in B-cells . Recent studies suggested that EBNA3C expression can regulate the stabilization of various oncoproteins , transcription factors and cellular kinases [5 , 28 , 29] . We performed stability assays for E2F6 to determine whether EBNA3C has a direct role in stabilizing E2F6 and observed in presence EBNA3C , the E2F6 protein was substantially stabilized . The mechanism of EBNA3C-induced E2F6 stability is likely to be through the control of the ubiquitin proteasome degradation pathway . However , the specific E3 ligase is yet to be identified . Our study also demonstrated that EBNA3C and E2F6 strongly co-localizes in nucleus . Earlier reports suggested the role of EBV latent antigens in regulation of E2Fs , for example , the EBV protein BRLF1 activates S phase entry through inducing E2F1 expression [30] . Other reports stated that E2Fs are involved in LMP1 mediated downregulation of p27KIP1 transcriptional activity [31] . Nuclear export of E2F4/E2F5 by LMP1 blocked p16INK4a–Rb pathway and regulated cell proliferation [32] . Previously our lab demonstrated that E2F1 mediated apoptosis induced by the DNA damage response was blocked by EBNA3C in lymphoblastoid cells [6] . We now clearly demonstrate that EBNA3C and E2F6 are both responsible for reduction of E2F1 transcriptional activity . Moreover , E2F1 forms a protein complex with EBNA3C and E2F6 , and EBNA3C competes with E2F1 for E2F6 binding . E2F6 is also recruited by EBNA3C to bind the E2F1 promoter , and is critical for EBNA3C-mediated cell proliferation ( Fig 11 ) . Therefore , in the context of EBV infection , EBNA3C-mediated E2F6/E2F1 regulation may offer novel insights to further understand the important role of EBV latent antigens in E2Fs-related cellular functions or cancer development . Recently , E2F6 was identified as a component of mammalian Bmi1-containing polycomb complex [33] . The polycomb multi-protein complexes maintain epigenetic regulation of cell type specific gene expression patterns which are critical for cellular identity via histone modification . Dysregulation of these activities promotes oncogenesis by causing inappropriate expression of cell identity , differentiation and proliferation related genes . Interestingly , endogenous E2F6 was found associated with other polycomb group proteins , including RYBP , Ring1 , MEL-18 , mph1 and Bmi1 [33 , 34] , suggesting that E2F6 may modulate cell signaling pathways through recruitment of the polycomb transcription factors [9] . Polycomb group complex proteins regulate different cellular functions with PRC1 and PRC2 complex . PRC1 complex is recruited to chromatin through the recognition of H3K27me3 by chromobox ( CBX ) proteins [35 , 36] , and PRC2 complex is recruited to chromatin which allows tri-methylation of histone H3 Lys27 by EZH2 [35] . RNF2/RING1 homologs that are included in PRC1 complex can catalyze H2A monoubiquitylation at Lys119 [37] . Then RYBP repressor protein recognizes H2A monoubiquitylation and represses transcription activity [38] . Based on these studies , it is also important to suggest the EBNA3C may affect the functions of these cellular antigens and related pathways through E2F6 , which may provide new insights for targeting E2F transcription factors in EBV-associated cancers as potential therapeutic intervention strategies . In regard to EBV-mediated lymphomagenesis , many signaling pathways have been regulated by latent antigens during EBV latent infection [39] . EBV latent antigen EBNA3C , one of the essential proteins for in vitro primary B-cell transformation , was identified as a transcriptional modulator that can efficiently regulate the transcription of both cellular and viral genes [40 , 41] . It was reported that EBNA3C associates with RBP-Jκ and regulates Notch-induced transcription which was very important for LCL propagation [1] . Additionally , EBNA3C was also shown to interact with a range of transcription factors or regulators , including c-Myc [42] , SUMO1 , SUMO3 [43] , HDAC1 [44] , CtBP [45] , DP103 [46] , p300 , Prothymosin-α [47] , Nm23-H1 [48] , p53 [4] , Mdm2 [49] , ING4 and ING5 [50] . As a master transcriptional co-regulator , many canonical domains of EBNA3C are essential for protein-protein interactions . Thus , it is possible for EBNA3C to bind with several factors at once or temporally to regulate important cellular pathways including the cell cycle . These functional domains will help us understand how EBNA3C regulates numerous cellular pathways by interacting with multiple proteins . A comprehensive virus-host interaction network implies a complicated regulatory mechanism for EBV-mediated tumorigenesis . Besides , the identification of EBV super-enhancer provides further evidence for the interaction network [51]; suggesting that EBV latent antigens could interact with many proteins at once and regulate special gene-regulatory sites in order to govern cellular functions . The ongoing deep-sequencing data analysis will provide further information to deconstruct the complex virus-host interaction network important for these cellular pathways including gene regulation . Overall , our work has demonstrated that E2F6 has adopted intricate strategies to target E2F1 for its transcriptional repression and that leads to deregulation of its cellular activities in EBV-mediated pathogenesis . Our study now show a role for EBNA3C in regulating E2F6 expression as well as E2F6 mediated modulation of the activities of E2F1 transcription factor . Furthermore , these results demonstrate how EBNA3C-mediated deregulation of E2F6 alters the cellular proliferative function of B-cells to be transformed into malignant cells . Our current study strongly demonstrates that the EBNA3C-E2F6 interaction promotes cellular proliferation by regulating E2F1 expression . This also offers a novel therapeutic target in E2F6 for targeted killing of EBV-associated B-cell lymphomas .
PBMC ( human peripheral blood mononuclear cells ) were obtained from UPENN immunology core donated by different de-identified healthy donors [52 , 53] . All the protocols were approved by the Institutional Review Board ( IRB ) and conducted according to the Declaration of Helsinki . And each donor gave written , informed consent . Myc-tagged EBNA3C full length and truncated mutants , and GFP-tagged EBNA3C expression plasmids were described previously [5] . Flag-tagged and GST-tagged human E2F6 expression construct were generated by inserting human E2F6 gene in pA3F and pGEX2T vector with EcoR1 and Not1 restriction digestion respectively . Wild-type Flag-tagged-E2F1 was described earlier [6] , full-length pGL2-basic wild-type human E2F1 promoter construct was kindly provided by Dr . Joseph R . Nevins ( Duke University ) [54] . The plasmids expressing E2F2 , E2F3a , E2F4 and E2F6 were generously provided by Andrew D . Wells ( University of Pennsylvania , Philadelphia , PA ) and used to generate Flag-tagged expression constructs . Wild-type BAC-GFP-EBV and EBNA3C null mutant ( ΔEBNA3C BAC-GFP-EBV ) were described previously [52 , 55] . Mouse anti-Flag ( M2 ) antibody was purchased from Sigma-Aldrich ( St . Louis , MO ) . Hybridomas for mouse anti-Myc ( 9E10 ) , and anti-EBNA3C ( A10 ) were previously described [6 , 56] . Mouse anti-E2F6 ( TFE61 ) antibody was purchased from Abcam ( Cambridge , MA ) . Rabbit anti-E2F1 ( C-20 ) antibody was obtained from Santa Cruz Biotechnology ( Santa Cruz , CA ) . Mouse anti-GAPDH antibody was bought from US-Biological ( Swampscott , MA ) . HEK-293 ( human embryonic kidney cell line ) and pRb null Saos-2 ( human osteosarcoma cell line ) were obtained from Jon Aster ( Brigham and Woman's Hospital , Boston , MA ) . U2OS ( human osteosarcoma cell line ) was purchased from the American Type Culture Collection ( ATCC ) . MEF ( mouse embryonic fibroblast cell line ) was a gift from Xiaolu Yang ( University of Pennsylvania ) . HEK-293 , Saos-2 , U2OS and MEF cells were maintained in Dulbecco's modified Eagle's medium ( DMEM; Hyclone , Logan , UT ) supplemented with 5% fetal bovine serum ( FBS ) , 25 U/ml penicillin , 50μg/ml streptomycin , and 2mM L-glutamine . EBV-negative Burkitt’s lymphoma cells Ramos and BJAB were kindly provided by Elliot Kieff ( Harvard Medical School , Boston , MA ) . Mutu I and Mutu III cells were kindly provided by Yan Yuan ( School of Dental Medicine , University of Pennsylvania , Philadelphia , PA ) . BJAB stably expressing EBNA3C cells ( BJAB7 , BJAB10 ) were prepared by transfecting pZipneo EBNA3C into BJAB cells followed with neomycin selection [57] . LCL1 and LCL2 cells were EBV-transformed immortalized LCLs ( lymphoblastoid cell line ) generated in our laboratory [58] . Lentivirus-mediated stable EBNA3C knockdown ( sh-EBNA3C ) or scramble control ( sh-ctrl ) LCL1 cells generated in our laboratory were previously described [53 , 59] . PBMC , Ramos , BJAB , BJAB7 , BJAB10 , LCL1 and LCL2 were grown in RPMI 1640 media ( Hyclone , Logan , UT ) supplemented as described above . Bio-Rad Gene Pulser II electroporator was used to transfect HEK-293 , Saos-2 , MEF and B-cells by electroporation . The electroporation condition is 210 V and 975 μF for HEK-293 , Saos-2 , MEF cells , and 220 V and 975 μF for Ramos cells . For immunofluorescence assay , cells were transfected with jetPRIME ( Polyplus Transfection , Illkirch , France ) according to the standard protocols . As previously described [52] , 10 million Peripheral blood mononuclear cells ( PBMC ) were incubated with BAC-GFP-EBV supernatant , with a multiplicity of infection ( MOI ) of 5 , in 1ml of RPMI 1640 media with 10% FBS . Cells were collected after 4 hours incubation at 37°C , re-suspended in 2ml of complete RPMI 1640 media , and cultured in 6-well plate . EBV GFP expression was measured by fluorescence microscopy . After indicated times of post-infection , the cells were harvested and detected expression levels . The results were derived from three independent experiments . Quantitative real-time PCR ( qPCR ) analysis was carried out as previously described [49] . Briefly , total RNA was isolated using Trizol reagent ( Invitrogen , Inc . , Carlsbad , CA ) , treated with Dnase I ( Invitrogen ) , and reversed to cDNA by using High-Capacity cDNA Reverse Transcription Kit ( Applied Biosystems Inc . , Foster City , CA ) , according to the manufacturer’s instructions . Real-time PCR was performed using a Power SYBR Master Mix kit ( Applied Biosystems ) and the Step One Plus Real-time PCR system ( Applied Biosystems ) . The results were normalized to the endogenous control , GAPDH . Each sample was performed in triplicate . The primers used in this study are listed in S1 Table . The constructs of full length EBNA3C and its different truncated mutants were used to produce in vitro translated proteins using the T7-TNT transcription/translation kit ( Promega ) . Cell lysates from different B-cells ( BJAB , BJAB7 , BJAB10 , LCL1 , LCL2 ) were harvested and prepared . The translated proteins or cell lysates were co-incubated with GST control protein or purified GST-E2F6 protein from bacteria , followed by addition of GST beads . Then the protein mixtures were washed three times with binding buffer ( 0 . 1% NP-40 , 0 . 5 mM DTT , 10% glycerol , and protease inhibitors in 1X PBS ) , analyzed by 10% SDS-PAGE , and detected with western blot using mouse anti-EBNA3C antibody ( A10 ) . Cells were washed with 1X Phosphate Buffered Saline ( PBS ) for twice , and lysed in RIPA buffer ( 10 mM Tris pH 7 . 5 , 0 . 5% NP-40 , 2 mM EDTA , 150 mM NaCl ) with 1 mM PMSF and protease inhibitors for 1 hour at 4°C . After pre-cleaning with normal mouse/rabbit serum and 35 μl protein-A/G ( 1:1 mixture ) -coupled Sepharose beads , cell lysates were incubated with appropriate antibody ( 1 μg/ml ) to capture the interest proteins by rotating overnight at 4°C . Immunoprecipitations were washed three times in RIPA buffer . To perform western blot analysis , the immunoprecipitated samples and 5% of total cell lysates were separated by SDS-PAGE and transferred to nitrocellulose membranes , followed by incubating with appropriate antibodies and scanning using the Odyssey scanner ( LiCor Inc . , Lincoln , NE ) . U2OS or Saos-2 cells were transfected with different expression constructs using jetPRIME ( Polyplus Transfection , Illkirch , France ) . After 36 hours post-transfection , cells were fixed and permeabilized with 4% paraformaldehyde ( PFA ) and 0 . 1% Triton X-100 followed by blocking with 5% BSA at room temperature . Flag-tagged E2F6 was incubated with mouse anti-Flag ( M2 ) antibody and Alexa Fluor 594-conjugated secondary antibody ( Molecular Probes , Invitrogen , Carlsbad , CA ) , while GFP-tagged EBNA3C was detected by GFP-fluorescence directly . BJAB , BJAB10 , and LCL1 cells were semi-air-dried on slides in culture hood and fixed as mentioned earlier . Endogenous EBNA3C and E2F6 proteins were detected with their specific primary antibodies and the corresponding secondary antibodies . Nucleus were stained with DAPI ( 4′ , 6′ , -diamidino-2-phenylindole; Pierce , Rockford , IL ) . Finally , the cells on slides were washed in 1X PBS for three times and mounted in an antifade mounting media . The images were collected by Fluoview FV300 confocal microscope and FLUOVIEW software ( Olympus Inc . , Melville , NY ) was used for image analysis . 10 million MEF or Saos-2 cells were transfected with combinations of constructs by using the BioRad electroporation system . After 36 hours post-transfection , the transfected cells as well as indicated B-cells were incubated with 40 μg/ml cycloheximide ( CHX ) at specific time points . Then cells were harvested , lysed with RIPA buffer , and analyzed by western blot . Odyssey 3 . 0 software was used to quantify the band intensities . Luciferase assays were performed as previously described with few modifications [27] . Briefly , 10 million HEK-293 or Saos-2 cells were transfected with luciferase reporter and different combination of plasmids as indicated . At 36 hours post-transfection , cells were harvested and lysed for luciferase assay using dual luciferase assay system ( Promega ) . The luciferase as well as renilla activities were measured using LMaxII384 luminometer ( Molecular Devices , Sunnyvale , CA ) . The results represent experiments performed in duplicate . Chromatin immunoprecipitation ( ChIP ) was performed as previously described [60] . Briefly , 30 million HEK-293 cells were transiently transfected by electroporation with E2F1 luciferase reporter plasmid or control vector , and expression constructs Myc-EBNA3C or Flag-E2F6 . 48h of post-transfection , cells were cross-linked by 1% formaldehyde , collected , sheared DNA to an average length of 600bp by sonication , as confirmed by agarose gel electrophoresis . Cross-linked DNA was immunoprecipitated with anti-E2F6 antibody or normal IgG and subjected for Real-time PCR analysis with primers designed for specific regions of E2F1 promoter . The primers used in this assay are: for exogenous E2F1 promoter 5'-GGTACCATCCGGACAAAG-3' and 5'-GGTTCCTATTGGCTTTAACG-3'; for endogenous E2F1 promoter 5'-GCAGCAGTGGGCAATAGA-3' and 5'-CACCGGAATCCCTGTAAT-3' [10] . For B cell lines , about 50 million cells were harvested , immunoprecipitated with either E2F6 specific antibody or normal IgG and processed as above . The experiments were performed in duplicate and standard deviations ( SDs ) were indicated by error bars . The two sense strands of E2F6 shRNA are 5’-tcgagtgctgttgacagtgagcgaAAGGATTGTGCTCAGCAGCTGtagtgaagccacagatgtaCAGCTGCTGAGCACAATCCTTgtgcctactgcctcggaa–3’ ( sh-E2F6-1 ) , and 5’- tcgagtgctgttgacagtgagcgaTTGATGTATCGCTGGTTTATTtagtgaagccacagatgtaAATAAACCAGCGATACATCAAgtgcctactgcctcggaa–3’ ( sh-E2F6-2 ) , respectively . The upper-cases designate E2F6 target sequences , while lower cases specify hairpin and sequences . These sense stranded oligos were annealed with their respective anti-sense stranded oligos , and then cloned into pGIPZ vector using Xho I and Mlu I restriction sites . Besides , a sh-ctrl plasmid including control shRNA sequence 5’-TCTCGCTTGGGCGAGAGTAAG–3’ ( Dharmacon Research , Chicago , IL ) was used as a negative control . Lentivirus production and transduction of B-cell lines has been described previously [53] . B-cells were collected and suspended in 1 X PBS at a concentration of 10 X 106 cells/ml . Then CFSE solution was added to make a final concentration of 5μM . Equal volume of 1X PBS containing 5% FBS was added after 10 mins incubation at room temperature . Cells were washed three times with 1X PBS containing 5% FBS and equally divided into several plates for incubation . At different time points ( 0 , 24h , 48h , 72h ) , cells were harvested , washed and fixed in 4% PFA . The cells were washed 3 more times with ice-cold 1 X PBS and resuspended in 5000μl 1 X PBS , then analyzed on FACScalibur cytometer ( Becton-Dickinson Inc . , San Jose , CA ) and FlowJo software ( Treestar , Inc . , San Carlos , CA ) . Data represented are as the mean values with standard errors of means ( SEM ) or standard deviation ( SD ) . Statistical significance of differences in the mean values was analyzed using the 2-tailed student's t-test . P-value below 0 . 05 was considered here as significant ( *P < 0 . 05; **P < 0 . 01; NS , not significant ) . Epstein-Barr virus ( EBV ) genome , strain B95-8-GenBank: V01555 . 2 , EBNA3C ( Human herpesvirus 4 ) -NCBI Reference Sequence: YP_401671 . 1 , E2F1 ( Homo sapiens ) -NCBI Reference Sequence: NM_005225 . 2 , E2F2 ( Homo sapiens ) -NCBI Reference Sequence: NM_004091 . 3 , E2F3 ( Homo sapiens ) -NCBI Reference Sequence: NM_001949 . 4 , E2F4 ( Homo sapiens ) -NCBI Reference Sequence: NM_001950 . 3 , E2F6 ( Homo sapiens ) -NCBI Reference Sequence: NM_198256 . 3 .
|
EBV is associated with a broad range of human cancers . EBV-encoded nuclear antigen 3C ( EBNA3C ) is one of the essential latent antigens important for deregulating the functions of numerous host transcription factors which play vital roles in B-cell immortalization . The family of E2F transcription factors are involved in diverse cellular functions . More specifically , E2F6 is one of the E2F family members with a unique property of transcriptional repression . Our current study now demonstrates that EBNA3C can enhance E2F6 repressive functions , and is also responsible for increased E2F6 protein expression in EBV-transformed LCLs . EBNA3C directly interacts with E2F6 at its amino and carboxy terminal domains . Additionally , E2F6 was stabilized by EBNA3C and co-localized in nuclear compartments . Our study also demonstrated that EBNA3C and E2F6 expression resulted in decreased transcriptional activity of E2F1 , and that EBNA3C , E2F6 and E2F1 can form a protein complex , and EBNA3C competes with E2F1 for E2F6 binding . The recruitment of E2F6 by EBNA3C was also shown to be important for its related cell proliferation . These results showed a crucial role for EBNA3C-mediated deregulation of E2F6 and its impact on the activities of other E2F family members . Our findings also provide new insights for targeting these E2F transcription factors as potential therapeutic intervention strategies in EBV-associated cancers .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"blood",
"cells",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"affinity",
"chromatography",
"gene",
"regulation",
"regulatory",
"proteins",
"immunology",
"cell",
"processes",
"dna-binding",
"proteins",
"plasmid",
"construction",
"dna",
"transcription",
"immunoprecipitation",
"transcription",
"factors",
"dna",
"construction",
"molecular",
"biology",
"techniques",
"glutathione",
"chromatography",
"research",
"and",
"analysis",
"methods",
"white",
"blood",
"cells",
"amino",
"acid",
"specific",
"chromatography",
"cell",
"proliferation",
"animal",
"cells",
"proteins",
"gene",
"expression",
"molecular",
"biology",
"precipitation",
"techniques",
"chromatographic",
"techniques",
"biochemistry",
"antibody-producing",
"cells",
"cell",
"biology",
"b",
"cells",
"genetics",
"biology",
"and",
"life",
"sciences",
"cellular",
"types"
] |
2016
|
EBV Nuclear Antigen 3C Mediates Regulation of E2F6 to Inhibit E2F1 Transcription and Promote Cell Proliferation
|
There is clear empirical evidence that environmental conditions can influence Ascaris spp . free-living stage development and host reinfection , but the impact of these differences on human infections , and interventions to control them , is variable . A new model framework reflecting four key stages of the A . lumbricoides life cycle , incorporating the effects of rainfall and temperature , is used to describe the level of infection in the human population alongside the environmental egg dynamics . Using data from South Korea and Nigeria , we conclude that settings with extreme fluctuations in rainfall or temperature could exhibit strong seasonal transmission patterns that may be partially masked by the longevity of A . lumbricoides infections in hosts; we go on to demonstrate how seasonally timed mass drug administration ( MDA ) could impact the outcomes of control strategies . For the South Korean setting the results predict a comparative decrease of 74 . 5% in mean worm days ( the number of days the average individual spend infected with worms across a 12 month period ) between the best and worst MDA timings after four years of annual treatment . The model found no significant seasonal effect on MDA in the Nigerian setting due to a narrower annual temperature range and no rainfall dependence . Our results suggest that seasonal variation in egg survival and maturation could be exploited to maximise the impact of MDA in certain settings .
Soil-transmitted helminth infections affect approximately 1 . 5 billion people worldwide [1] , with periodic mass deworming playing a key role in control and elimination efforts . More efficient allocation of control effort resources therefore has the potential to improve the lives of many millions of people , with studies like the DeWorm3 initiative working to determine the feasibility of interrupting transmission [2] . Ascariasis , infection of the small intestine by the parasite Ascaris lumbricoides , is one of the most common of these infections [1] and the life cycle of the parasite involves egg exposure to environmental conditions during larval stage development [3] . Experimental studies on Ascaris suum eggs , a closely-related species of ascarid , have shown that changes in temperature can affect maturation , viability and mortality [4–6] . It is likely that temperature also affects A . lumbricoides eggs , and that data from this and related species can be used to predict climatic effects on ascariasis [7–9] . For a related ascarid in pigs , A . suum , high temperatures are associated with a trade-off between faster maturation and higher mortality [5] , such that an optimum temperature exists for maximum viability . This optimum temperature has been estimated for the ascarid of dogs , Toxocara canis , as around 25°C [10] , whereas at temperatures below 10°C little or no evidence of development was recorded for either A . suum or T . canis , even after multiple months of observation [6 , 10] . Rainfall is also expected to impact the life-cycle and onward transmission , but there is greater uncertainty around the magnitude and mechanism of this effect . It appears that minimal rainfall is needed to maintain soil water content above a required threshold for development of A . suum larvae [11] . Moisture requirements are better characterised for strongylid nematodes of livestock , for which fecal matter often already contains sufficient moisture for rainfall to not be considered a limiting factor to development , at least in temperate climates [12 , 13] . There is some evidence that excess water can lead to accelerated development of ascarid larvae [11] and that survival rates are higher in environments with higher moisture [14] , but it is possible that the greatest impact of rainfall on the infection cycle is through transmission . Rain is associated with greater sequestration of eggs through the soil and studies have shown that soil samples taken during rainy seasons often produce the highest yield of viable A . lumbricoides ova [15] . Climates which exhibit wet and dry seasons may also see changes in human behaviour that could impact transmission during these periods; for example , consumption of pickled vegetables during the late autumn to winter season has previously been suggested as a driver of reinfection in South Korea [16] . Historical field studies of ascariasis have found seasonal peaks in prevalence [17 , 18] and reinfection rates [16] . One study , treating at different times of year from 1977 to 1978 across six hamlets in Kyunggi Do province , Korea , found that the highest peak in transmission occurred in early spring; a difference of 23 . 5% was observed between the highest and lowest reinfection rates [16] . Strong seasonal variation in reinfection rates has also been recorded in Saudi Arabia [15] , with the optimal period for larval survival and transmission coinciding with cooler temperatures and a brief rainy season . A more recent field study of 477 individuals in Sri Lanka [19] found positive correlations between wet-days per month and both infection and re-infection rates . In contrast to parasite control programmes in humans , anthelmintic treatment of livestock populations routinely takes account of seasonal variation in infection pressure . Gastrointestinal nematode infection typically peaks in summer in temperate areas [20] and during the rainy season in arid and semi-arid regions [21] . Management factors such as winter housing and concentration of birthing in spring or rainy seasons , when grass availability is highest , modify these seasonal patterns [22] . Nevertheless , effects of climatic drivers , especially temperature and rainfall , on the development and survival of infective larvae are well documented [23] and explain seasonal variation in levels of infection [24] . Models in which climate drives infection pressure are able to predict observed seasonal patterns [25–27] . Treatment generally aims to protect animals during periods of heightened risk , or to eliminate egg output in advance of conditions suitable for larval development . Thus , suppression of egg output is widely used as a management tool , and is most effective when calibrated to local climatic conditions [28 , 29] . In seasonally arid regions , treatment during periods hostile for free-living parasite stages was once recommended in order to minimise reinfection; however , this favours the development of anthelmintic resistance [30] . Improved ability to predict nematode infection risk for livestock in terms of climate has led to model-driven farmer decision support tools , which are sensitive to seasonal variation in infection pressure [31 , 32] . For A . suum , egg maturation driven by summer temperatures and prolonged survival in the winter forms the basis for recommended seasonal control strategies in pigs [33] . Despite the precedent set in the veterinary sector , the majority of public health programs have yet to adopt seasonal timing of mass drug administration ( MDA ) for A . lumbricoides control due to a lack of empirical evidence on the expected impact of such a move . Drugs are distributed through existing infrastructures , such that adjusting procedure can incur significant financial and operational costs , which the benefit of seasonal treatment would have to outweigh . However , the gain in reduced public health burden from seasonally targeted treatments could be high for certain climates , with areas that see large variations in temperature and rainfall likely to display the most pronounced differences . The key aims of this theoretical study are to propose a novel model for A . lumbricoides transmission that incorporates some of the seasonal elements of the system , and in doing so to demonstrate the potential impact seasonally-timed treatment could have in different climates and prevalence settings .
A model reflecting four key stages of the A . lumbricoides life cycle , is used to describe the level of infection in the human population alongside the environmental egg dynamics ( see Fig 1 . The full equations describing the model can be found in the Supporting Information ) . This is a new model framework inspired by the well-established two-stage delay differential equation model developed by Anderson and May that considers the interaction between the mean worm burden ( M ) and the number of infective larval stages present in the immediate environment ( L ) [34] . Using a similar approach to Fowler et al [35] , this new framework is easier to describe , implement and fit as it removes the need for delays . Here we additionally consider the mean number of juvenile worms per host ( J ) and the total count of immature eggs in the environment ( E ) . This allows the removal of delays from the system such that maturation of both juvenile stages in the host and eggs in the environment can be represented as rates: 1/τ1 and 1/τ2 respectively , where τ1 and τ2 represent average maturation times . Death rates μ1 and μ2 for the within-host stages , J and M , incorporate both parasite and host mortality . For the environmental stages , E and L , death is taken to occur at rates γ1 and γ2 respectively . The excretion rate of eggs into the environment , sNλM , is calculated using the worm gender ratio , s = 0 . 5 , the worm fecundity , λ , the human population size , N , and the current level of mean worm burden , M . Ingestion occurs at rate βL per host , removing eggs from the environment at rate βNL , where β represents an ingestion uptake rate . All biological processes occurring during environmental stages are considered to be affected by seasonal factors; parameters for egg maturation time ( τ2 ) , egg mortality ( γ1 ) and infective stage mortality ( γ2 ) are linked to temperature through experimental data , whereas transmission ( β ) is taken to vary with rainfall . Values of model parameters are given in Table 1 . To form an evidence-base for relationships between biological model parameters and temperature we have drawn on three different experimental studies considering A . suum eggs . Two of these studies have been used to parameterise the average time taken for eggs to mature into infective larvae ( τ2 ) across temperatures ranging from 5-35°C [4 , 6] . The third study was used for seasonal parameterisation of the immature egg and infective larval death rates ( γ1 and γ2 ) , with a temperature range of 15-35°C [5] . The first study [4] investigated the rate of development to infectivity of a suspension of A . suum eggs in flasks placed inside a pig barn in Saskatchewan , western Canada . Recorded temperatures in the barn ranged from 16 . 8-25 . 5°C and increased rates of maturation were seen at higher temperatures; it took an between 21-28 days to observe development for temperatures above 23 . 5°C , whereas a development time of 77-84 days was recorded for a mean barn temperature of 16 . 8°C . This data was used as the main basis for the relationship between temperature and egg maturation time ( τ2 ) . The second study [6] recorded the developmental stages of eggs in a coarse sand medium in an environmental chamber with 50% humidity under three temperature conditions: 5°C , 25°C and 30°C . As the humidity was maintained it can be assumed that this was not a limiting factor in development , giving temperature as the sole determinant . No development was observed at 5°C in the first month , with only marginal development being recorded across the three-month time-span of the study; no eggs reached infectivity . At 25°C and 35°C it took 19 and 17 days respectively for eggs to display successful embryonation . This data was used to extend the previous dataset to consider a wider range of temperatures in fitting τ2 . For the two external stage death rates a third study was used [5] that considered larval viability post-development and larval death rate across a temperature range of 16-34°C ± 1 . Eggs were incubated in flasks containing a H2SO4 solution so moisture is also assumed to be sufficient for development and survival . Higher temperatures recorded lower viability and faster time to 90% mortality; larvae were observed as living for up to 150 days at temperatures of around 20°C , but above 25°C this quickly drops to below 50 days and above 30°C larvae survived for fewer than 10 days . The study also considered development rate , but recorded the time until 90% of the eggs had reached maturity rather than the average; this was used to provide a qualitative validation of the fitted relationship for τ2 but not considered for fitting purposes . All seasonal egg relationships were fitted using fminsearch in Matlab 0 . 0 . 21 to minimise the squared error between the model and the data . An exponential decay curve was fitted to the maturation time data as a function of temperature; limits of function parameters give τ2 bounded below by a non-zero limit and the exponential relationship reflects the assumption that development occurs either very slowly or not at all for low temperatures [6] . The proportion of eggs successfully reaching maturity ( d2 ) was fitted to a quartic relationship for higher temperatures and capped at a fitted maximum for lower temperatures . Immature egg death rates ( γ1 ) were then calculated as values which would give the associated survival proportions; γ 1 = ( 1 d 2 - 1 ) / τ 2 . Larval death rates ( γ2 ) were derived by solving a simple differential relationship to get γ2 = −ln ( 0 . 1 ) /m where m is the time taken to achieve 90% mortality , which was fitted to an inverse tangent relationship with temperature . Records of mean monthly temperature ( °C ) and rainfall ( mm ) relevant to the dates and settings we chose to investigate are taken from web archives [37 , 38] and used to fit setting-specific functions . The main requirement of these functions is annual periodicity , hence a sinusoidal function provides the best approximation . The first dataset used to fit and validate the model originates from a field study conducted between April 1977 and September 1978 in Gyeonggi Province , South Korea [16] . The study was conducted across six hamlets ( labeled A-F ) , each consisting of approximately 100 inhabitants , that were considered far enough apart to have independent transmission . Three rounds of biannual testing and chemotherapy were applied in each location , with intervention dates offset by a month for each hamlet to monitor different seasonal responses . The drug used was pyrantel pamoate . The second dataset , used to investigate an alternative setting , is taken from a field study based in Osun State , Nigeria , between 2006 and 2007 [39] . The study followed two groups of 194 children , aged 12-60 months , across a period of 14 months . The treatment group received albendazole every 4 months for a year , with a follow-up assessment at 14 months; the control group received no treatment but prevalence was measured at the same intervals . The model was coded and run using Matlab R2015b , with the function ode45 used to compute numerical solutions to the differential equations . For each simulation the model was run for a 30 year period with a time step of 0 . 5 days to equilibrate the initial conditions before any intervention strategy was applied . Administration of anthelmintic drugs was implemented as a proportional reduction in mean worm burden; this proportion depended on efficacy , taken from the literature , and coverage , taken from the data . Treatment using albendazole was assumed to have an efficacy of 88% [40]; pyrantel pamoate is taken to have the same efficacy [40] , although this is likely to be a conservative estimate [41] . In both settings transmission rate ( β ) was originally considered as an inverse tangent function of rainfall; three parameters are taken to describe the magnitude ( β0 ) , slope gradient ( a1 ) , and horizontal shift ( a2 ) of the function . The slope gradient and shift are permitted to take a range of values to allow for either a positive or negative relationship to reflect conflicting views in the literature and as it is possible that this will change between settings due to influence from human behavioural characteristics , such as the effects of rice planting during the rainy season . Parameters were fitted to the epidemiological datasets using approximate Bayesian computation ( ABC ) , followed by a regression-based conditional density estimation method [42] . Uniform priors for each of the fitted parameters were defined and simulations were run using values sampled from these distributions . Simulation outputs were compared to prevalence data , filtered , and then linear regression is used to correct model inputs , resulting in a posterior distribution . Simulation outputs were filtered ( keeping 1000 of 75K realisations ) to maximise the binomial log-likelihood and weights were calculated using the squared errors between the model output and the data . Model outcomes were obtained from 1000 model runs sampling parameters from the posterior distributions and credible intervals were calculated by taking the 2 . 5% and 97 . 5% quantiles of the outputs . For the South Korean dataset villages A-C were used for fitting and then model outputs for villages D-F were compared to the other half of the data for validation . Prevalence was calculated from mean worm burden using a standard negative binomial relationship [34] , with aggregation parameter k = 0 . 45 [43 , 44] , to capture the expected heterogeneity of infection intensity . For the Nigerian dataset the model was fitted to the first four data points for each group and then the model predictions were compared to the fifth data point in each case . Due to results displaying a very low overall impact of rainfall on transmission the model was also fitted assuming a constant transmission rate , β , and Akaike Information Criterion ( AIC ) values were used for model selection . A negative binomial relationship was also assumed between prevalence and mean worm burden , but in this case the aggregation parameter was also fitted to the data .
The experimental data for all three environmental egg parameters showed strong dependence on temperature , as seen in Fig 2 . The biggest effects are seen in maturation for temperatures below 20°C , for infective stage mortality above 25°C . The proportion of eggs that develop into viable larvae is mostly constant unless temperatures reach above 30°C , which is only relevant in some climates . All fitted seasonal relationships can be seen in Table 2 . Standard transmission models for human ascariasis would expect maturation time to be in the range of 10 to 30 days and a free living infective stage life expectancy of 28 to 84 days [34] . The fitted relationships fall in the 10-30 day range for temperatures above approximately 22 . 5°C , but exhibit a dramatic increase for lower temperatures . For mid-range temperatures the model predicts time to 90% mortality for infective stages to be between 40 and 120 days , which equates to a life expectancy of 17-52 days and falls within the expected range . The fitted parameters for transmission rate in South Korea are: β0 = 3 . 30 ( 2 . 89 , 3 . 82 ) × 10−9; a1 = −3 . 46 ( −3 . 97 , −3 . 12 ) ; a2 = −66 . 8 ( −85 . 3 , 56 . 0 ) ; such that the transmission rate β = a 1 π ( arctan ( a 2 R + a 3 ) + π / 2 ) . This reflects an inverse relationship between rainfall and transmission , perhaps due to high rainfall resulting in infective stages being washed away from areas where uptake is likely to occur . Fig 3 shows both the fitting and validating outcomes for the South Korean dataset . There is excellent agreement between the model and the data used for fitting from villages A-C . Model outcomes for villages D and E are seen to provide a reasonable fit to the data when considering overlap between the respective 95% confidence intervals , with some discrepancy between the model and the data for village F . ABC results for fitting the model to the Nigerian data returned posteriors that allowed for a range of marginal positive and negative relationships between rainfall and transmission , indicating a lack of evidence to support this element of model structure , unlike the significant relationship found for the South Korean data . Comparing the Akaike Information Criterion of this model ( Arain ) to that of the reduced model ( Aβ ) , considering constant transmission , leads us to reject the combined model with rainfall in favor of one relying only on temperature . ( Arain = 434 . 6; Aβ = 429 . 6; Aβ < Arain; the relative likelihood of the rainfall model is 0 . 082 . ) This implies that we do not have enough evidence to suggest rainfall is a significant predictor of disease in Nigeria . A transmission rate of β = 7 . 93e − 10 ( 7 . 86e − 10 , 8 . 00e − 10 ) and a parasite density aggregation of k = 0 . 16 gives the best fit to the data . Fitting outcomes for the constant transmission model capture the overall magnitude and trend of the data in both cases ( see Fig 4 ) , with reasonable agreement between the model and testing data points ( August 2007 ) . The model appears unable to capture the observed peak in cases seen across both groups during February 2007 , which suggests that this increase was driven by additional factors; it is possible that sampling biases caused by behavioural change among the target population could influence such a peak . Investigating model outcomes for the South Korean setting shows that seasonal timing of MDA could result in a 74 . 5% difference in the number of days the average individual is infected with worms ( mean worm days ) across the 12 months following cessation of MDA; the best and worst case scenarios are March and June respectively . This represents a significant improvement in worm burden across the population , which could be expected to link with similar decreases in morbidity and infection intensity . Similar improvements for prevalence and levels of infectious larvae in environment are detailed in Table 3 . It is also interesting to note that whilst the seasonal trend is much more noticeable in the external larval population than the mean worm burden , there is still potential for a large seasonal impact on intervention ( see Fig 5 ) . In South Korea June represents a peak in levels of infectious larvae and the beginning of an uptake in transmission across the following months ( see Supporting Information for estimated seasonal transmission levels ) , causing faster reinfection . Bringing down prevalence through MDA also results in low egg output until new adult worm infections have developed ( approximately 2-3 months ) . As larval numbers will be naturally declining in this period it is expected that artificially reducing egg output through mass treatment will have a less marked effect on the overall population . Contrastingly , in March , infectious larval counts are close to an annual minimum and transmission is on the decline . As the temperature picks up through April and May the larval population should experience a sharp increase , hence treating at this time is likely to limit the resulting peak and dampen future reinfection potential . In Nigeria there is still a seasonal peak in environmental levels of infective larvae , but the model does not indicate much difference between MDA outcomes for treatment at differing times of the year ( Fig 5 , bottom right ) . This comes partially from the lack of rainfall dependence , but also due to the narrow temperature range in the region; for temperatures above 25°C both the egg maturation and larval death rates show very little variation , resulting in a reduced seasonal effect—see Supporting Information for estimated parameter values across the year . In both cases the best time for annual treatment is predicted to occur just before the main upswing in infective larvae , the worst time coinciding with the peak . Bringing down infection levels whilst larval numbers are low starves the larval population , causing larger reductions in future infection levels; decreased transmission due to high rainfall in the summer months in South Korea exaggerates this effect . A deterministic macroparasite model has been used to investigate known relationships between temperature , rainfall and A . lumbricoides transmission . Model parameters were fitted to egg data from lab experiments , as well as prevalence data for two settings ( South Korea 1977-78 , Osun Province , Nigeria 2006-07 ) and used to predict the impact of these relationships on control strategies . Our results show that there could be large undetected fluctuations in the infective larval population , impacting transmission , without these effects being necessarily evident through untargeted surveys of human infection . In South Korea fitting resulted in a negative relationship between rainfall and transmission , with the higher rainfall in the summer months causing a steep decline in transmission rate . The temperate South Korean climate is expected to provide sufficient soil moisture for year-round egg development so it is plausible that low rainfall doesn’t negatively impact the larval population . A transmission decrease due to high rainfall could be explained by the possibility eggs and larvae are being washed away through drainage systems , reducing host exposure to infection . Osun State is located in South-Western Nigeria , where rainfall is abundant across the year; there is no dry season , as experienced by the Northern areas of the country . The lack of dependency on rainfall displayed by fitting the model to data from this region indicates that the factors influencing disease dynamics differ from those in South Korea . The infection data is not seasonally structured , and hence gives only partial information on the seasonal trends , but the peak of infection in February does imply that there could be an additional level of seasonal variation that is not captured by the model . This could be indicative of seasonal changes in population behaviour or eating habits , or other climatic factors such as humidity and soil water-content . The model implies that optimal timing for MDA could coincide with minima in the environmental larval population , with the best treatment time predicted to be just preceding the annual upswing . These results agree with veterinary practices that advise treatment coinciding with hostile environmental conditions for the free-living stages , but we would expect a similar need for caution in this approach due to the potential for selecting for anthelmintic resistance [45] . For South Korea the much wider temperature range , as well as the inclusion of rainfall-influenced transmission in the model , led a predicted comparative decrease of 74 . 3% in mean worm burden between the best and worst MDA timing . In comparison , the model predicted only a 12 . 8% decrease for Nigeria . The climate data used was taken from as close a geographical location as possible to each study , although the monthly temperature averages used to fit these relationships will undoubtedly conceal daily fluctuations that would be expected to result in a more variable seasonal trend . Analysis of these two contrasting settings demonstrates that the importance of seasonal factors for A . lumbricoides control is expected to vary dramatically between different locations , depending on local climatic and transmission patterns . It is possible that different efficacies of treatment could lead to differences in the optimal time of year for treatment , or that changing the time of year between treatments could be beneficial in some settings . Frequency and number of MDA rounds could also impact our results , but benefits from treating at a seasonally optimal time of year are expected to be cumulative . In temperate climates , like South Korea , high ranges of temperatures may allow for significant fluctuations in larval stage development across the year and could lead to important knock-on effects for MDA programs . Although the consistent temperature pattern in Nigeria results in low predicted seasonal differences and the data presented here shows no evidence for rainfall-dependence , it is possible that rainfall could still play an important role in other settings . Although current results are subject to further evidence , we can still use the findings to gain insight into the types of settings where we might expect seasonal effects that have the potential to impact the efficacy of MDA programs . For example , the DeWorm3 trials , which aim to test the feasibility of interrupting the transmission of STH using intensified MDA programs , are based in three countries with heterogeneous weather profiles: Benin , India and Malawi [2] . In Benin the temperature range ( monthly averages of 25-30°C [46] ) is narrower than that of Nigeria , so one would expect any seasonal drivers to be behavioural or rainfall-related . Temperatures are similarly high in Vellore , India , ( monthly averages of 23-33°C [47] ) but with a steeper drop off in the cooler months that may introduce more seasonal variation . The third setting , Malawi , exhibits a fairly narrow but much lower temperature profile ( monthly averages of 17-24°C [48] ) and , depending on rainfall effects , this is where we would expect seasonal MDA to have the highest impact due to the steep increase in maturation time as temperature drops under 20°C . Therefore in this setting it would be prudent to carefully consider the implications of seasonally-timed intensified MDA , as model results suggest that treatment during the cooler months could deliver maximum impact on A . lumbricoides transmission . All results are subject to uncertainty , through the Bayesian fitting framework , and under the assumptions made during model construction and selection . In addition , the egg survival data used to fit the model originates from experiments on Ascaris suum life stages; there may still be some variation that has been unaccounted for , although previous studies have shown strong parallels between A . suum and A . lumbricoides eggs [7] . Preferred epidemiological data would include more frequent measurements , with treatment at different times of the year in parallel communities across at least four years to provide greater insight into the long term infection dynamics . The model succeeds in qualitatively describing the biological components of the system and exhibits a good fit to both datasets , but caution must still be taken when interpreting predictions . Although the model is adapted from a well-established literature base there are still some limitations . For simplicity of calculation the helminth sex ratio within a host is not considered; infections consisting of only male or only female parasites should not result in any egg output . The assumed negative binomial relationship between mean worm burden and prevalence is also an approximation and not a true conversion . Additionally it could be worth considering the uncertainty around where transmission occurs; if infection is driven by hot-spots , such as community latrines , then these may have their own micro-climate that is less affected by the environmental conditions . Taking seasonality into account when planning control programs can also be difficult; even in settings with clear seasonal trends there are likely to be additional complications when determining and successfully executing the optimal treatment timing . For example , the presence of other parasitic diseases in the human population could impact MDA outcomes and interfere with control measures; treatment often targets multiple STH infections and the best timing for one species may not be ideal for another . In addition , the logistics of treating at a particular time of year may be disproportionately costly or difficult for the benefits gained; it may be much easier to treat at particular times of year and moving MDA outside of these windows could result in lower coverage and hence worse overall outcomes . If achievable , timing treatment to maxmise impact now may create future problems further down the line; veterinary experience shows that timing MDA during low periods of larval density in the environment can magnify the risk of drug resistance by imposing additional selection pressures on the system . Although anthelmintic resistance has not been definitively identified using currently available tools for human STH infections , it is still important to be cautious of any action that may encourage resistance to spread . Any seasonal recommendations for treatment timing should therefore be considered alongside the potential resistance development risk and further analysis would need to be done to inform any actions taken . Nonetheless , our results suggest that variation in egg survival and maturation could be exploited to maximise the impact of MDA . Practically , we face the challenges of feasibility , caused by factors such as school term times and potential seasonal accessibility in hard-to-reach areas , but optimising treatment timing may be worth considering in some areas . Even though the evidence base in humans is weak there is enough grounds , combined with the depth of veterinary literature suggesting significant advantages to seasonally targeted anthelmintic therapy , to warrant further investigation .
|
Soil-transmitted helminth infections affect 1 . 5 billion people worldwide and mass drug adminstration ( MDA ) is one of the key public health measures for reducing the burden of infection . A number of experimental studies have demonstrated links between temperature and the dynamics of Ascaris spp . eggs , which are exposed to the environment during maturation . Field studies in a number of countries , including Sri Lanka , Saudi Arabia and South Korea have demonstrated seasonal variation in reinfection rates; in Sri Lanka significant correlations were found between reinfection and both temperature and the number of wet days recorded . The impact of these variations on transmission , and therefore control , are unknown . Using a deterministic macroparasite model we demonstrate that seasonal changes in maturation and death rates of free-living stages , as well as transmission , could result in large seasonal swings in the infective larval population and mean worm burden in the human population . Although the seasonal effects are likely to vary between species of STH due to disparate life-cycles , this could have large knock-on consequences for timings of MDA programs and expected public health outcomes for A . lumbricoides infections in certain settings , indicating the need for further investigation in this area .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"and",
"discussion"
] |
[
"death",
"rates",
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"helminths",
"geographical",
"locations",
"parasitic",
"diseases",
"animals",
"parasitology",
"seasons",
"developmental",
"biology",
"ascaris",
"ascaris",
"lumbricoides",
"population",
"biology",
"africa",
"life",
"cycles",
"nigeria",
"people",
"and",
"places",
"population",
"metrics",
"eukaryota",
"asia",
"south",
"korea",
"earth",
"sciences",
"nematoda",
"biology",
"and",
"life",
"sciences",
"larvae",
"organisms",
"parasitic",
"life",
"cycles"
] |
2018
|
Seasonally timed treatment programs for Ascaris lumbricoides to increase impact—An investigation using mathematical models
|
Cell water permeability and cell wall properties are critical to survival of plant cells during freezing , however the underlying molecular mechanisms remain elusive . Here , we report that a specifically cold-induced nuclear protein , Tolerant to Chilling and Freezing 1 ( TCF1 ) , interacts with histones H3 and H4 and associates with chromatin containing a target gene , BLUE-COPPER-BINDING PROTEIN ( BCB ) , encoding a glycosylphosphatidylinositol-anchored protein that regulates lignin biosynthesis . Loss of TCF1 function leads to reduced BCB transcription through affecting H3K4me2 and H3K27me3 levels within the BCB gene , resulting in reduced lignin content and enhanced freezing tolerance . Furthermore , plants with knocked-down BCB expression ( amiRNA-BCB ) under cold acclimation had reduced lignin accumulation and increased freezing tolerance . The pal1pal2 double mutant ( lignin content reduced by 30% compared with WT ) also showed the freezing tolerant phenotype , and TCF1 and BCB act upstream of PALs to regulate lignin content . In addition , TCF1 acts independently of the CBF ( C-repeat binding factor ) pathway . Our findings delineate a novel molecular pathway linking the TCF1-mediated cold-specific transcriptional program to lignin biosynthesis , thus achieving cell wall remodeling with increased freezing tolerance .
Freezing temperature is an important environmental factor that determines the natural geographical distribution of plants and limits crop productivity [1] . Sudden exposure to freezing temperature causes intracellular freezing , membrane damage and cell death [1–3] . To better survive freezing low temperature , plants have evolved coping mechanisms through initiating cold acclimation when the temperature gradually drops lower in autumn in nature . Many signal transduction cascades are involved in this physiological adaptation process . In Arabidopsis , expression profiling of cold-treated plants revealed that up to 20% of genes in the genome are regulated by cold . Characterization of a group of cold-regulated ( COR ) genes which are highly induced by cold stress using forward and reverse genetics has led to identification of a key CBF ( C-repeat binding factor , also known as dehydration-responsive element-binding protein 1 or DREB1 ) signaling pathway . CBF transcription factors ( CBF1 , CBF2 , CBF3 ) can activate expression of the COR genes by binding to cis-elements in their promoters and induce cold acclimation and freezing tolerance [4–6] . Several regulators of CBF genes have been identified , such as Inducer of CBF expression 1 ( ICE1 ) , calmodulin binding transcription activator 3 ( CAMTA3 ) , MYB15 and Ethylene Insensitive 3 ( EIN3 ) [7–10] . Most recently , it has been shown that OPEN STOMATA 1 ( OST1 ) , a central component in ABA signaling pathway , plays a crucial role in plant response to cold . OST1 is induced by cold and cold-activated OST1 can interact and phosphorylate ICE1 to enhance the stability of ICE1 , resulting in increased plant tolerance to freezing [11] . However , multiple studies have reported that the CBF signaling pathway is not the sole mechanism modulating plant cold acclimation and cold tolerance , because only 12% of the cold responsive genes are regulated by CBF transcription factors [12] . The prominent example is HOS15 , which regulates freezing tolerance through modification of histone acetylation [13 , 14] . In addition , SFR2 was found to modulate freezing tolerance through lipid remodeling of the outer chloroplast membrane [15] . Most recently , it was found that AtHAP5A modulates freezing stress resistance in Arabidopsis independent of the CBF pathway [16] . Through these signaling pathways , a wide variety of antifreeze/stress-related proteins and compounds are accumulated to minimize intracellular ice formation , to increase tolerance to dehydration caused by water outflow , and to maintain cell membrane stability and integrity that is considered central to the ability of plants to survive freezing [2 , 15 , 17–19] . The plant cell wall is the extracellular matrix consisting of cellulose , hemicellulose and lignin . It plays essential roles in plant growth and adaptive responses to adverse environmental conditions [20–22] . The cell wall integrity ( CWI ) and structures are dynamically regulated during plant development and are capable of being remodeled in response to various environmental stresses [23–26] . Fine-tuning regulation of the proportions and the amounts of each matrix component within the cell wall determines its nature and functions . Remarkably , deposition of lignin , phenylpropanoid polymer , which is highly hydrophobic in the cell wall , determines cell wall stiffness and permeability to water [27–29] . In yeast , the CWI signaling pathway plays a vital role in adjusting the cell wall thickness and composition to environmental cues , in particular freezing temperature and osmotic stress [30 , 31] . In plants , similar processes are employed for controlling cell wall integrity and performance during development , drought and defense [32] , but the precise mechanisms remain unclear . Previous studies have shown that the expression of genes related to cell wall biosynthesis and remodeling is dramatically altered under cold treatment [33] . Using cryo-scanning electron microscopy ( cryo-SEM ) , several studies have revealed that both cell membrane and cell wall properties play equally important roles in cold acclimation and freezing tolerance [18] . Most strikingly , cell wall thickness and rigidity have been linked to dynamic water heterogeneities during cold acclimation and extra-/inter- or intracellular freezing upon freezing/thawing process . Variations in cell wall rigidity and composition in different types of plant tissues and cells ( e . g . xylem , phloem , living fibers and mesophyll cells ) showed altered intracellular freezing , tension-induced cavitation and cell viability during freezing/thawing [34–36] . Thus , resistance to freezing temperatures is dependent on the capacity for water outflow from the cells during cold acclimation and freezing and water reabsorption during thawing , on the capacity to accommodate growth of ice crystals in extra-/intercellular spaces , and on the ability of cell wall elasticity to respond to cellular shrinkage . Investigation of the roles of the cell wall in cold acclimation and freezing/thawing has been limited to correlation of cellular changes in various plant species or different types of tissues and cells using cryo-SEM; mechanistic analysis of the regulatory genes and signaling cascades underlying cell wall mediated water movement and freezing tolerance is lacking . Lignin is a major component of the plant secondary cell wall , and the amounts of lignin are altered after cold treatment in various species [37 , 38] . In the past decades , some genes that regulate lignin biosynthesis have been identified [39–43] . Among them , Phenylalanine ammonia-lyase 1–4 ( PAL; EC 4 . 3 . 1 . 5 ) encoding the enzymes that catalyze the first step in the phenylpropanoid pathway regulate biosynthesis of lignin and secondary metablites ( e . g . flavonoids and salicylic acid ) in Arabidopsis thaliana [39 , 44–48] . Arabidopsis thaliana blue copper binding gene ( BCB ) is another positive regulator of lignin synthesis , and AtBCB overexpression substantially increases lignin content in Arabidopsis roots [49] . It has been shown that PAL1-PAL4 and BCB genes are responsive to a variety of environmental stimuli , including pathogen infection , wounding , nutrient depletion , UV irradiation , and extreme temperature , etc . [49 , 50] , suggesting their roles in plant stress resistance . However , it remains unknown how these genes mediate plant responses to biotic and abiotic stresses . Here we report that Tolerant to Chilling and Freezing 1 ( TCF1 ) , a gene encoding a RCC1 family protein , is required for chromatin based gene regulation of cold responsive genes in a CBF-independent pathway . Importantly , we reveal that lignin content in leaves is directly related to freezing tolerance , and that TCF1 plays a critical role in the adjustment of lignin accumulation through modulation of expression of BCB and downstream effectors PAL1/3/4 genes during cold acclimation and freezing tolerance in Arabidopsis .
To identify the genetic loci that regulate specifically plant cold acclimation and freezing tolerance through chromatin condensation and remodeling , we examined cold responses of the genes encoding Regulator of Chromatin Condensation 1 ( RCC1 ) family proteins from AtGenExpress Data [51] . The gene At3g55580 which is specifically responsive to cold was identified ( S1A Fig ) , and designated Tolerant to Chilling and Freezing1 ( TCF1 ) based on the phenotypes of its mutant . RT-PCR analysis and GUS assay of TCF1pro::GUS lines validated induction of TCF1 expression in response to cold but not to osmotic stress or ABA ( Fig 1A–1D ) . The TCF1 gene encodes a protein containing six predicted tandem RCC1 repeats that shows similarity to RCC1 in yeast and human [52 , 53] ( S1B and S1C Fig ) . To determine whether TCF1 is localized in the nucleus like RCC1 [54 , 55] , we made translational fusions with GFP and expressed them in tcf1-1 plants using the native promoters . Examination of independent transgenic lines revealed that GFP-TCF1 fluorescence was present in the nucleus ( Fig 1E ) , and the level of the fusion protein GFP-TCF1 was also induced by cold ( Fig 1E ) . RCC1 is a guanine nucleotide exchange factor ( GEF ) for the small GTP-binding protein Ran . RCC1 is constitutively localized in the nucleus , binds to chromatin , and generates a Ran-GTP/Ran-GDP gradient across the nuclear envelope that is required both to drive nucleo-cytoplasmic transport and to regulate processes associated with progression of the cell cycle and mitosis [54 , 55] . We then ask whether TCF1 has the similar roles of RCC1 . To address whether TCF1 interacts physically with chromatin , an in vitro assay was performed . The fusion protein GST-TCF1 expressed in E . coli bound strongly to a histone agarose column ( Fig 2A ) . To further investigate which histone TCF1 interacts preferentially with , in vitro translated Myc-tagged TCF1 was pre-incubated with each kind of purified core histone in 20-fold excess followed by incubation with histone-agarose . As shown in Fig 2B , H3 and H4 were the histones that can compete effectively to diminish the binding of Myc-TCF1 to histone-agarose , suggesting high affinity binding between TCF1 and histones H3 and H4 . Further yeast-two-hybrid results confirmed that TCF1 can indeed interact with the specific histone H4 ( HFO2 , at5g59690 ) , but not other tested histone H3 and H4s ( HTR9 , At5g10400; HFO4 , At1g07820 ) ( Fig 2C ) . To exam whether TCF1 has GEF activity , an in vitro assay was performed . The results showed that TCF1 exhibited less than 1% of the GEF activity with human Ran as substrate of that measured for RCC1 ( Fig 2D ) . Also expression of TCF1 did not complement the phenotype of the yeast prp20 mutant lacking yeast RCC1 ( S2 Fig ) . The results indicate that TCF1 is not the ortholog of RCC1 in Arabidopsis , but it does associate with chromatin via its interaction with histones . To verify the function of TCF1 in cold and/or freezing response , tcf1-1 and tcf1-2 were analyzed ( Fig 3A and 3B ) . tcf1-1 with T-DNA insertion at the first exon of TCF1 showed no TCF1 transcript in response to cold treatment ( Fig 3B ) , but tcf1-2 having a T-DNA insertion in the 3’-UTR region of TCF1 ( Fig 3A ) exhibited similar TCF1 expression to that of the wild-type under cold treatment ( Fig 3B ) . Thus , the phenotypic analysis of tcf1-1 was shown thereafter . All of the F1 plants from tcf1-1×wild-type cross were resistant on MS medium containing 5 mg/L bialaphos . The F2 progeny of the selfed F1 plants segregated in a 3: 1 ratio ( From 3130 plants , 2322 conferring resistant to bialaphos compared with 808 plants showing sensitive phenotype , x2 = 1 . 11 < 3 . 841; x2 test with one degree of freedom ) . Analysis of the bialaphos resistance revealed the presence of a single functional T-DNA that is inserted in the genome of the tcf1-1 mutant . The results indicated that the tcf1-1 mutation is recessive in a single nuclear gene . To evaluate the effect of the tcf1-1 mutation on freezing tolerance , we performed the whole-plant freezing test . Without cold acclimation , tcf1-1 shows a slightly higher survival rate than wild-type , but there was no significant difference between wild-type and tcf1-1 ( S3A Fig ) . When the plants were acclimated at 4°C for 7 days , 45 . 2% of the tcf1-1 plants survived freezing temperature as low as -10°C , but only 17 . 4% of the wild-type plants survived the treatment ( Fig 3C and 3D ) . The electrolyte leakage assay confirmed the freezing tolerance of tcf1-1 plants under cold acclimation ( Fig 3E ) . To verify the role of TCF1 in cold acclimation , we generated TCF1 RNA interference ( TCF1-RNAi ) lines and two TCF1-RNAi lines ( TCF1-RNAi-2 and TCF1-RNAi-6 ) with reduced expression of TCF1 were used in freezing response assay ( S3B Fig ) . Without cold acclimation , both TCF1-RNAi lines showed a similar survival rate to the wild-type ( S3C Fig ) . When the plants were acclimated at 4°C for 7 days , these TCF1-RNAi lines also displayed significantly higher survival rates than the wild-type plants at -10°C ( S3D and S3E Fig ) . The percentages of electrolyte leakage of TCF1-RNAi lines were also decreased under cold acclimation ( S3F Fig ) , suggesting a role of TCF1 during cold acclimation and freezing tolerance in Arabidopsis . To further determine that freezing tolerance of tcf1-1 was due to loss of function in TCF1 , we expressed the GFP-CDS of TCF1 under the control of its native promoter in the tcf1-1 background ( Fig 1E ) . Three independent transgenic lines ( tcf1-1TCF1-3 , tcf1-1TCF1-12 and tcf1-1TCF1-13 ) with increased levels of TCF1 showed a cold-sensitive phenotype compared with tcf1-1 mutant and had a similar response to the wild-type under freezing treatment with cold acclimation ( Figs 3F and S4 ) , thereby confirming that expression of TCF1 complemented the freezing tolerance phenotype of the mutant . Because TCF1 associates with chromatin , we questioned whether TCF1 regulates freezing tolerance by modulating expression of CBF1-3 and the targeted genes . However , the expression levels and patterns of CBF1-3 under cold treatment in tcf1-1 did not show any significant difference from that in the wild-type ( Fig 4A–4C and S5A ) . No significant changes in transcript levels of TCF1 in cbf2 ( S5B Fig ) were detected under cold ( Fig 4D ) . In addition , the expression of the CBF regulon genes such as COR15A and COR47 was also not changed in tcf1-1 under cold ( Fig 4E and 4F ) . Notably , TCF1 was not in the list of the CBF1 , CBF2 and CBF3 coregulated genes in the previous study [12] . These results indicate that TCF1 may influence cold/freezing tolerance through a mechanism different from the CBF-COR cascade . Because TCF1 reached the highest level of expression and loss of function in TCF1 enhanced freezing tolerance with a 7-day cold acclimation , the expression profiles of tcf1-1 and the wild-type grown under normal condition with a 7-day cold acclimation and without cold acclimation were compared by the GeneChip Array . Under normal conditions , only 20 genes up-regulated or down-regulated ( > 1 . 5 fold ) in tcf1-1 were identified and they are annotated to encode proteins with diverse cellular functions ( S1 Table ) . Five genes that should be repressed during cold exposure were activated in tcf1-1 , and 12 down-regulated genes are involved in diverse stress response processes . After cold acclimation , expression of 36 genes was varied ( 13 genes up-regulated and 23 down-regulated ) in tcf1-1 ( S2 Table ) . Among them , nineteen genes were cold responsive and are involved in different cell functions , but none of them appears to involve the CBF regulon . The highest and lowest two expression genes in microarray data under cold acclimation were validated by qRT-PCR ( S6 Fig ) . The data supports that TCF1 functions in a novel pathway independent of the CBF-COR cascade . HOS15 has been reported to function in a CBF-independent pathway to regulate cold acclimation and freezing tolerance [14] . HOS15 interacts with H4 and represses RD29A expression by facilitating H4 deacetylation through association with the RD29A promoter . To determine whether TCF1 functions in the same way as HOS15 , RD29A expression was analyzed in tcf1-1 plants under cold treatments . However , there were no changes in RD29A transcript levels detected in tcf1-1 plants under stress treatment ( Fig 4G ) . Importantly , in contrast to hos15 the level of nuclear tetra-acetylated histone H4 in tcf1-1 was similar to that in wild-type plants ( Fig 4H ) . Together with the results of expression of RD29A , which is elevated in hos15 plants in response to cold treatment [14] but not in tcf1-1 ( Fig 4G ) , we conclude that TCF1 functions differently from HOS15 . To investigate whether TCF1 represses or activates the expression of target genes through direct interaction with chromatin , we performed chromatin immunoprecipitation ( ChIP ) assay using TCF1 complementational line ( tcf1-1TCF1-3 ) that were cold-treated for 7 days . Seven genes ( At1g23150 , At1g69120 , At1g75040 , At5g50720 , At5g10760 , At5g20230 and At2g22500 ) , which were responsive to cold in tcf1-1 in the microarray dataset were selected and three pairs of primers covering the entire genomic sequences of the candidate genes were designed . Using ChIP assay with an anti-GFP antibody , we found that among the tested genes , TCF1 was only associated with a chromatin fragment containing the coding region ( BCBc1 , CDS sequence +87 to +347 ) of the ( BLUE-COPPER-BINDING PROTEIN ) BCB gene ( At5g20230 ) , which is cold inducible ( Fig 5A ) , whereas ACT2 whose expression level is not changed in the mutant was not immunoprecipitated ( Fig 5B ) . The results suggest that the association of TCF1 with chromatin at the BCB locus mediates cold-induced and TCF1-regulated expression of BCB gene . To get a better understanding of the relationships between histone modifications and BCB gene expression , we analyzed histone modifications across the BCB locus in non-stressed and stressed wild-type , tcf1-1 and TCF1 complementational lines ( Figs 5C and S7A ) . As histone H3K4me2 has been demonstrated to play widespread roles in activation of gene expression , we first analyzed the H3K4me2 status of the BCB gene . In the wild-type and TCF1 complementation lines , BCB was induced by cold treatment . Accordingly , the level of H3K4me2 increased in the transcribed region of BCB ( BCBc1 fragment ) after cold treatment ( Figs 5C and S7A ) . Under normal condition , tcf1-1 had comparable level of H3K4me2 to that of the wild-type . When exposed to cold for 7 days , tcf1-1 also exhibited an increase in H3K4me2 , but the level of H3K4me2 at the BCB locus was lower than that of the wild-type and TCF1 complementation lines ( Figs 5C and S7A ) . Trimethylation of histone H3 at lysine 27 ( H3K27me3 ) is a histone mark associated with gene silencing . We then analyzed the status of H3K27me3 at the BCB locus ( BCBc1 fragment ) in non-stressed and stressed wild-type , tcf1-1 and tcf1-1TCF1 seedlings . In the non-stressed wild-type and tcf1-1TCF1 lines , the level of H3K27me3 was relatively high . However , the level of H3K27me3 was significantly decreased when the seedlings were treated at low temperature for 7 days ( Figs 5C and S7A ) . In sharp contrast , loss of TCF1 function caused an opposite trend in H3K27me3 . Under normal condition , the tcf1-1 mutant displayed a similar level of H3K27me3 compared with that of the wild-type and TCF1 complementational lines ( Figs 5C and S7A ) . As expected , the level of H3K27me3 at the BCBc1 region of BCB was higher in mutant compared with wild-type when exposed to low temperature for 7 days . We also tested H3K36me3 and H3K9me2 ( for gene repression ) , H3K14ac and H3K9ac ( for gene activation ) and AcH4 ( global histone H4 tetra-acetylation at K5/K8/K12/K16 , which is also associated with gene activation ) levels in the BCBc1 fragment and found there were no significant differences between WT and tcf1-1 plants ( S7B–S7E Fig ) . The results indicate that TCF1 may regulate freezing tolerance of plants through modulation of histone modification and subsequent expression of target genes . To test whether BCB is a functional target of TCF1 in plant response to cold , we used artificial microRNA ( amiRNA ) method to generate the BCB knock-down transgenic plants . Two transgenic lines with similar reduction in BCB expression to tcf1-1 mutant ( named amiR-BCB4-9 and amiR-BCB9-2 ) ( S8 Fig ) were used for phenotypic analysis . Three week-old seedlings of the BCB knock-down lines and the wild type control with cold acclimation were treated at -8°C and -10°C for 2 h and were then grown under normal conditions for 7 days . The results showed that the survival rates of the amiR-BCB transgenic lines were markedly increased compared with that of the wild type ( Figs 6A and 6B and S9A and S9B ) . The electrolyte leakage assay revealed that knock-down of BCB significantly reduced the electrolyte leakage from the treated plant cells ( Fig 6C ) , confirming that BCB gene plays an important role in freezing tolerance of Arabidopsis plants during cold acclimation . However , we noticed that the average survival rate for the amRNAi-BCB lines ( about 35% ) is significantly lower than the tcf1-1 ( 45 . 5% ) under -10°C treatment ( Figs 3D and 6B ) , indicating that in addition to BCB , alteration of other genes may also contribute to the freezing tolerance of tcf1-1 . The immediate question is how BCB regulates plant freezing tolerance . Since a previous study has shown that overexpression of BCB results in increased lignin accumulation in Arabidopsis [49] , we attempted to test whether BCB regulates plant freezing tolerance through modification of lignin content of plants . To this end , we analyzed the lignin levels of 3-week-old seedlings of the amiRBCB transgenic lines and the wild-type that were treated at 4°C for 7 days . As shown in Fig 6D , the lignin content of the wild type seedlings was not affected by cold treatment , by contrast , the lignin levels of the amiRBCB transgenic lines was significantly reduced compared with WT under cold treatment ( Fig 6D ) . The results suggest that BCB is responsible for maintaining steady lignin content in Arabidopsis plants under cold stress conditions . Taken together , these results indicate that TCF1 may modulate freezing tolerance through a BCB-dependent mechanism that positively regulates lignin biosynthesis . To investigated whether the genes responsible for lignin biosynthesis showed any differential expression in tcf1 in response to cold treatment . The transcript levels of PAL genes , which encode isoforms of a key enzyme Phe ammonia lyase involved in lignin biosynthesis were analyzed . In wild-type seedlings , the transcript levels of PAL1 and PAL3 remained not changed after cold treatment , but the PAL2 and PAL4 gene was induced by cold treatment ( Fig 7A and 7B ) . In the tcf1-1 mutant , expression of PAL1-4 was similar to that of wild type under normal condition , but PAL1 , PAL3 and PAL4 exhibited significantly decreased transcript accumulation after cold treatment . PAL2 showed a similar level of transcript in the non-stressed and stressed tcf1-1 mutant to that of the wild-type ( Fig 7A and 7B ) . We also checked the PALs expression in amiRNA-BCB transgenic lines , we found that PAL2 showed a similar level of transcript in the non-stressed and stressed amiRNB-BCB mutant to that of the wild-type , but PAL1 , PAL3 and PAL4 exhibited significantly decreased transcript accumulation after cold treatment in the amiRNB-BCB mutant ( Fig 7C and 7D ) . These results suggest that TCF1 and BCB positively regulates lignin content in rosettes under low temperature and that this may be due to regulation of genes involved in lignin biosynthesis . To test whether TCF1 regulates plant freezing tolerance through modulating lignin biosynthesis , we measured the lignin contents in tcf1 and wild-type plants with cold treatments . As expected , at the end of a 7-day cold treatment , the tcf1 rosettes accumulated a significantly lower level of lignin compared with the wild-type ( Fig 8A ) , suggesting a role of TCF1 in lignin biosynthesis . As the loss of function in TCF1 causes reduced lignin content and increased freezing tolerance , we proposed that under freezing temperature , reduced lignin may protect plant cells from freezing stress . To test this possibility , we analyzed the freezing tolerance of the pal1pal2 ( pal1-2/pal2-2 ) double mutant with or without cold acclimation . Under normal condition , the pal1pal2 rosettes contained the lowest level of lignin reduction compared with the wild-type and tcf1-1 mutant ( Fig 8A ) . The substantial low level of lignin in the pal1pal2 double mutant observed in this study is consistent with previous reports [56] . Based on our hypothesis , the less lignin the plants accumulate , the higher freezing tolerance they have . Similar sized wild-type and pa11pal2 seedlings were subjected to freezing treatment with or without cold acclimation . As expected , the pal1pal2 mutant displayed the highest freezing tolerance ( Fig 8B–8D and S10A and S10B ) . The stronger effect of pal1pal2 mutant on freezing tolerance than tcf1 and wild-type suggests that the reduced lignin level is correlated to plant freezing tolerance .
The present study reports isolation and functional characterization of a novel nuclear protein TCF1 as a determinant of cold acclimation and freezing tolerance . We show that TCF1 specifically activated by cold associates with chromatin and regulates a specific set of genes that are involved in cold acclimation and adaptation to freezing temperature . Importantly , TCF1 regulates lignin accumulation during acclimation/freezing and affects plant freezing tolerance via a CBF independent pathway . Thus , our study not only identifies a novel TCF1-mediated signaling cascade that plays a key role in cold acclimation and freezing tolerance , but also reveals a critical role of cell wall remodeling , in particular lignin homeostasis in cell wall in protecting cells from freezing damage . TCF1 specifically responded to cold stress via both transcript and protein accumulation ( Fig 1 ) , and tcf1-1 plants displayed specific enhancement of freezing tolerance after cold acclimation ( Fig 3C and 3D ) . These results suggest that TCF1 is functions as a negative regulator in cold acclimation and freezing tolerance in Arabidopsis . Recently , extensive attention has been paid to the CBF signaling pathway , however , several studies have shown that the CBF signaling pathway is not the sole mechanism modulating plant cold acclimation and cold tolerance , because at least 28% of the cold-responsive genes were not regulated by the CBFs [12] . TCF1 may act independently of the CBF-COR signaling pathway , because expression of the CBF genes and CBF-regulated genes was unaltered in tcf1-1 ( Fig 4A–4C , S1 and S2 Tables ) , and TCF1 expression remained unchanged in cbf2 ( Fig 4D ) and in the CBFs-overexpression plants [12] . The TCF1-mediated pathway is also distinct from HOS15 , because in tcf1-1 , cold induction of RD29A and the level of nuclear tetra-acetylated histone H4 were not affected ( Fig 4G and 4H ) . Thus , we conclude that TCF1 regulates plant cold acclimation and tolerance through at least one additional regulatory pathway . Further genetic analysis will be necessary to determine the genetic relationship between TCF1 and CBFs under cold acclimation and freezing tolerance . TCF1 belongs to a family of RCC1-like proteins in Arabidopsis . RCC1 functions as the GEF for the small G-protein Ran and is critical for maintaining the RanGTP/RanGDP gradient across nuclear envelope [57] . Our data show that unlike RCC1 , TCF1 had very low Ran-GEF activity ( Figs 2D and S2 ) , indicating that TCF1 is not the ortholog of RCC1 in Arabidopsis . However , TCF1 shares several important features with UVR8 , another RCC1 family protein: very low GEF activity , function in nucleus ( Fig 1E ) and histone/chromatin association ( Fig 2A–2C ) , although UVR8 preferentially interacts with histone H2B [40 , 58] . UVR8 regulates plant responses specifically to UV-B through modifying expression of HY5 [40] . Thus , we hypothesized that TCF1 may regulate cold acclimation and freezing tolerance through a similar regulatory mechanism . Indeed , we found that TCF1 directly interacted with the coding regions of BCB ( Fig 5B ) . Further we show that activity of BCB correlates with the enrichment of the positive mark H3K4me2 and reduction of the repressive mark H3K27me3 as cold acclimation is initiated ( Figs 5C and S7A ) . Most importantly , TCF1 is required to modulate levels of both H3K4me2 and H3K27me3 at the BCB locus and regulate BCB transcription ( Figs 5C and S7A ) . Therefore , H3K4me2 and H3K27me3 appear to synergistically regulate transcription activation of BCB , pointing to a critical role of active and repressive marks in cold acclimation . It is clear that cold activates TCF1 to induce or repress a set of target genes through a chromatin based mechanism . Interestingly , BCB encodes a Blue Copper Binding protein , which is a glycosylphosphatidylinositol-anchored protein ( GAP ) targeted to the cell surface [59 , 60] and seems to be responsible for lignin accumulation and cell wall-based resistance to aluminum and bacteria [49 , 61] . Therefore , our data support the hypothesis that TCF1 regulates cold acclimation and freezing tolerance through modulating BCB to adjust lignin accumulation and consequent cell wall remodeling ( Fig 6 ) . We observed a TCF1 dependent reduction in BCB expression and lignin content in rosette leaves of tcf1-1 during cold acclimation ( Fig 8A ) . Importantly , we found that expression of PAL1 , PAL3 and PAL4 was reduced under cold in tcf1-1 and BCB knock-down transgenic lines ( Fig 7 ) , suggesting that the transcriptional activity of the genes in leaves is also influenced by TCF1 and BCB . Our data reveals that TCF1 directly influences BCB activity and affects PAL1 , PAL3 and PAL4 expression and lignification , although it is unclear whether BCB directly regulates PAL genes and lignification or whether TCF1 mediated alteration of chromatin state during cold acclimation indirectly affects PALs expression . Therefore , our data define TCF1 as a key factor able to bind to chromatin and epigenetically regulate the cold-specific GAP leading to establishment of cold specific transcriptional programs and consequently lignin and extracellular matrix remodeling . Although BCB level was almost the same in tcf1-1 and amiRNA-BCB lines , tcf1-1 showed higher survival rate than amRNAi-BCB lines under freezing treatments ( Figs 3D and 6B ) , pointing out that BCB was not the only gene to involve in TCF1-mediated freezing tolerance . Further identification of the target gene ( s ) of TCF1 will help us to uncover the molecular mechanism underlying TCF1-mediated plant cold acclimation and freezing tolerance . Since UVR8 can also modulate plant response to UV-B through direct interaction with COP1 [62] . We do not exclude the possibility that TCF1 regulates cold acclimation and freezing through association with the key regulator ( s ) in the cold signal transduction pathway . It is well known that lignin fills the spaces in the cell wall to reduce water permeability and increase the stiffness of the cell wall [45 , 50] . Extensive cellular studies have shown that freezing tolerance is directly related to cell permeability and cell wall properties , in particular lignin content , so that water outflows and ice forms in the extracellular spaces without damaging cellular structures [18 , 35 , 63–65] . Although it is still technically difficult to measure the cell wall permeability of plants during cold acclimation , the important role of lignin in plant cold acclimation and freezing tolerance has been well documented . For example , the freezing tolerant Miscanthus contains lower lignin content and higher PAL gene expression than the freezing sensitive ecotype [66] . Therefore , it is conceivable that lignin content is closely related to the cell wall permeability and freezing tolerance . However , the molecular mechanism by which lignin content is regulated during cellular adaptation to low temperature still remains a mystery . Here we demonstrate that during cold acclimation tcf1-1 , amiRNA-BCB and pal1pal2 had increased freezing tolerance which correlates with their reduced lignin contents ( Figs 6 and 8 ) . Therefore , reduction of lignin deposition within the cell wall of the tcf1-1 , amiRNA-BCB and pal1pal2 plants during acclimation and freezing may increase cell wall permeability and protect the cells from freezing damage . Reduction of lignin may also enhance elasticity of the cell wall to increase the capacity to accommodate growth of ice crystals with less damage to both the dehydrated cell and cell wall . Together , our data reveal a novel regulatory mechanism in cold acclimation and freezing tolerance in Arabidopsis that involves chromatin based regulation of lignification and cell wall remodeling . The immediate question is what the biological significance of TCF1 induction is because absence of TCF1 confers freezing tolerance . It is known that cold hardy ( freezing tolerant ) plants frequently employ extracellular freezing to cope with the freezing temperature . Arabidopsis Col-0 is not hardy plant although it has moderate freezing tolerance compared with other ecotypes [67] . It is possible that TCF1-mediated signaling is activated to maintain lignin content of cells that can enhance cell rigidity and reduce cell expansion , which is required for plant growth arrest under low temperature . Thus , it is conceivable that TCF1-mediated signaling modulates plastic development of the plants during cold acclimation , but not freezing tolerance . A mechanistic working model is presented in Fig 9 . The fact that absence of TCF1 enhances freezing tolerance of Col-0 plants suggests that low expression or TCF1 absence may be related to plant freezing tolerance . Thus , our study also identifies a new gene that can be used for genetic improvement of plant freezing tolerance . It will be interesting to see whether TCF1 gene acts differently in cold hardy plants . Given that natural variation in TCF1 may contribute substantially to cold acclimation and freezing tolerance among Arabidopsis accessions , the role of TCF1 expression and its mediated signaling in cold acclimation and freezing tolerance is worthy of special focus . Furthermore , further study will also help to differentiate the molecular mechanisms in plant cold acclimation and freezing tolerance .
A . thaliana ( L . ) Heynh Col-0 and tcf1-1 ( SAIL_97_D05 ) and tcf1-2 ( SAIL_749_A09 ) from ABRC were germinated on MS medium containing 2% sucrose , pH 5 . 7 . Three-week-old tcf1-1 and wild-type plants grown in soil with cold-acclimation ( 4°C for 7 days ) and non-cold acclimation were used for whole plant freezing tolerance tests and an electrolyte leakage assay ( ELA ) as previously described [13] . Briefly , the plants were grown in soil under a long-day photoperiod ( 16 h light/8 h dark ) for 3 weeks in a growth chamber ( 22°C ) , and then placed in a temperature chamber ( ZSP-A0160 , ZHCHENG ) at -6°C , -8°C or -10°C for 2 hours . Freezing tolerance was determined as the capacity of plants to resume growth after 7 days under control conditions . For ELA , the rosettes were placed in a temperature chamber starting at 0°C . The temperature was reduced by 2°C after 30 min and maintained for 1 h . Then an identical timing sequence was used for successive 2°C decreases until -10°C was reached . The percentage of EL was calculated as the percentage of the conductivity before autoclaving over that after autoclaving . Total RNA was extracted from plants’ leaves by using the TRizol Reagent . First-strand cDNA synthesis was performed according to standard procedures using reverse transcriptase ( Promega , 18064–014 ) following the manufacturer’s instructions . Semi-quantitative RT-PCR and qRT-PCR were done as described previously [68] . A positive control was provided by a parallel analysis based on the ACT2 gene , and three independent replicates were performed per experiment . Gene-specific primers for CBF1 , CBF2 , CBF3 etc are shown in S3 Table . For Microarray analysis , 3-week-old plants were treated with or without low temperature at 4°C for 7 day , 2 μg of total RNA was used to produce cyanine dye-tagged cRNA ( cy5-WT , cy3-tcf1 ) and was hybridized to JingXin Array ( CapitalBio Company ) containing 29K Arabidopsis transcripts . Three biological replicates with 6 Microarray slides were used to check differential expression genes . Data from the GeneChip arrays were scanned on a GeneChip Scanner 3000 and analyzed using GeneChip Operating software ( GCOS 1 . 4 ) . The Significant Analysis of Microarray software ( SAM ) was used to identify significantly differentially expressed genes between tcf1 and WT groups . Different genes were determined to be significantly differentially expressed with a selection threshold of false discovery rate , FDR = 5% and fold change > 1 . 5 in the SAM output result . The Microarray data had submitted to GEO ( Accession number: GSE70682 ) at NCBI website . For histochemical analysis of TCF1 expression , a genomic fragment including 1 , 337 bp upstream of the translation initiation codon was amplified by PCR and cloned into the binary vector pCAMBIA1391 between the HindIII and BamHI sites . β-glucuronidase ( GUS ) activity was assayed as previously described [68] . To make the TCF1pro::GFP-TCF1 fusion , the TCF1 CDS was first amplified by PCR and cloned in-frame into the binary vector pEZR ( K ) -LC [40] between the EcoRI and SalI sites , and then the 35S promoter was replaced by the 1 , 337 bp TCF1 promoter between the HindIII and SacI sites to generate TCF1pro::GFP-TCF1 . The construct was introduced into wild-type and/or tcf1-1 plants through A . tumefaciens-mediated transformation . At least three independent homozygous T3 lines were tested for TCF1pro::GUS expression , protein subcellular localization , and gene expression analysis . The TCF1 coding region was amplified by PCR and cloned in-frame between the EcoRI and BamHI sites of pGBKT7 . HTA1 [histone H2A ( AT5g54640 ) ] and HTR9 [histone H3 ( AT5g10400 ) ] coding regions were amplified and cloned in-frame into pGADT7 between the NdeI and XhoI sites to generate prey constructs . pACT2-HTB1 ( histone H2B ) and pACT2-HFO1 ( histone H4 ) were from Dr . R . A . Bressan at Purdue University . For analysis of specific histone H4 interaction with TCF1 , TCF1 coding region was cloned in-frame between the EcoRI and BamHI sites of pGADT7 , another two histone H4 variants ( At5g59690 and At1g07820 ) with high expression in tcf1-1 were introduced into pGBKT7 at EcoRI and SalI sites . Plasmid DNA of bait and prey constructs was transformed into the S . cerevisiae strain Y190 . Individual transformants were streaked on plates containing a synthetic , minimal ( SD ) medium lacking tryptophan and leucine and grown for 24 h . Yeast cells were transferred onto a filter paper , and β-galactosidase ( β-gal ) filter assays were performed [14] . Nuclei were isolated as described by [14] . Twenty micrograms of nuclear protein and 1 μg of purified core histones from chicken ( Upstate Biotechnology , 13–107 ) were separated by SDS/PAGE and blotted onto a PVDF membrane ( Millipore , GVPPEAC12 ) . Anti-tetra-acetylated-histone H4 ( 1:1 , 000 ) or anti-histone H4 ( 1:100 , 000 ) ( Upstate Biotechnology ) primary antibodies were used to detect acetylated and unacetylated histone H4 , bands were visualized by using the BCIP/NBT Kit ( Invitrogen ) . Data shown are representative of six independent experiments . GEF activity assay was performed as described in [40] . Briefly , RCC1 , TCF1 , and human Ran were expressed in Escherichia coli as fusions with GST . The Ran clone was provided by Dr . Murray Stewart ( Medical Research Council Laboratory for Molecular Biology , Cambridge , U . K . ) . Assays of guanine nucleotide exchange activity were performed by using [3H]GDP to load 30 pmol GST-Ran and subsequent incubation with 0 . 5 nM recombinant RCC1 or TCF1 for 3 min . The exchange activity was calculated as ln ( Ct/C0 ) , where C0 and Ct are radioactive counts at the start and end of the reaction , respectively . The GEF assays were repeated four times . Analysis of TCF1-histone interaction was performed as previously described [40] . Competition assays between TCF1 and histones were conducted as previously described by Cloix and Jenkins [58] . Chromatin was isolated and the chromatin immunoprecipitation assay was carried out [40 , 69] by using an anti-GFP antibody ( Invitrogen A-11122 ) . Before antibody treatment , the samples were precleared with protein A Dynabeads ( Dynal Biotech , Great Neck , NY , 100 . 02 ) . The immunoprecipitated DNA was used in PCR reactions to amplify fragments from the BCB and ACT2 genes , using primers shown in S3 Table . Data shown are representative of three independent experiments . Cell wall fractions were isolated as described by Tokunaga et al [70] . Briefly , three-week-old plants at rosette stage were ground in liquid nitrogen , and then washed with 95% ethanol and ethanol:hexane ( 1:2 , v/v ) in turn . The washed pellet was allowed to air-dry at 70°C overnight . Lignin content was measured according to the method of Fukuda and Komamine [71] with some modifications . Five mg of the air-dried samples suspended in a 1 ml aliquot of 25% acetyl bromide in acetic acid were treated at 70°C for 30 min . After cooling down , 0 . 9 ml of 2 M NaOH , 5 ml of acetic acid , 0 . 1 ml of 7 . 5 M hydroxylamine hydrochloride , and 3 ml of glacial acetic acid were added . The 10 ml samples were centrifuged and the absorbance of the supernatant was measured at 280 nm to determine the lignin content .
|
Cold acclimation is a well-known adaptive process through which plants can dramatically increase their tolerance to freezing temperature . Modifications of cell wall have been recognized as a key characteristic during plant acclimation to low temperature . However , the molecular mechanism responsible for such cellular adaptation still remains a mystery . Here , we report an unexpected regulatory role of TCF1 on lignin content during cold acclimation in Arabidopsis . TCF1 is specifically induced by cold and is required for chromatin based gene regulation of cold responsive genes such as BCB ( a GAP ) that regulates lignin genes . Further evidence shows that reduction in lignin dramatically increases plant freezing tolerance , while lignin maintenance required for cold acclimation is regulated by TCF-mediated signaling . Thus , our study has revealed , for the first time , lignin remodeling as a key function of cold acclimation and freezing tolerance . The findings provide the first direct molecular evidence that freezing tolerance is directly related to cell wall properties during cold acclimation and extra/intercellular freezing upon and freezing/thawing process .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
The Arabidopsis RCC1 Family Protein TCF1 Regulates Freezing Tolerance and Cold Acclimation through Modulating Lignin Biosynthesis
|
APOBEC3G ( A3G ) is a potent antiretroviral deoxycytidine deaminase that , when incorporated into HIV virions , hypermutates nascent viral DNA formed during reverse transcription . HIV Vif counters the effect of A3G by depleting intracellular stores of the enzyme , thereby blocking its virion incorporation . Through pulse-chase analyses , we demonstrate that virion A3G is mainly recruited from the cellular pool of newly synthesized enzyme compared to older “mature” A3G already residing in high-molecular-mass RNA–protein complexes . Virion-incorporated A3G forms a large complex with viral genomic RNA that is clearly distinct from cellular HMM A3G complexes , as revealed by both gel filtration and biochemical fractionation . Unexpectedly , the enzymatic activity of virion-incorporated A3G is lost upon its stable association with HIV RNA . The activity of the latent A3G enzyme is ultimately restored during reverse transcription by the action of HIV RNase H . Degradation of the viral genomic RNA by RNase H not only generates the minus-strand DNA substrate targeted by A3G for hypermutation but also removes the inhibitory RNA bound to A3G , thereby enabling its function as a deoxycytidine deaminase . These findings highlight an unexpected interplay between host and virus where initiation of antiviral enzymatic activity is dependent on the action of an essential viral enzyme .
APOBEC3G ( A3G ) is a highly active antiretroviral deoxycytidine deaminase that greatly impairs HIV spread in cultures of activated CD4 T cells provided the HIV Vif protein is absent [1] . In these activated cells , the antiviral action of A3G involves its effective incorporation into budding virions and subsequent hypermutation of nascent viral DNA formed during the next round of infection [2–6] . Vif has been proposed to block the incorporation of A3G into HIV virions by targeting this enzyme for accelerated degradation in the 26S proteasome [7–12] and partially blocking its de novo synthesis [7 , 13] . A different situation occurs in resting CD4 T cells and likely monocytes , which are not permissive for HIV infection . In these cells , a low-molecular-mass ( LMM ) form of cellular A3G is present , and it functions as a potent postentry restriction factor for HIV by blocking late reverse transcription [14] . This antiviral action of A3G is unchecked by Vif because insufficient quantities of Vif are present in the incoming virions and the virus has not progressed far enough into its life cycle to synthesize new Vif . Thus , the growth of wild-type ( WT ) HIV is effectively restricted in these cells by LMM A3G . Incorporation of A3G into virions budding from HIV-infected CD4 T cells has been proposed to involve assembly with the nucleocapsid ( NC ) component of the Gag polyprotein and/or viral genomic RNA [15–22] . Recent studies with highly divergent Gag proteins [23] or treatment with RNase A [16 , 18 , 19 , 22] suggest that Gag binding may be indirect , involving an RNA intermediate . Following the entry of A3G-containing virions into new target cells , A3G deoxycytidine deaminase activity targets the minus-strand DNA product of reverse transcription , leading to the appearance of deoxyuridines in lieu of deoxycytidines at canonical sites of deamination ( 5′CC; the residue targeted for A3G-mediated deamination is italicized ) [1–6 , 24] . The nontemplated action of various DNA repair enzymes , including uracil N-glycosylase , may mediate DNA strand cleavage [25] , although a recent study suggests that uracil-N-glycosylase 2 is dispensable for the antiviral action of A3G [26] . If plus-strand synthesis proceeds , dA residues are introduced at sites of dC deamination , which results in dG-to-dA hypermutation in the viral coding strand . These mutations may compromise HIV infectivity by altering various viral open reading frames and introducing inappropriate translation termination codons . In contrast to the LMM form of A3G in resting CD4 T cells , A3G in activated CD4 T cells principally resides in high-molecular-mass ( HMM ) RNA–protein complexes [14] . These complexes include Staufen RNA transporting granules and Ro/La ribonucleoprotein ( RNP ) complexes containing Alu and hY retrotransposon RNA [27 , 28] . These complexes lack detectable deoxycytidine deaminase activity in vitro but interrupt Alu retrotransposition by sequestering the retroelement RNA in the cytoplasm away from the requisite nuclear LINE machinery . Treatment of these complexes with RNase A promotes complex disassembly and generates the LMM , enzymatically active form of A3G . Thus , the cellular forms of A3G in resting and activated CD4 T cells are remarkably different . The recruitment of A3G into HMM RNA–protein complexes during the course of T-cell activation likely explains why cellular A3G fails to function as a postentry restriction factor for HIV in these activated cells . The purposes of this study were to analyze the form of A3G that is incorporated into HIVΔVif virions and to assess its enzymatic activity . Since virion A3G is readily able to mediate hypermutation of viral DNA formed during reverse transcription , we anticipated that enzymatically active forms of A3G would predominate in virions . We have found that newly synthesized A3G , not preexisting A3G already assembled into the inactive cellular HMM complexes , is encapsidated into budding virions . We also found that A3G recruited into virions assembles with viral RNA to form a large intravirion A3G complex ( IVAC ) that is enzymatically inactive . Finally , we have demonstrated that the action of viral RNase H during reverse transcription ultimately releases A3G from its state of inhibition , allowing hypermutation of the minus-strand viral DNA . Thus , activation of the enzyme-dependent antiviral action of A3G appears to critically depend on the action of an HIV enzyme , RNase H .
Initially , we sought to identify transfection conditions for the generation of A3G-containing HIV virions that would recapitulate the virion encapsidation of A3G that naturally occurs in T cell lines and primary T cells infected with HIVΔVif viruses . Activated blood-derived primary CD4 T cells or H9 T cells expressing endogenous A3G were spinoculated with HIVΔVif and emergent viruses were harvested 2 d postinfection . In parallel , 293T cells were cotransfected with a fixed dose of proviral expression plasmid DNA and increasing doses of A3G expression plasmid DNA . Virions were similarly collected from the transfected cells 2 d later , after purification of virions by ultracentrifugation through iodixanol cushions . Virion lysates were subjected to immunoblotting to determine the amount of A3G incorporated relative to p24-CA content . These virion preparations were not contaminated with significant cellular material or microvesicles , as determined by immunoblotting with anti–14-3-3γ and anti-CD45 antibodies ( Figure S1 ) . Transfection of increasing amounts of A3G expression plasmid resulted in increasing amounts of A3G virion incorporation relative to p24-Capsid ( CA ) ( Figure 1A and 1B ) . However , when compared to virions produced from infected primary CD4 T cells or H9 T cells , the 293T transfection conditions that best recapitulated the “natural” packaging levels of endogenously expressed A3G were achieved at a plasmid microgram ratio of 2 ( HA-A3G ) to 60 ( pNL4–3ΔVif ) , which equals a molar ratio of 1 ( HA-A3G ) to 12 . 5 ( pNL4–3ΔVif ) . Thus , this condition was used for the production of A3G-containing virions in all the subsequent experiments unless otherwise indicated . To determine if A3G is packaged into the core of HIV virions and to assess the localization of the additional A3G packaging that occurred at higher transfection doses , virions were subjected to biochemical fractionation . The virion envelope was removed by brief solubilization with Triton X-100 , as previously described [29] , yielding separable virion cores containing p24-CA , integrase ( IN ) , reverse transcriptase ( RT ) , and NC ( Figure 1C ) . Fractionation of viruses containing A3G at levels comparable to those of virions budding from primary CD4 T cells revealed that A3G is indeed packaged into virion cores ( Figure 1C ) . However , virions derived from cells expressing higher levels of A3G ( for example , 20 μg of HA-A3G:60 μg of pNL4–3ΔVif ) packaged a similar amount of A3G into virion cores as well as additional A3G that fractionated into the gp41-containing supernatant after Triton X-100 solubilization ( Figure 1A–1C ) . These findings suggest that when A3G is overexpressed in virus-producing cells , considerable amounts of the enzyme reside outside the viral core , likely in the viral matrix region , between the core and outer envelope ( for example , 1:3 μg ratio ) . The appearance of approximately half the virion p24-CA in the Triton X-100 supernatant after solubilization similarly reflects excess Gag packaged into virions that does not contribute to core formation [30] . Thus , A3G may gain access to HIV virion cores through a specific interaction with viral RNA and/or Gag ( NC ) . However , under conditions of overexpression , a lower-affinity interaction , perhaps directly between A3G and Gag , results in the recruitment of additional enzyme into the detergent-sensitive matrix space , into which excess Gag is also packed [30] . Under steady-state conditions in activated CD4 T cells , cellular A3G resides in an HMM RNase A–sensitive complex of 5 to 15 MDa [14 , 27] . The observation that HA-A3G incorporated into virions packages into the virion core suggested several possible cellular sources of the enzyme . The first possibility , albeit unlikely , is that entire 5- to 15-MDa cellular A3G complexes are recruited into the virion core . Alternatively , HIV RNA and Gag may promote release of A3G from the cellular HMM A3G complex , allowing its recruitment into the virion , with or without a limited subset of cellular cofactors , as has been suggested [28] . Finally , newly synthesized LMM A3G not yet assembled with cellular cofactors or RNA may be recruited into the virion through its association with viral RNA and Gag . To determine if newly synthesized A3G , more “mature” A3G , or both serve as cellular reservoirs for virion recruitment , we performed pulse-chase radiolabeling studies . First , the time course for recruitment of newly synthesized HA-A3G into cellular HMM complexes in the absence of proviral gene expression was determined . Cells were pulsed with radiolabel for 10 min , followed by chases of 30 min to 3 h . Size-fractionation ( Figure S2A ) of the pulse and chase lysates identified the presence of pulse-labeled HA-A3G initially in low and intermediate mass fractions ( Figure 2A , t = 0 fractions 6 and 7 ) that was chased into HMM complexes within 30 min ( Figure 2A , A3G in fractions 4 and 5 at 0 . 5 and 1 h ) . The radiolabeled A3G remained stably associated with the cellular HMM complex during longer chase periods ( Figures 2A and S2B ) . Thus , newly synthesized A3G is initially LMM and recruited within 30 min into stable cellular HMM RNA–protein complexes . When HIVΔVif was coexpressed , we observed that the presence of viral RNA and proteins did not alter the ability of newly synthesized LMM A3G to assemble into cellular HMM RNA–protein complexes ( Figure 2B ) . To assess whether newly synthesized or more “mature” preexisting cellular HMM A3G is recruited into virions , we performed similar pulse-chase radiolabeling studies in cells producing HIV . To enhance radiolabeling and detection of intravirion A3G , the length of the pulse was extended from 10 min to 30 min . The longer pulse time did not affect A3G assembly into cellular HMM complexes , in either the absence ( Figure 2C ) or presence ( Figure 2D ) of HIVΔVif expression; however , it did mask the chase of LMM A3G into HMM A3G . Expression of HIVΔVif also did not affect the turnover of radiolabeled A3G ( Figure S2D ) . Both producer cells and their supernatants containing virions were collected and analyzed simultaneously for radiolabeled HA-A3G and p24-CA content . Since all of the virus-containing supernatants were collected at the indicated time points , the radiolabel present in virion p24-CA or HA-A3G reflects labeling events occurring during the discrete intervening time points . In each of the three independent experiments shown , incorporation of radiolabeled p24-CA into virions increased with increasing chase time over the first 1 to 2 h and then declined by 4 h ( Figure 2E , lower panels ) . The incorporation pattern of p24-CA over time is consistent with previous reports [31 , 32] showing increased accumulation of radiolabeled p24 in virions with long cumulative chase times . In contrast , HA-A3G incorporation into virions displayed a sharp spike between 30 and 60 min after the pulse ( Figure 2E , middle panels ) , even though large cellular pools of radiolabeled A3G were present both before and after this time point ( Figure 2E , upper panels ) . Specifically , despite the presence of radiolabeled cellular HA-A3G at the 2- and 4-h time points , these pools of A3G were not effectively incorporated into virions compared to the 1-h time point . The distinct peak of incorporation of radiolabeled A3G at 1 h after pulse also was not due to a relative loss of radiolabeled cellular A3G available for virion incorporation at the later collection times , since normalization by the available radiolabeled pool of cellular enzyme did not alter the distinct early kinetic pattern for A3G incorporation into virions ( Figure S2C ) . These findings indicate that newly synthesized A3G less than 1 . 5 h old is incorporated HIV virions and that older , more “mature” A3G in HMM complexes is apparently less available for virion incorporation during the time course examined here . Using a modified version of these experiments ( infection with HIVΔVif instead of transfection of a proviral DNA plasmid ) and extended chase times , we observed a similar trend of newly synthesized radiolabeled A3G incorporation into virions and low-to-undetectable levels of radiolabeled A3G in virions up to 9 h after the pulse period ( Figure S2E ) . The finding that newly synthesized A3G is packaged into virion cores ( Figure 1C ) coupled with the observation that newly synthesized A3G rapidly forms HMM complexes in cells ( Figure 2A–2D ) led us to next examine whether intravirion A3G resolves as monomers/dimers or instead as a larger complex . We hypothesized that A3G in virions might remain in an enzymatically active LMM form , because it ultimately deaminates the viral minus-strand DNA synthesized during reverse transcription . Lysates derived from virions containing HA-tagged A3G were size-fractionated by fast protein liquid chromatography ( FPLC ) . Each fraction was then analyzed by SDS-PAGE and immunoblotting with anti-HA monoclonal antibodies ( Figure 3A ) . Surprisingly , virion-incorporated HA-A3G was detected almost exclusively in the HMM region , eluting in the void volume of the Superose 6 column ( Figure 3A ) . This result was not due to incomplete lysis of the virion cores since the p24 viral capsid ( p24-CA ) was detected only in the expected LMM fractions . To determine if the IVAC contains an RNA component , virion lysates were treated with RNase A ( Figure 3B ) . Under these conditions , HA-A3G shifted to LMM fractions consistent in size with monomers and/or dimers of the enzyme . Therefore , reminiscent of the cellular forms of A3G present in activated CD4 T cells [14] ( and in HIV-producing cells , as shown below ) , A3G incorporated into virion assembles into large RNA–protein complexes that are distinct from the cellular HMM complexes ( see below ) . As an alternative approach to examine whether A3G in virion cores were indeed freely soluble or associated with other factors , cores purified by solubilization of whole virions ( Figure 1C ) were biochemically disassembled by exposure to a low pH “STE” buffer at 37 °C , as previously described [33] . This treatment resulted in release of p24-CA into the supernatant of a pelletable viral RNP complex consisting of IN , NC , and viral genomic RNA . Under these conditions , RT is more readily released from the RNP upon biochemical fractionation of the cores ( Figure 3C and as previously described [34 , 35] ) . Analysis of A3G-containing virion cores revealed that IVAC A3G cofractionated with the viral RNP proteins ( Figure 3C ) , suggesting a continued association with the viral genomic RNA and/or NC protein . Since virion-derived A3G ultimately exerts deoxycytidine deaminase activity during reverse transcription [1–6] , we considered the possibility that A3G might remain enzymatically active even when bound to RNA in the IVAC . However , in an in vitro deoxycytidine deaminase assay , the IVAC HAA3G exhibited no detectable enzymatic activity ( Figure 4A ) . Like the cellular HMM A3G complex ( Figure 4D and [14] ) , deoxycytidine deaminase activity was readily detected when the virion HA-A3G immunoprecipitates were pretreated with RNase A before assessment ( Figure 4A ) . Analysis of whole-virion lysate similarly showed that RNase A treatment was required for detection of A3G enzymatic activity in vitro ( Figure 4B ) . These findings indicate that HA-A3G is incorporated into virions as an enzymatically latent large RNP complex . Of note , previous reports have observed readily detectable enzymatic activity from A3G-containing virions . We believe this may be due to the presence of additional , noncore LMM A3G packaged into the matrix space of virions upon A3G overexpression in cells ( Figure 1C ) . When virions containing increasing amounts of A3G ( Figure 1A ) were tested for in vitro deaminase activity , substrate was readily deaminated by those virions which contained higher proportions of A3G to p24-CA than is normally packaged by infected CD4 T cells ( Figure 4C ) . In addition , Yu et al . [6] reported that virion-derived A3G was enzymatically active in the absence of RNase treatment . However , in that study , virions were extracted in buffers containing EDTA [6] , an agent that disrupts some RNA–protein complexes [36 , 37] . When we analyzed virion HA-A3G extracted in EDTA-containing buffers , we also detected deoxycytidine deaminase activity , suggesting that this treatment likely activated A3G by promoting its dissociation from inhibitory RNA ( s ) ( Figure S3 ) . Likewise , it has been observed that the addition of salts like magnesium that promote and stabilize RNA tertiary structure enhance the activity of recombinant A3G purified from insect cells while inducing a shift of A3G from large to intermediate-sized complexes [38] . Together these findings support the notion that RNA binding inhibits A3G enzymatic activity , likely by occluding the catalytic site . The reported recruitment of A3G into virions through viral RNA , the cofractionation of intravirion A3G with viral RNP proteins ( Figures 1C and 3C ) , the assembly of A3G into a large RNase A–sensitive complex within virions ( Figure 3A ) , and the RNase A–dependent in vitro enzymatic activity of intravirion A3G ( Figure 4A and 4B ) strongly suggested that HIV genomic RNA may be an important constituent , perhaps even a nucleating factor for IVAC assembly . To test this possibility , we immunoprecipitated A3G from IVAC fractions , purified the RNA , and subjected it to RT-PCR with primers that specifically amplify HIV genomic RNA . Genomic HIV RNA was readily detected in the IVAC as well as in A3G immunopre0cipitates prepared from virus-producing cells ( Figure 5A ) . Next , we compared the FPLC fractionation profile of HIV RNA derived from virions that contained or lacked HA-A3G . In the presence of HA-A3G , HIV RNA was detected in fractions that contained the IVAC , indicative of A3G-dependent assembly of large HIV RNA–protein complexes in virions ( Figure 5B ) . In the absence of HA-A3G , viral RNA was detected in lower fractions 11 through 14 . The fractions were determined to contain full-length genome by the production of PCR products using probes that amplify across various regions of the genome . However , these gel filtration experiments were associated with some RNA fragmentation , particularly in the absence of A3G ( Figure 5B , TAR/Gag amplicons were observed in fractions 11 through 20 ) . We suspect that such fragmentation also occurred in A3G-containing virions but that the RNA fragments continued to resolve in the IVAC fractions through persistent association with A3G . Of note , virion NC , which also binds HIV RNA , was detected in the FPLC fractions that contained IVAC ( Figure 5C ) . Conversely , RT , which is more readily released from HIV RNA upon biochemical manipulation/lysis of virions [34 , 35] ( and Figure 3C ) , did not display a strong shift into the IVAC fractions in the presence of HA-A3G upon gel filtration . As expected , the non–RNP-associated gp41 viral protein resolved independently of A3G . In the absence of A3G , NC resolved in lower FPLC fractions , consistent with pools of protein that either have dissociated from the viral RNP or remain associated with RNA fragments ( possibly caused by the gel filtration conditions , as discussed above ) . In both the absence [14] and presence of HIV gene expression ( Figure 5D ) , cellular A3G resolves as HMM upon gel filtration . Next our studies focused on how the latent deoxycytidine deaminase activity of A3G present in the IVAC is ultimately activated . As shown in Figure 4A and 4B , the simple addition of single-stranded DNA ( ssDNA ) substrate was insufficient for triggering its activity . In view of the effects of RNase A treatment in vitro , we were intrigued by the possibility that viral RNase H enzyme might play a role in the activation of A3G enzymatic activity . RNase H resides near the C-terminus of the large subunit of the p66-p51 RT heterodimer [39 , 40] . The DNA-dependent action of RNase H is required for commencement of second-strand synthesis and concomitantly generates the free , minus-strand ssDNA substrate that is targeted by A3G for deamination [2–6] . We hypothesized that viral RNase H action might not only generate the substrate for A3G-mediated deamination but also reverse the RNA-mediated inhibition of A3G deoxycytidine deaminase activity by degrading the genomic RNA bound to A3G . To examine this possibility , we established in vitro conditions that lead to RNase H activity and assessed the effects of active RNase H on A3G deaminase activity . Both recombinant purified RT and virion-derived RT cleaved an end-labeled RNA oligonucleotide from an RNA–DNA hybrid substrate ( Figure 6A and 6B ) . Importantly , RNase H activity was magnesium dependent [41] , was inhibited by a variety of small molecules including Compound I [42 , 43] , and was compromised by specific mutations within its catalytic domain , for example , E478Q ( [44] and Figure 6A and 6B ) . We used these various properties of RNase H and reagents to probe the potential involvement of RNase H as an activator of latent A3G deoxycytidine deaminase activity within virions . First , we tested whether stimulation of endogenous reverse transcription in virion lysates by the addition of magnesium and deoxynucleotide triphosphates promoted activation of A3G enzymatic activity measured in the deaminase assay . Such treatment effectively induced readily detectable deoxycytidine deaminase activity , suggesting a link between reverse transcription and A3G activation ( Figure 6C ) . Of note , the appearance of deoxycytidine deaminase activity was blocked in a dose-dependent manner by Compound I , an RNase H inhibitor . Importantly , Compound I did not impair the deoxycytidine deaminase activity of A3G induced by prior RNase A treatment ( Figure 6D ) , supporting inhibition of RNase H as the cause of Compound I–mediated inhibition of A3G enzyme activation . Additionally , the introduction of a point mutation in the catalytic core of RNase H ( E478Q ) , which compromised RNase H activity ( Figure 6B ) , also impaired activation of A3G deoxycytidine deaminase activity under conditions permissive for endogenous reverse transcription ( Figure 6E ) . HA-A3G was otherwise equally active upon RNase A treatment of viruses bearing either WT or compromised RNase H domains ( Figure 6F ) . These findings demonstrate that , in addition to generating the substrate for A3G-mediated deamination , HIV-1 RNase H plays a central role in triggering the activity of the latent virion-associated A3G enzyme .
Our observation that A3G packages into HIV virion cores is not unexpected given that RNA and/or the NC region of Gag recruit the enzyme into the virus [15–22] . Further , A3G antiviral activity is ultimately manifested during reverse transcription and therefore proximity to reverse transcription complexes would be anticipated . Indeed , Khan et al . [18] reported core localization of A3G . However , it is clear from the virion fractionations that additional enzyme may gain access to the virion when A3G is overexpressed in virus-producing cells ( Figure 1 ) . This additional A3G is not specifically recruited into virion cores . It has been previously demonstrated that roughly twice as much Gag than that present in the virion core is incorporated into immature particles [30] . Based on the RNase A sensitivity of the interaction between Gag ( NC ) and A3G [16 , 18 , 19 , 22] and the ability of A3G to interact with highly divergent Gag proteins [21 , 23 , 45–47] , it has been suggested that A3G may recognize an NC–RNA interface that promotes virion incorporation [23 , 47] . Alternatively , the conflicting reports regarding A3G virion recruitment by Gag/NC in an RNA-dependent [16 , 18 , 19 , 22] or RNA-independent [15 , 17 , 20 , 21] manner could be a consequence of the relative amounts of additional , non–core packaging occurring under the conditions of assay . The RNase-sensitivity of Gag interaction may be observable only at lower ( endogenous-like levels ) A3G concentrations where affinity is perhaps governed primarily by RNA interactions . At higher expression levels of A3G , a lower affinity but direct interaction with Gag independent of RNA may occur . Whether virion core–incorporated A3G , like NC , coats the viral RNA or whether A3G binding is restricted to certain regions [18] remains to be determined . Differences in the absolute amount of A3G packaged into virions may also contribute to apparent disparities in prior studies of A3G . Within this study , for example , the additional packaging of A3G under conditions of A3G overexpression masked the RNase-dependent “activation” of virion A3G ( Figure 4C ) , highlighting the importance of establishing conditions that closely recapitulate physiological levels of A3G incorporation into virions . Several observations in this study support the notion that A3G incorporated into HIV virion cores is assembled into a large RNA–protein complex that we have termed the IVAC . First , A3G was incorporated into virion cores ( Figure 1C ) , which contain viral RNP complexes consisting of viral genomic RNA , NC , IN , and Vpr . IVAC A3G both coimmunoprecipitated viral genomic RNA ( Figure 5 ) and cofractionated with the virion core proteins ( Figure 3C ) . Additionally , the shift in viral genomic RNA from lower to higher mass FPLC fractions upon HA-A3G expression supports the notion that RNA is critical for virion packaging of A3G and suggests its possible central role in nucleation of the IVAC . Importantly , although the resolving power of our fractionation is currently not able to differentiate the sizes of the cellular HMM A3G complexes and IVAC ( all resolve at or near the void volume of the Superose 6 column ) , these complexes are not identical . For instance , the cellular HMM complexes form in the absence of viral genomic RNA in activated but uninfected CD4 T cells [14] , while IVAC A3G interacts with HIV RNA ( Figure 5 ) . Second , our preliminary results indicate that many of the protein components of the cellular HMM A3G complex [27 , 28 , 48] are not corecruited into HIV virions ( unpublished data and [28] ) . Finally , the activation of IVAC A3G by in vitro endogenous reverse transcription ( Figure 6 ) suggests that viral RNA inhibits IVAC A3G enzymatic activity unless removed by RNase H , a virally encoded enzyme that acts on the RNA component of RNA–DNA hybrids . Of note , the level of A3G activation obtained when endogenous reverse transcription is stimulated ( Figure 6E ) was consistently less robust than the level of enzymatic activity observed when the IVAC was treated with RNase A ( Figure 6F ) . We suspect this finding reflects a more complete clearance of RNA from IVAC A3G by exogenously added RNase A than occurs with RNase H activation under conditions of endogenous reverse transcription . Alternatively , A3G incorporated into HIV virions may bind both HIV RNA and non-HIV RNA ( for example , the tRNA-Lys3 primer ) ; however , since only the viral RNA genome is reverse transcribed , thereby forming a substrate for RNase H activity , only viral RNA-bound A3G may become “activated” during reverse transcription . The pulse-chase radiolabeling studies of A3G in cells revealed that newly synthesized , LMM A3G is rapidly ( within 30 min ) recruited into cellular HMM complexes ( Figure 2A ) and that proviral gene expression has little , if any , effect on A3G assembly into HMM RNA–protein complexes ( Figure 2B ) . Extension of the pulse radiolabeling time did not impede assembly but masked detection of the rapid assembly of newly synthesized LMM A3G into HMM complexes , in both the absence ( Figure 2C ) and the presence ( Figure 2D ) of viral gene expression . Notably , these complexes appear to be stable for at least several hours ( Figure S2B ) . The assembly of intravirion A3G into a large RNP complex could result from recruitment of any of the cellular HMM A3G complexes ( Staufen RNA transporting granules , Ro/La RNPs ) into virions or by HIV RNA extraction of A3G from cellular HMM A3G complexes . Alternatively , newly synthesized A3G not yet assembled into fully mature cellular HMM complexes could bind to HIV RNA , which in turn targets the enzyme for encapsidation into HIVΔVif virions . As noted , A3G assembled into HMM A3G complexes or A3G assembled with HIV RNA and core proteins sieve near the void volume of the Superose 6 columns and thus these different types of complexes cannot be distinguished by FPLC . To investigate whether newly synthesized A3G or older , “mature” A3G already assembled into HMM complexes is recruited into HIV virions , virus-producing cells were subjected to pulse-chase radiolabeling studies . In each of three experiments , we observed the appearance of a peak of radiolabeled A3G in virions at one discrete time point , occurring between 0 . 5 h and 1 h after the pulse ( Figure 2E , middle panels ) . Radiolabeled A3G incorporation into virions decreased dramatically after this peak , despite the persistence of substantial pools of radiolabeled A3G in the producer cells from which the virions were derived ( Figure 2E , top panel ) . The loss of radiolabeled A3G in virions after this peak also could not be explained by a sharp decline of radiolabeled A3G in the producer cells ( Figure S2C ) . These findings suggest that once A3G assembles into the cellular HMM A3G complexes [27 , 28 , 48] , it may no longer serve as a major reservoir of enzyme for virion encapsidation . Pulse radiolabeling of A3G before the peak of viral Gag expression and extension of the collection times up to 9 h after the pulse further confirmed that mature HMM cellular A3G does not form a major pool of enzyme for incorporation into HIVΔVif virions ( Figure S2E ) . Instead , it appears that newly synthesized A3G is preferentially recruited into HIV virions within 1 . 5 h after synthesis ( Figure 2E , middle panels ) . Interestingly , the appearance of radiolabeled HA-A3G in virions appeared to be slightly delayed since samples collected ( 1 ) immediately after the pulse and ( 2 ) in the first 30-min chase contained relatively little radiolabeled A3G compared to virions collected over the second 30-min chase ( Figure 2E , 1-h collection time point ) . This is clearly within the time required for assembly of newly synthesized A3G into cellular HMM complexes ( within 30 min ) . Although virions were budded during the 30-min pulse , newly synthesized Gag did not yet contribute to these virions until the first 30-min chase period ( Figure 2E , lower panels ) . Thus , the recruitment of newly synthesized A3G into virions may be intimately tied to the synthesis and assembly of the viral genome and/or Gag ( NC ) and the budding of these new virions . Indeed , A3G assembles with viral RNA in producer cells ( Figure 5 and [28] ) and maintains this association within virion cores ( Figures 3 and 5 ) . Since newly synthesized A3G assembles into HMM complexes in cells within 30 min in the presence or absence of HIV RNA ( Figure 2A–2D ) , we cannot determine in these experiments whether the newly synthesized A3G ( less than 1 . 5 h old ) recruited into virions represents A3G newly assembled into any one specific cellular complex , a viral-specific HMM complex , or a combination of cellular and viral complexes . However , several other observations in conjunction with these pulse-chase data strongly support a model in which newly synthesized cellular A3G not yet fully assembled into cellular HMM complexes forms the major pool for recruitment into HIVΔVif virions . First , A3G incorporation into virions is mediated by assembly with viral determinants for encapsidation , including the viral RNA genome and/or Gag [15–22] , and virion-incorporated A3G ultimately forms a large RNP complex ( IVAC ) with viral RNA , NC , and IN ( Figures 3 and 5 ) . Thus , virion-bound A3G forms a complex distinct from the cellular HMM complexes , at the very least distinguished by the presence of the viral encapsidation determinants ( viral RNA genome and/or Gag ) . Second , since none of the cellular cofactors identified in the cellular HMM A3G complexes are corecruited with A3G into virions in an A3G-specific manner ( [28] and unpublished data ) , viral determinants for A3G virion incorporation would have to extract A3G out of mature multisubunit complexes if they do serve as a reservoir for virion incorporation . If such a mechanism is employed , it is difficult to explain why A3G incorporation into virions is not also readily detected at much later time points in the pulse-chase radiolabeling studies . Finally , overexpression of A3G in cells leads to packaging of additional amounts of A3G into virions that localize outside of the virion core ( Figure 3C ) , and this form of A3G is enzymatically active in vitro in the absence of addition of RNase A ( Figure 4C ) . Thus , this additional extra-core A3G appears to be the LMM monomer/dimer that forms upon RNase A treatment [14] and could possibly arise from ( 1 ) newly synthesized LMM A3G not yet assembled into HMM complexes or ( 2 ) LMM A3G not assembled into HMM complexes due to saturation of cellular cofactors upon A3G overexpression . We thus favor a model in which , upon translation , newly synthesized LMM A3G assembles with viral RNA and protein factors to gain access to newly assembling virions ( and , in so doing , forms an IVAC-like complex ) . Since viral genomic RNA is subject to cellular processing that may be common to RNA that nucleates the cellular HMM complexes , a subset of common cellular RNA-binding factors can be predicted to be found in both the cellular HMM complexes [27 , 28 , 48] and viral RNP complexes . For example , RNA helicase A , a component of the Staufen-containing HMM A3G complex , has been reported to be packaged into virions [49] , but its incorporation occurs independently of A3G and is unaffected by A3G virion incorporation ( unpublished data ) . However , we cannot completely exclude the possibility that A3G is recruited into virions by viral cofactors from very recently assembled HMM complex ( es ) , as recently suggested [28] . However , because we do not observe virion incorporation from older HMM A3G complexes , we do not favor such a model . One limitation of the pulse-chase studies is that we cannot calculate the percentage of radiolabeled ( newly synthesized ) to unlabeled ( mature ) A3G that is in virions at any of the given collection times . The recruitment of A3G into HIV virions is ultimately detrimental to the virus , underscoring the essential function of the HIV Vif protein in blocking encapsidation of the deaminase . The principal mechanism by which Vif abrogates antiviral A3G activity is believed to involve proteasome-mediated degradation of A3G , most of which is resident in HMM A3G complexes . The observation that newly translated A3G ( less than 1 . 5 h old ) is preferentially recruited into virions ( Figure 2B ) implies that Vif must also effectively target this newly synthesized pool of cellular A3G . Recently , it has been reported that more Vif binds to A3G in the presence of RNase that in its absence [28] , suggesting that LMM A3G unbound to RNA may be a good target for Vif . Our prior studies have shown that Vif expression promotes polyubiquitinylation of A3G that resolves as HMM [14] . Whether Vif activity leads to A3G ubiquitylation before , during , or after the assembly of this enzyme into cellular HMM complexes remains to be determined . Similarly , it remains to be determined whether ubiquitylated A3G resolving into HMM fractions upon FPLC represents the modification of A3G in the recently identified cellular complexes [27 , 28 , 48] or a separate complex of A3G , Vif , Cul5 , SCF , and the proteasome . Perhaps Vif targets ribosome-associated A3G , thus destroying newly synthesized A3G and removing the key pool of enzyme that is selectively incorporated into virions . Such a scenario is consistent with the observation that Vif partially inhibits the synthesis of A3G [7 , 50] . Alternatively , Vif could target A3G bound for virion incorporation by targeting viral RNA–associated A3G . Indeed , Vif has been reported to interact with viral genomic RNA , suggesting a mechanism by which it might preferentially target virion-bound A3G [51–53] . Similarly , the reported association of Vif with the plasma membrane [54 , 55] , the site of virion assembly , could localize this viral protein in proximity to A3G undergoing active encapsidation . While Vif expression ultimately depletes cells of all A3G , others have suggested that such global degradation of the enzyme may not be strictly required for Vif to exert its countereffects on A3G [56 , 57] . Regardless , our findings suggest that an important target for Vif is the newly synthesized pool of A3G , rather than A3G already assembled into cellular HMM complexes . Because A3G can hypermutate the nascent minus-strand DNA of HIV , we found it surprising that intravirion A3G is inactive in in vitro deoxycytidine deaminase assays . Indeed , we observed that the binding of HIV RNA to A3G within virions prevented ssDNA binding and/or occluded the A3G catalytic site ( s ) . As shown in Figure 4A , the addition of free ssDNA substrate to virion A3G proved insufficient to compete RNA binding and/or access the catalytic pocket ( s ) . Rather , the inhibitory RNA had to be removed before A3G enzymatic activation was observed . In view of the emerging findings that A3G can also exert antiviral activity independent of deoxycytidine deamination [14 , 46 , 58–62] , we propose that virion A3G may employ two mechanisms acting sequentially to produce its full antiviral effect . First , the enzymatically latent form of A3G bound to HIV RNA may impair the generation of minus-strand DNA by physically blocking the movement of RT on its viral RNA template . Indeed , short interfering RNA–mediated knockdown of endogenous A3G in resting CD4 T cells enhances the synthesis of late reverse transcription products [14] and the generation of both early and late reverse transcription products is reduced by the presence of A3G in virions [63] . However , because this inhibition is incomplete , minus-strand viral DNA is occasionally generated , setting the stage for the second , enzyme-dependent antiviral action of A3G . During reverse transcription , we now show that RNase H degrades the viral RNA that impairs A3G activity , allowing the enzyme to extensively deaminate the minus-strand DNA . Perhaps incomplete inhibition of reverse transcription by A3G is caused by the occasional ability of RT displace A3G off the RNA template . Although this could result in the generation of enzymatically active A3G , A3G may also be able to rebind the RNA–DNA duplex , reestablishing the inactivated state and dependence upon RNase H for enzymatic activation . Indeed , the lack of A3G activity induced by reverse transcription but under conditions where RNase H activity is inhibited ( Figure 6E ) suggests that if A3G is displaced by RT , it rapidly rebinds inhibitory nucleic acid . These events of initial inhibition and subsequent activation of A3G enzymatic activity by various components of the virus highlight an unexpectedly complex but interesting interplay between HIV and its cellular host . Such a dual strategy for A3G inhibition of retroviral replication could account for its potent antiviral activity and explain reports of both enzyme-dependent and -independent antiviral activities . This model could also explain the conflicting results concerning the ability of A3G to inhibit the replication of hepatitis B virus . In some cells , A3G acts independently of deoxycytidine deamination [46 , 58] , while in others , prominent DNA mutation is evident [64–66] . Cell-type differences in the relative effectiveness of these two sequential antiviral actions of A3G could underlie these findings .
The 293T cells were maintained in DMEM supplemented with 10% FBS ( Gemini Bio-Products , http://www . gembio . com ) . H9 cells were maintained in RPMI supplemented with 10% FBS . Primary CD4 T cells were isolated from fresh human peripheral blood mononuclear cells on CD4 magnetic microbeads ( Miltenyi Biotec , http://www . miltenyibiotec . com ) . The isolated CD4 T cells were then activated by 36-h treatment with PHA ( 5 μg/ml ) followed by 36-h IL-2 treatment ( 20 U/ml; Roche , http://www . roche . com ) in RPMI supplemented with 10% FBS , 100 μg/ml streptomycin , and 100 U/ml penicillin . Virions were generated by calcium phosphate–mediated cotransfection of subconfluent 293T in T175 flasks with a proviral plasmid ( 60 μg ) , pCMV4-HA-A3G vector ( 0 to 20 μg ) , and/or pCMV4-HA ( 0 to 20 μg ) . The medium was changed after 16 h , and the supernatant and cells collected were collected after 48 h . The virus-containing supernatant was clarified by low-speed centrifugation , filtered through a 0 . 22-μm membrane , and sedimented by ultracentrifugation over a 2-ml cushion of 8 . 4% iodixanol at 20 , 000 rpm using an SW28 rotor ( Beckman Coulter , http://www . beckmancoulter . com ) for 2 h at 4 °C . The virus-containing pellet was resuspended in 1 ml of PBS , DNase-treated ( RNase-free; Roche ) , underlaid with a 100-μl cushion of 8 . 4% iodixanol and ultracentrifuged at 20 , 000 rpm in an HFA 22 . 1 rotor ( Heraeus , http://www . thermo . com ) for 1 h at 4 °C . Unless otherwise indicated , 0 . 1 U of RNase A inhibitor ( RNaseOUT; Invitrogen , http://www . invitrogen . com ) was added to virion pellets , which were then immediately lysed or flash-frozen on liquid nitrogen and stored at −80 °C until lysis . The addition of RNaseOUT had no effect on the intrinsic activity of HA-A3G ( Figure S4 ) . Cells were washed with PBS , and the pellet was either immediately lysed or flash-frozen on liquid nitrogen and stored at −80 °C until use . To generate VSV-G–pseudotyped ΔVif virions , 293T cells were cotransfected with expression vectors for the ΔVif provirus and the envelope of VSV-G . At 48 h after transfection , supernatants were cleared by low-speed centrifugation and filtration as described above and then used directly on fresh H9 or primary CD4 T cells . The T cells were spinoculated with the pseudotyped virion-containing supernatant as previously described [67] . Briefly , 0 . 4 × 106 cells/well of a 48-well plate were centrifuged at low speed for 2 h at room temperature with VSV-G–pseudotyped viruses . Cells were then washed five times with cold medium and returned to complete media for an additional 40 h . Supernatants and cells were then collected and processed as described above for the transfected 293T cells . The proviral clone of pNL4–3ΔVif used to generate the HIV-1ΔVif virions has been previously described [68 , 69] . pNL4–3ΔVifH− ( E478Q ) contains a point mutation in the catalytic site of the RNase H domain of RT that compromises RNase H activity . This plasmid was generated by first subcloning the SpeI-EcoR1 Pol-containing restriction fragment of pNL4–3ΔVif into pEF1A . The mutagenesis primer 5′–ACAACAAATCAGAAGACTCAGTTACAAGCAATTCATCTAGC–3′ and its complement ( Operon , http://www . operon . eu . com ) were used to generate the E478Q mutation in the subclone using the QuikChange site-directed mutagenesis kit ( Stratagene , http://www . stratagene . com ) . The mutation was confirmed by DNA sequencing . The pol region in the subclone was then recloned back into pNL4–3ΔVif . The introduction of the E478Q mutation into NL4–3ΔVifH– ( E478Q ) was confirmed by sequencing . pCMV4-HA and pCMV4-HA-A3G [7] expression vectors were cotransfected with pNL4–3ΔVif to generate HIV-1ΔVif virions lacking or containing HA-tagged A3G . Virion cores were obtained using a previously published method [29] . Briefly , virion pellets were resuspended in MOPS Buffer I ( 200 mM NaCl , 100 mM MOPS [pH 7 . 0] ) and Triton X-100 added to a final concentration of 0 . 5% for 2 min at room temperature . The cores were then pelleted from the solubilized enveloped by spinning the samples at 14 , 000g for 8 min at 4 °C . The core pellets were then washed twice with MOPS buffer II ( 100 mM NaCl , 50 mM MOPS [pH 7 . 0] ) . The cores were then either analyzed by immunoblotting or further fractionated to remove the p24-CA shell , as previously described [33] . Briefly , cores were resuspended in STE buffer ( 10 mM Tris [pH 6 . 7] , 1 M NaCl , 0 . 5 mM EDTA ) , incubated at 37 °C for 4 h and subsequently centrifuged at 14 , 000g to pellet the RNP complex . Virions present in the supernatants of 293T cells were transiently transfected with HIV proviral plasmids , and the cells themselves were lysed in ice-cold lysis buffer ( 50 mM HEPES [pH 7 . 4] , 125 mM NaCl , 0 . 2% NP-40 , and 1× EDTA-free protease inhibitor cocktail [Calbiochem/EMD Biosciences , http://www . emdbiosciences . com] ) . Lysates were clarified by sedimentation , quantified with a protein assay ( Bio-Rad , http://www . bio-rad . com ) , and applied to a calibrated Superose 6 HR 10/30 gel filtration column run by an FPLC apparatus ( AKTA; Amersham Biosciences , http://www . amersham . com ) . One column-volume ( 24 ml ) using FPLC running buffer ( 50 mM HEPES [pH 7 . 4] , 125 mM NaCl , 0 . 1% NP-40 , 1 mM dithiothreitol , and 10% glycerol ) was collected in 1-ml aliquots . Equal volumes of collected fractions were either directly run on SDS-PAGE gels or concentrated with YM-3 Microcon filters with a cutoff of 3 , 000 Da ( Millipore , http://www . millipore . com ) before running on SDS-PAGE after normalization for resultant concentrate volume . The size-separated proteins were then transferred to nitrocellulose membranes and immunoblotted . To test nuclease sensitivity , the lysates were pretreated with 50 μg/ml RNase A ( DNase-free; Roche ) and/or 20 to 200 U/ml DNase ( RNase-free; Roche ) for 1 h at 37 °C before gel filtration . Polyclonal antibodies against A3G [7] and Vpr [70] have been previously described . Through the National Institutes of Health AIDS Research and Reference Reagent Program , HIV-1 RT monoclonal antibody ( 8C4 ) was obtained from Dr . Dag E . Helland , polyclonal antiserum to HIV-1 IN ( 757 ) was obtained from Dr . Duane Grandgenett , and HIV-1 gp41 human antibody ( No . 50–69 ) was obtained from Dr . Susan Zolla-Pazner . Anti–NC-p7 antibody was generously provided by Dr . Robert J . Gorelick ( National Cancer Institute , Frederick , Maryland , United States ) . Mouse monoclonal anti-p24 Gag ascites was generously provided by Beckman Coulter . Other antibodies used include polyclonal anti-HA antibody Y11 , monoclonal anti–14-3-3γ antibody C-16 , monoclonal anti-CD45 antibody 2D-1 , and polyclonal anti-GFP antibody ( FL ) ( all Santa Cruz Biotechnology , http://www . scbt . com ) and monoclonal anti-HA antibody HA . 11 unlinked or linked to beads ( Covance , http://www . covance . com ) . Immunoblot analysis of proteins was performed using horseradish-linked secondary antibodies followed by ECL detection ( Pierce Biotechnology , http://www . piercenet . com ) . In Figure 1 , A3G and p24-CA were detected and quantified by using fluorescently linked secondary antibodies ( LI-COR Biosciences , http://www . licor . com ) . Blotted proteins were then detected and quantified using the Odyssey infrared imaging system and software ( LI-COR ) . Four plates of 293T cells were transfected with pCMV4-A3G-HA alone or with pNL4–3ΔVif . After 36 h , the cells were rinsed once and incubated for 1 h with pulse-radiolabeling medium ( DMEM without methionine and cysteine; GIBCO , http://www . invitrogen . com ) plus 10% dialysed FBS ) . The cells were pulse labeled for 10 min with 500 μCi/ml EasyTag XPRESS 35S Protein Labeling Mix ( Perkin Elmer , http://www . perkinelmer . com ) containing radiolabeled methionine and cysteine in fresh pulse-radiolabeling medium . At the end of the pulse-radiolabeling period , the radiolabel was removed and one plate of cells harvested . The remaining radiolabeled samples were incubated with chase medium ( DMEM supplemented with 10% FBS , 4 . 02 mM methionine [20×] , and 3 mM cysteine [15×] ) . Cells were harvested following incubation for 0 . 5 , 1 , or 2 h . Cells pellets were lysed in ELB lysis buffer . Each lysate was size-fractionated on gel filtration columns packed with Sepharose CL-6B beads , which crudely separate HMM from LMM proteins ( Figure S1 ) . For each sample , ten fractions of 300 μl each were collected , and equal volumes of each fraction were immunoprecipitated with anti-HA antibody . The immunoprecipitates were run on SDS-PAGE , and the signal was detected by autoradiography . The signal from each radiolabeled A3G-HA band was quantitated using Scion Image for Windows software ( Version 1 . 62; Scion Corporation , http://www . scioncorp . com ) and divided by the sum of the total signal , in order to assign a relative percent density versus the fraction number for every chase time point sample . For the pulse-chase analysis of virus-producing cells , 293T cells were cotransfected with pNL4–3ΔVif , pCMV4-HA-A3G , and pEGFP-C1 to generate HA-A3G–containing HIV-1ΔVif virions . After 48 h , the medium was changed , and the cells were rinsed and incubated for 1 h with pulse-radiolabeling medium as described above . The cells were then pulse-radiolabeled for 30 min with 125 μCi/ml EasyTag XPRESS 35S Protein Labeling Mix ( Perkin Elmer ) in fresh pulse-radiolabeling medium . At the end of the pulse-radiolabeling period , the radiolabel was removed . Supernatant from the initial pulse-labeled samples ( t = 0 , pulse ) was harvested , and radiolabeled cells were incubated with chase medium ( DMEM supplemented with 10% FBS , 4 . 02 mM methionine [20×] , and 3 mM cysteine [15×] ) for 0 . 5 h . Again , supernatant was collected ( t = 0 . 5 h ) , and the cells were incubated with chase medium for a further 30 min to generate the t = 1 h sample . The process was repeated twice more with an incubation of 1 h and 2 h to generate the t = 2 h and t = 4 h samples . At all time points , a fraction of radiolabeled cells were also collected , washed with PBS , pelleted by centrifugation , flash-frozen on liquid nitrogen , and stored at −80 °C . The virus-containing supernatants were filtered through a 0 . 22-μm membrane , and virions were sedimented by ultracentrifugation over a 2-ml cushion of 8 . 4% iodixanol at 20 , 000 rpm in an SW28 rotor ( Beckman ) at 4 °C . The pellets were resuspended in 1 ml of PBS , underlaid with a 100-μl cushion of 8 . 4% iodixanol , and ultracentrifuged at 20 , 000 rpm in an HFA 22 . 1 rotor ( Hereaus ) for 1 h at 4 °C . The resultant virion pellets were flash-frozen on liquid nitrogen and stored at −80 °C . After virion and cell pellets had been obtained for all time points , the samples were lysed in the lysis buffer described above . Lysates were clarified by sedimentation and quantified with a protein assay ( Bio-Rad ) , and immunoprecipitations were set up at equal protein concentration/volume in the presence of monoclonal anti-p24 ascites or monoclonal anti-HA antibody and incubated for 2 h at 4 °C . The immunoprecipitates were washed once with lysis buffer and subjected to SDS-PAGE . The proteins were transferred to nitrocellulose and immunoblotted for GFP or HA with polyclonal antibodies or for p24 with monoclonal antibody . GFP , HA-A3G , and p24-CA identified by immunoblotting were excised from the membranes and subjected to scintillation analysis . Bands were first identified by immunoblotting since Gag and HA-A3G coimmunoprecipitate with each other [15 , 17 , 20 , 21] and are close in size . Scintillation counts were normalized to the amount of immunoprecipitated material assessed , determined with ImageJ ( http://rsb . info . nih . gov/ij ) . The normalized counts were divided by the sum of the total counts to assign a relative percent density for every sample . No GFP was detected in virions ( unpublished data ) . In an alternate approach ( Figure S2E ) , 293T cells were first transfected with HA-A3G expression vector DNA using Fugene ( Roche ) followed by infection of these cells with VSV-G-pseudotyped NL4–3ΔVif for 12 h . The cells were then pulse-radiolabeled with 125 μCi/ml EasyTag , as described above . Also as described above , after the pulse , cells were chased with cold medium and cells and virions were harvested at 1 , 3 , 5 , and 9 h after the pulse-radiolabeling period . In these experiments , samples were subjected to denaturing lysis ( 50 mM Tris [pH 7 . 5] , 1% SDS , 5 mM dithiothreitol ) followed by anti-HA or anti-p24 immunoprecipitations ( 50 mM Tris [pH 7 . 5] , 250 mM NaCl , 5 mM EDTA , 0 . 5% NP-40 ) and immunoblotting or PhosphorImaging ( Bio-Rad ) , as indicated . Samples for analysis were either ( 1 ) whole virion lysates or ( 2 ) FPLC fractions from cell or virion lysates . FPLC fractions were immunoprecipitated with monoclonal anti-HA antibody to concentrate HA-A3G . In all cases , the amount of HA-A3G in the input samples was confirmed by immunoblotting before analysis . DNA oligonucleotides ( 5′-ATTATTATTATTCCCATTTATTTATTTATTTATGGTGTTTGGTGTGGTTG-3′ ) containing target sites for A3G deamination [italicized] were labeled at the 5′ end with [32P]ATP using T4 polynucleotide kinase ( New England Biolabs , http://www . neb . com ) or with an FITC fluorophore ( Operon ) . Labeled oligonucleotides and input samples were incubated in 20 μl of 50 mM Tris buffer ( pH 7 . 4 ) , with or without RNase A ( 1 μg ) at 37 °C for 3 h unless otherwise indicated . For incubations under conditions stimulating endogenous reverse transcription , KCl ( final concentration , 60 mM ) , MgCl2 ( final concentration , 4 mM ) , and dNTPs ( final concentration , 1 mM ) were added . The RNase H inhibitor Compound I , generously provided by Dr . Daria Hazuda ( Merck ) , was used at a final concentration of 0 . 1 , 1 , 10 , or 100 μM . To terminate the reactions and purify the labeled oligonucleotides , the reactions were subjected to G-25 Mini Quick Spin Columns ( Roche ) . Any uracil bases generated by A3G were converted to abasic sites by treatment of the purified oligonucleotides with 1 U of uracil DNA glycosylase ( New England Biolabs ) for 30 min at 37 °C . After 10 min of heat inactivation at 95 °C , the reactions were subjected to alkaline hydrolysis by the addition of NaOH ( final concentration , 0 . 2 M ) for 10 min at 95 °C . Cleavage products were resolved on 15% PAGE TBE-urea gels ( Bio-Rad ) and visualized with a Personal FX Imager ( Bio-Rad ) , for radiography or fluorescence . Samples for analysis were either virion lysates or recombinant protein . Recombinant HIV-1 RT , WT or E478Q , was generously provided by Dr . Matthias Gotte ( McGill University ) and used at a final concentration of 1 nM . Test substrates included ssDNA , ssRNA , RNA–RNA hybrid , and DNA–RNA hybrid . Unmodified 18-mer PAGE-purified complementary DNA and RNA oligonucleotides were from Operon and are based on the oligonucleotides 18-DAB-DNA and 18-FAM-RNA described by Shaw-Reid et al . [42] . In addition , an unmodified RNA oligonucleotide complementary to 18-FAM-RNA was used to generate the RNA–RNA hybrid ( Operon ) . All oligonucleotides were end-labeled with [32P]ATP using polynucleotide kinase ( New England Biolabs ) . Hybrids were formed by annealing hot oligonucleotide to cold complementary oligonucleotide . Samples were incubated in 20 μl of RNase H buffer ( 50 mM Tris-Cl [pH 8 . 0] , 60 mM KCl ) either with or without 5 mM MgCl2 , as indicated , for 10 min at 37 °C . Compound I was added to a final concentration of 0 . 1 , 1 , 10 , or 100 μM . Radiolabeled substrate ( single-stranded or hybrid ) was added to a final concentration of 100 nM , and the reactions were allowed to continue at 37 °C for 30 min . After the addition of loading dye to stop the reactions , the cleavage products were resolved on 20% PAGE-TBE-urea gels and were visualized with a Personal FX PhosphorImager ( Bio-Rad ) . FPLC samples and immunoprecipitates for RNA analysis were first treated with 20 U of DNase ( RNase-free; Roche ) at 37 °C . The RNA was then extracted with the QiaAmp RNA purification kit ( Qiagen ) according to the manufacturer's instructions . Viral genomic RNA was detected by reverse transcription with a primer complementary to the gag region of HIV-1 ( 5′-TGCTATGTCACTTCCCCTTGG-3′ , generously provided by Jerry Kropp [Gladstone Institute of Virology and Immunology] ) followed by PCR using primers complementary to the R ( F496; nucleotides 496–517 ) and U5 ( R573; nucleotides 552–573 ) or U5 ( F592; nucleotides 592–613 ) and PBS ( R666; nucleotides 645–666 ) regions of HIV-1 . All these primers have been described [14] . In addition , reverse transcription was also performed using an antisense primer complementary to the Vpu region ( 5′-TCATTGCCACTGTCTTCTGCTCT-3′ ) followed by PCR using the Vpr primer and a primer complementary to Pol ( 5′-GTAATATGGGGAAAGACTCCT-3′ ) .
|
APOBEC3G ( A3G ) is a cellular enzyme that promotes DNA mutagenesis and can restrict infection by HIV-1 . However , HIV counters the antiviral effects of A3G through the action of its Vif protein . In the absence of Vif , A3G is effectively incorporated into virions , where it mutagenizes the first DNA copy ( cDNA ) generated during reverse transcription of the viral RNA genome . A3G also appears to be able to inhibit HIV via nonenzymatic mechanisms . A3G and related deoxycytidine deaminases can also inhibit the growth of retroviruses other than HIV and protect the cellular genome from endogenous mobile retroelements . In this study , we analyzed the recruitment and enzymatic activity of A3G incorporated into HIVΔVif virions . Unexpectedly , we found that the binding of A3G to viral genomic RNA led to inactivation of the enzyme . However , latent A3G was ultimately activated through the action of HIV RNase H , which degrades the RNA genome during reverse transcription . These findings highlight an unexpected interplay between a host enzyme and HIV , where the antiviral enzymatic activity of the host factor ( A3G ) is dependent on the action of an essential HIV enzyme ( RNase H ) . The strong interaction with viral RNA also suggests a potential mechanism by which A3G could exert antiviral activity in the absence of enzymatic activity , by physically impeding reverse transcription .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"viruses",
"biochemistry",
"infectious",
"diseases",
"virology",
"in",
"vitro"
] |
2007
|
Newly Synthesized APOBEC3G Is Incorporated into HIV Virions, Inhibited by HIV RNA, and Subsequently Activated by RNase H
|
Myxobacteria are social bacteria that upon starvation form multicellular fruiting bodies whose shape in different species can range from simple mounds to elaborate tree-like structures . The formation of fruiting bodies is a result of collective cell movement on a solid surface . In the course of development , groups of flexible rod-shaped cells form streams and move in circular or spiral patterns to form aggregation centers that can become sites of fruiting body formation . The mechanisms of such cell movement patterns are not well understood . It has been suggested that myxobacterial development depends on short-range contact-mediated interactions between individual cells , i . e . cell aggregation does not require long-range signaling in the population . In this study , by means of a computational mass-spring model , we investigate what types of short-range interactions between cells can result in the formation of streams and circular aggregates during myxobacterial development . We consider short-range head-to-tail guiding between individual cells , whereby movement direction of the head of one cell is affected by the nearby presence of the tail of another cell . We demonstrate that stable streams and circular aggregates can arise only when the trailing cell , in addition to being steered by the tail of the leading cell , is able to speed up to catch up with it . It is suggested that necessary head-to-tail interactions between cells can arise from physical adhesion , response to a diffusible substance or slime extruded by cells , or pulling by motility engine pili . Finally , we consider a case of long-range guiding between cells and show that circular aggregates are able to form without cells increasing speed . These findings present a possibility to discriminate between short-range and long-range guiding mechanisms in myxobacteria by experimentally measuring distribution of cell speeds in circular aggregates .
Myxobacteria are social bacteria that upon starvation form multicellular fruiting bodies whose shape in different species can range from simple mounds to elaborate tree-like structures consisting of 105 − 106 cells [1 , 2] . The development of fruiting bodies is a result of collective movement of flexible rod-shaped cells in close contact with one another on a solid surface . After the movement of cells within the fruiting body has stopped , cells differentiate into dessication-resistant spores . Since collective cell motility during morphogenesis is also common in higher organisms [3] , myxobacteria serve as a relatively simple model organism to study multicellular movement , organization and development . In the course of development of myxobacteria , groups of cells move in circular or spiral patterns to form aggregation centers that can become sites of fruiting body formation [4 , 5] . Such cell aggregates are dynamic , i . e . they can disperse , split , merge with other aggregates , or stabilize and form a fruiting body [6] . Nascent cell aggregates grow as new cells enter in multicellular streams , where cells are aligned and move in concert [4 , 5 , 7] . Remarkably , circular and spiral patterns of cell movement are conspicuous during different stages of myxobacterial morphogenesis and can be observed on different spatial scales from several cells to large streams [8–16] . Several adjacent streams can move circularly within the fruiting body in opposite directions [17] . Spores in the fruiting body of Myxococcus xanthus , the most studied myxobacterium , have been shown to be organized in spiral patterns , presumably as a result of such movements [5] . The mechanisms of formation of streams and circular or spiral aggregates are not well understood . Circular aggregates can form by a stream of cells trapping itself [8] . Cells have been observed to travel long distances in streams and enter distant aggregates rather that the ones nearby , suggesting that aggregation is not caused by a long-range diffusible signal emitted from aggregation centers [18] . Further , it has been shown that myxobacteria development is regulated by the C-signal that is passed from cell to cell through end-to-end contact [19] . These findings resulted in a hypothesis that myxobacteria aggregation and development depends on short-range contact-mediated communication between cells , i . e . cell aggregation does not require long-range signaling in the population . Recent studies on aggregate merging and dispersal dynamics further argues against the presence of long-range diffusible molecules to signal the aggregation process [6] . Vegetative cells in swarms reverse their direction of gliding by switching leading and trailing poles approximately once every 10 min [20 , 21] . In the course of development , due to C-signaling , reversal frequency of cells is reduced and gliding speed is increased . Therefore , cell movement becomes essentially unidirectional at the final stages of development [17 , 22 , 23] . Søgaard-Andersen and Kaiser [24] proposed that streams form when reversal frequency of cells is decreased due to C-signaling as they come into end-to-end contact . As a result , collective cell movement in roughly the same direction becomes favored . However , this model does not explain what keeps cells in the chain . Moving cell masses and streams can turn and swirl [8] , but cells appear to follow one another over long distances and not escape the stream due to random fluctuations in cell orientation [25] or contacts with surrounding cells . A guiding mechanism seems to be present for streams to be stable , i . e . for cells to continue following one another and move as one unit . One possible mechanism of such stability could be a long-range guiding system other than a signal diffusing from aggregation centers . For example , at low cell population densities , cells are often observed to follow slime trails laid down by other cells [26] . This could establish a long-range order required to guide cells into aggregation centers . However , whether and how slime trails could persist in a high-density population , which is the usual state of myxobacteria communities , let alone in three-dimensions , is not clear . Alternatively , cells could employ a short-range guiding mechanism whereby guiding forces act only when cells are in contact or very close to one another . Possible hypothetical mechanisms for short-range guiding could include following slime immediately extruded by another cell , response to a diffusible signal from another cell , physical adhesion between cells or attachment with type IV pili [27] . A number of modeling studies investigated myxobacteria motility ( e . g . , see [28–30] ) . However , few of them examined mechanisms of circular motility patterns . Lattice cell simulations showed that streams and ring-shaped aggregates , where cells move in circular tracks , could form as a result of local , short-range contact mediated interactions between cells , whereby rod shaped cells would preferentially turn towards maximizing end-to-end contacts [31–33] . However , modeling approach used there is not mechanically accurate , as cells are perfectly rigid ( i . e . cannot bend ) , can overlap in space and move only in limited number of directions . In this study , by means of a more mechanically accurate two-dimensional ( 2D ) computational mass-spring model developed earlier [34] , we investigate how different types of short-range guiding interactions between the leading pole of one bacterium and the trailing pole of another bacterium could affect the formation of patterns in myxobacteria population . In addition , we consider a case of long-range guiding between cells analogous to slime-trail following and compare the resulting patterns with the ones of short-range guided populations .
To model guiding interactions between cells , we use a 2D mass-spring model previously described in [34] with changes to collision response algorithm presented in S1 Text . In brief , a rod-shaped cell is modeled as an array of particles connected by linear and angular springs . Linear springs maintain the distance between particles , and thus the length of a bacterium , whereas angular springs ( characterized by angular spring constant ka ) resist bending of a cell . Cells glide on a substratum powered by engine forces and change their direction of movement as a result of collisions with other cells . Here , only the distributed engine is considered ( i . e . engine with forces distributed along the whole length of a cell ) , given recent evidence strongly supporting its existence [35–37] . In addition to the features described in the basic model , here we introduce and study three kinds of short-range guiding forces ( Fig 1 ) . First , adhesion between the leading pole ( “head” ) of one cell and the trailing pole ( “tail” ) of another cell is considered . Thereby , adhesion forces between a pair of line segments that connects particles in bacteria are introduced only when the head of one bacterium and the tail of another ( or the same ) bacterium are involved . If both interacting cells have polarity ke = 1 , the head of bacterium j is the point P1 = 0 on segment Q1j , and the tail of bacterium l is the point P2 = 1 on segment Q ( N − 1 ) l ( see [34] for notation ) . Thus , when the smallest distance between the two segments d is W < d < dg , where dg is maximum guiding ( in this case , adhesion ) distance and W is cell width , adhesion forces to respective head and tail particles of interacting bacteria are introduced ( Fig 1 , forces marked F H g and F T g ) . The adhesion forces are described by the same 4 equations that govern collision response [34] , with kc replaced by k g = F max g / d g , where F max g is the maximum magnitude of the guiding force ( exerted when two segments are separated by distance dg ) . Essentially , adhesion in the model is collision response working in reverse , i . e . attracting cells when d becomes larger than W . As a result of these forces , the head of the trailing cell will tend to turn towards the tail of the leading cell when the distance between them is small enough , due to the normal component of adhesion force on the head particle ( F H g , n in Fig 1 ) . In addition , the component of adhesion force along the tangent of trailing bacterium body ( F H g , t in Fig 1 ) will result in increased speed of the trailing bacterium ( i . e . the leading cell will pull the trailing cell forward ) . As adhesion forces work in action-reaction pairs , the tail of the leading cell will also turn towards the head of the trailing cell and the speed of the leading cell will tend to decrease ( due to normal and tangent component of adhesion force respectively , F T g , n and F T g , t in Fig 1 ) . Overall , when d < W ( i . e . when cells overlap ) , only collision forces would act to separate the cells , and there will be no adhesion forces . Collision and adhesion forces would be zero when the head of the trailing cell touches the tail of the leading cell ( i . e . when d = W ) . As distance between head and tail particles d increases beyond W , the magnitude of adhesion forces increases linearly until distance dg , where the magnitude of adhesion forces gets its maximal value F max g . Beyond dg , adhesion forces would be zero and thus cells will have no guiding interactions . A second type of short-range guiding force represents active following , whereby the force described for adhesion is added only to the head of the trailing cell , but not to the tail of the leading cell ( i . e . only F H g in Fig 1 is added ) . It models the effect of the trailing cell responding to the presence of the tail of the leading cell by actively moving in its direction , but having no effect on the movement of the leading cell . A third type of short-range guiding force is passive following ( steering ) , whereby the force to the head particle of the trailing cell is added only in the direction normal to bacteria body n ^ ( F H g , n in Fig 1 ) . By this , only the steering effect on the head of the trailing cell is modeled , i . e . turning the tip of the cell left or right with respect to the normal trajectory of the cell , but having no effect on cell speed . In addition , we also consider a case of long-range guiding that is analogous to slime trail following by myxobacteria cells . To model slime trails , a square grid with elements of side Δx is defined on the substratum . Each grid element can contain a unit vector s indicating slime trail direction at that location , or a zero vector , if no slime trail is present [38] . When the head particle of a bacterium glides over a grid element containing a slime trail s , the guiding force on the head particle is introduced to reorient the leading tip of that cell along the slime trail . Since a cell can glide in both directions along the slime trail , the guiding force is defined to turn the cell by an acute angle [26] . Thus , if the orientation of the leading tip o is defined as the tangent to bacterial body at the leading particle ( i . e . t ^ 1 when engine direction ke = 1 and t ^ N when ke = −1 ) and F max s is the maximal magnitude of guiding force , a guiding force sgn ( o · s ) F max s s is found and its component in the direction of n ^ is added to the leading particle . The applied force is similar to passive following described above , because the force only orients the tip of the cell along the slime trail , without affecting cell speed along tangent t ^ . After each integration step , the deposition of slime by the rear of the bacterium is modeled by assigning a tangent t ^ at the rear particle to slime trail s at a grid element below . The deposition of slime overrides the previous value of slime trail direction at that grid location . Slime trails at each grid location persist until overridden by other cells . The parameters used in the simulations are the same as in [34] , with the addition of extra parameters describing guiding forces . The value F max g was chosen to be 200 pN , unless stated otherwise , and dg = 0 . 25 μm ( i . e . half of bacterium width W ) . Since guiding forces not only steer the head of the cell , but can also speed up the cell , the value of F max g was chosen in such a way that the speed-up due to the guiding force would be roughly within experimentally observed speed increase of myxobacteria cells during development , 1 . 5–2 . 5 times [23] . F max g = 200 pN results in 3-fold maximum increase of speed ( engine force of 100 pN and maximum guiding force of 200 pN results in maximal 3vb speed ) . For long-range guiding simulations , F max s was also set to 200 pN and Δx = 0 . 25 μm ( half of bacterium width W ) . Since bending stiffness of myxobacteria cells have not been experimentally determined , but only theoretically estimated for M . xanthus [34 , 39] , a wide range of angular spring stiffness values ka were studied in the simulations: 1 × 10−18 , 1 × 10−17 , 1 × 10−16 and 1 × 10−15 N·m . They correspond to cell bending stiffness ( B ) values of 7 × 10−25 J·m ( referred to in the text as “very flexible” ) , 6 × 10−24 J·m ( “flexible” ) , 6 × 10−23 J·m ( “rigid” ) and 6 × 10−22 J·m ( “very rigid” ) , respectively . Further , both low density 5 × 106 cells/cm2 and high-density 4 × 107 cells/cm2 populations were studied . For a low density population simulation , the collision stiffness between cells was set as in [34] , kc = 0 . 01 N·m−1 . For high density populations , the collision stiffness had to be reduced to kc = 0 . 002 N·m−1 , because high collision stiffness blocks the movement of cells in a crowded environment . To analyze cell movement , cell speeds and strain energies due to collisions between cells were shown for every line segment in the bacterium . Speed of a line segment was defined as an average speed of two particles at the ends of the segment . To find strain energies , for every two segments that overlap due to collision ( i . e . the when the smallest distance between segments d < W ) , potential energy of the collision response spring ( 1/2 ) kc ( d − W ) 2 was calculated and one half of the value was added to both segments involved .
Firstly , the effect of different types of short-range guiding forces on cell movement patterns of a low-density population of non-reversing cells was studied . All cells were initially placed on a planar substratum with random positions and orientations ( Fig 2A ) , and cell movement was simulated for 6 hours . A population of flexible non-guided cells at 6 hours formed clusters ( Fig 2B and S1 Video ) , whereas the presence of steering forces between cells ( passive following ) resulted in occasional chains of cells between clusters and small unstable circular structures that quickly dissipated ( Fig 2C and S2 Video ) . However , cells with active following and head-to-tail adhesion formed stable rotating circular aggregates ( Fig 2D and S3 Video , and Fig 2E and S4 Video , respectively ) . During the process of aggregate formation , streams of cells were first formed from randomly distributed cells . A stream could collide with other streams , turn , move in circular trajectories , close in upon itself and trap the leading cells . The rest of stream cells then swirled around the trapped cells . The seed of rotation could also be formed by several cells swirling around a fixed point . Later , additional cells or entire streams could join in to increase the size of the aggregate . Within the aggregate , cells were arranged spirally and new streams joined in by following the freely exposed tail of a cell at the aggregate edge . The decrease of guiding force F max g from 200 pN to 100 pN ( referred to as weak guiding ) resulted in a more dynamic population that was less likely to form stable rotating aggregates . In such a population circular aggregates were smaller , could dissipate , and streams could leave one aggregate and join another ( S5 Video ) . Interestingly , when a cell at the edge of the aggregate left , it often had a chain of trailing cells behind it . Stable circular aggregates also form in a population of rigid non-reversing cells with long-range guiding ( Fig 4A and S16 Video ) . However , in contrast to short-range guided aggregates , most of the cells traveled with equilibrium speed , including the ones at aggregate edges , as guiding forces affected only cell movement direction but not cell speed . Further , aggregates appear to be less tightly packed , i . e . they contain more voids than short-range guided aggregates . Stress accumulation patterns inside aggregates , however , are similar in both cases ( Fig 4B and Fig 2F ) . Interestingly , flexible cells also exhibited marked circular movement and produced small short-lived circular aggregates , but they were unstable ( Fig 4C and S17 Video ) . The instability could be explained by easier bending of flexible cells under stress inside nascent aggregates . As cells can travel in both directions on a slime trail , bent flexible cells can switch their movement direction to the opposite , squeeze in or be pushed through voids in the aggregate and thus disturb the circular arrangement of slime trails .
The mechanisms of myxobacteria aggregation during fruiting body formation are not well understood . Non-linear patterns of movement of myxobacteria cell masses and streams imply the existence of some sort of guiding mechanisms that keep the cells moving as one unit and direct them into aggregation centers [8] . It is not known whether these guiding mechanisms are long-range or short-range . In this study , by means of a computational mechanical mass-spring model , we demonstrate that short-range guiding between the head and the tail of two myxobacteria cells in close contact are sufficient to produce stable streams and circular or spiral aggregates in model myxobacteria populations . Many features of cell movement that are present in our short-range guiding simulations are also observed in experimental videos . Multicellular cell masses ( streams ) in the simulations can travel in straight lines , or , when colliding with other streams or clusters of cells , can wave and swirl ( S5 Video , [8 , 11 , 14] ) . A circular aggregate in the simulations often forms when a stream turns , closes in upon itself and traps the leading cells , a situation also observed in experimental videos ( S3 Video; [8] ) . Simulated circular aggregates exhibit rotational movements ( S3 Video ) . Similarly , circular and spiral movement is often observed in developing myxobacteria [9 , 12] . In fact , fruiting bodies often develop in places where such spiral aggregates initially form [5] . Additional streams of cells join existing aggregates to increase their size ( S3 Video , [23] ) . In the simulations , a circular aggregate sometimes forms from a rotation seed of several flexible cells ( S3 Video and S6 Video ) . Similar small rotating cell clusters have been observed experimentally [10 , 15 , 40 , 41] . Furthermore , a smaller magnitude of guiding forces results in a more dynamic aggregate behavior: simulated streams can travel from one aggregation center to another , aggregates can dissipate , split or join with other aggregates ( S5 Video ) . Our results show that stability of large aggregates increases with increasing cell rigidity , whereas more flexible cells tend to form separate rotation seeds inside aggregates due to easier bending and thus induce splitting of a large aggregate ( S9 Video and S10 Video ) . It has been experimentally observed that the size of initial aggregates of different myxobacteria species differs [42] . Our results suggest that it might be the result of different bending stiffness of cells of different species [43] . Finally , we also observed the formation of hollow aggregates and adjacent streams swirling inside aggregates in opposite directions ( S10 Video , [4 , 17] ) . The formation of ring-shaped aggregates , where cells move in circular tracks , both clockwise and counter-clockwise , was also observed in lattice cell modeling studies [31–33] . In these studies cells preferentially turn towards maximizing end-to-end contacts with other cells and therefore the interactions between cells are effectively similar to guiding forces in our model . Interestingly , short-range guided circular aggregates in our model rotate as rigid bodies , i . e . cells within the aggregate do not slide laterally past one another . As a consequence , the further from the rotation axis cells are located , the faster they travel . Furthermore , cells at the edge of a simulated aggregate and cells in the incoming streams move faster than the equilibrium cell speed ( S3 Video ) . Therefore , in order for stable rotating aggregates to form , short-range guiding must act in such a way that the trailing cell is both turned towards the tail of the leading cell and sped up to catch up with it . Otherwise , because guiding interactions are short-range ( half the cell width ) , the faster moving leading cell will escape the trailing cell and the short-range guiding interaction will be lost . It has been shown that during myxobacterial development average speed of cells does increase due to interaction with other cells [23 , 44] , but unfortunately it was not reported whether cell speed depended on cell location within circular aggregates or streams . Long-range guided populations were also able to form circular aggregates in the simulations , but in contrast to short-range guided populations , cell speed increase was not required , as most cells inside aggregates traveled with equilibrium speed , including the ones at aggregate edges ( S16 Video ) . This finding could be used to experimentally discriminate between short-range and long-range guiding mechanisms present in myxobacteria . Observation of currently available experimental videos does not allow to tell conclusively whether circular aggregates rotate fully or partially as rigid bodies and whether cell speed increases with increasing distance from the rotation axis [9 , 12] . Experimentally tagging part of the cells in the population with fluorescent markers could be used to obtain such data . Another experimentally testable prediction of the model is that cell speed at the short-range guided aggregate edge is independent of the aggregate size . This means that smaller aggregates rotate with larger angular speeds , and increase of the aggregate size due to incoming streams will result in decrease of the angular speed of aggregate rotation . To our knowledge , there is no experimental evidence about the existence of short-range guiding interactions between a head and a tail of two myxobacteria cells . The model proposed in this study does not imply any particular short-range guiding mechanism for myxobacteria aggregation , as long as the interaction would both steer the trailing cell and adjust its speed . One possibility could be mechanical adhesion force between a head and a tail of two myxobacteria , or physical link between cells by type IV pili . A gliding M . xanthus cell extends type IV pili that originate at the leading pole , attach to neighboring cells and pull to produce motility force [27 , 45] . Groups of myxobacteria cells are usually well aligned [46] , therefore it is likely that the extended pilus will attach to the rear of the leading cell . Alternatively , the leading pole of a trailing cell could respond to a diffusible substance or slime secreted from the rear of the leading cell . For a short-range interaction , such a substance should diffuse slow enough to form gradients on the spatial scale of cell width and should break down quickly not to interfere with signaling between other cells at the same location at a later time . For example , lipids could satisfy slow diffusion requirement [6] . In such a scenario , a trailing cell would turn and adjust its speed based on the concentration of the diffusing substance . It has been shown that bacteria are not too small for spatial sensing of chemical gradients [47] . Furthermore , it has been observed that in low-density populations myxobacteria cells tend to follow slime trails produced by other myxobacteria [26] . It is not clear , however , whether slime trails could persist for a long time in high-density population , a usual state of myxobacterial communities . At high cell densities , a particular spot on a substratum is continuously overrun by other cells and existing slime trails thus would be overridden . Therefore , it is possible that slime trails in high-density populations are short-lived and extend no longer than the distance between adjacent cells . For short-range guiding , the trailing cell should follow only the new slime immediately secreted by the leading cell , but not the old slime . If slime-contained signaling molecule were broken down quickly after slime extrusion from the cell rear , its concentration would show slime age , and therefore , the distance to the leading cell that produced it . If , further , a trailing cell responded to older slime by increasing speed , the situation would be akin to active following considered in our simulations , as cell speed would be dependent on the distance between interacting cells . Furthermore , type IV pili can also attach to slime left behind by other cells [45] . If the pili attached only to immediately extruded slime and the force of pulling were proportional to the length of extended pilus , it would also present active following . It has also been shown that myxobacteria development depends on contact mediated C-signaling [19] . C-signal is relayed by the end-to-end contact between cells [48] , and one of its effects is to decrease cell reversal frequency [23] . C-signal mutants are unable to aggregate , or the aggregates that form quickly dissipate [49] . These results are consistent with C-signal acting as a part of guiding mechanism suggested by our model . In our simulations , weak guiding forces resulted in the formation of very dynamic aggregates that could easily disperse ( S5 Video ) . Although a real fruiting body develops in three dimensions , at the initial stages of aggregation cells appear to move in independent monolayers that are stacked on top of one another [5 , 49] . This observation suggests the presence of forces that keep cells confined to two-dimensional sheets and do not allow them to escape crowded environment by moving upward . It also justifies a 2D model in this study and explains how cell trapping is possible when streams close in upon themselves . Further , our simulations show that mechanical stress accumulates inside circular aggregates because cells are trapped and squeezed . It has been experimentally observed that when a second layer forms on top of the original monolayer of M . xanthus cells , cells leave the base layer at one point [49] . Our results suggest that this phenomenon may occur when mechanical stress reaches a critical value at some point inside the aggregate and cells at that point are propelled upwards . Consistent with this idea is the observation that fruiting bodies develop at the places of traffic jams [50] or where spiral aggregates initially form [5] . During vegetative swarm phase of myxobacterial life cycle , cells reverse their direction of gliding by switching the leading and trailing pole approximately once every 10 min [20 , 21] . In the course of development , reversal frequency of cells decreases and cell movement become essentially unidirectional at the final stages of development [17 , 22 , 23] . Reversing cells , in contrast to non-reversing cells , are unable to form circular aggregates in our simulations . This result is in a good agreement with experimental observations that circular aggregates do not form during vegetative stage of myxobacterial life cycle [46] , but only during fruiting body development . Furthermore , our study suggests an extension of the conceptual model whereby cell streams form when cell reversal frequency is reduced due to contact-mediated C-signaling as two cells come into end-to-end contact [24] . It was proposed that collective cell movement in roughly the same direction would be favored as a result . However , this model does not address the question of what keeps cells in the chain . Streams can turn and swirl [8] , but cells appear to follow one another over long distances and not escape due to contacts with surrounding cells or random fluctuations in cell orientation [25] . Our results show that cells with only suppressed reversals would not be able to form streams from initially randomly distributed reversing cells . However , the presence of guiding interactions in addition to reversal suppression allows for the formation of stable streams and circular aggregates ( S15 Video ) . Interestingly , in our high-density population simulations , initially aligned but randomly oriented rigid cells could mechanically sort into adjacent streams of cells moving in the same direction ( S12 Video ) , but it is not clear whether this effect occured due to a relatively small simulation domain . Furthermore , bending stiffness of myxobacteria cells has not been determined experimentally , but evidence suggests that for M . xanthus it is closer to the “flexible” value used in our simulations [51] . Guiding forces allow cells to form stable streams and circular aggregates independently of bending stiffness value and initial cell configuration .
Supporting videos are encoded in H . 264 format . Please note that not all media players can handle this format by default . Installation of a proper codec is needed in such cases . For the best viewing experience , we recommend VLC Media Player , freely available for a number of different platforms ( http://www . videolan . org/vlc/ ) .
|
Myxobacteria are social bacteria that upon starvation form multicellular fruiting bodies whose shape in different species can range from simple mounds to elaborate tree-like structures . The formation of fruiting bodies is a result of collective cell movement on a solid surface . Since collective cell motility during biological morphogenesis is also common in higher organisms , myxobacteria serve as a relatively simple model organism to study multicellular movement , organization and development . In the course of myxobacterial development , groups of flexible rod-shaped cells form streams and move in circular or spiral patterns to form aggregation centers that can become sites of fruiting body formation . The mechanisms of such cell movement patterns are not well understood . In this study , by means of a computational mass-spring model , we demonstrate that the formation of streams and circular aggregates during myxobacterial development can be explained by short-range head-to-tail guiding between individual cells , whereby movement direction of the head of one cell is affected by the nearby presence of the tail of another cell . We suggest that such interactions between cells can result from physical adhesion , response to a diffusible substance or slime extruded by cells , or the action of cell motility engine .
|
[
"Abstract",
"Introduction",
"Model",
"Results",
"Discussion",
"Supporting",
"Information"
] |
[] |
2015
|
Short-Range Guiding Can Result in the Formation of Circular Aggregates in Myxobacteria Populations
|
Juvenile hormones ( JHs ) play a major role in controlling development and reproduction in insects and other arthropods . Synthetic JH-mimicking compounds such as methoprene are employed as potent insecticides against significant agricultural , household and disease vector pests . However , a receptor mediating effects of JH and its insecticidal mimics has long been the subject of controversy . The bHLH-PAS protein Methoprene-tolerant ( Met ) , along with its Drosophila melanogaster paralog germ cell-expressed ( Gce ) , has emerged as a prime JH receptor candidate , but critical evidence that this protein must bind JH to fulfill its role in normal insect development has been missing . Here , we show that Gce binds a native D . melanogaster JH , its precursor methyl farnesoate , and some synthetic JH mimics . Conditional on this ligand binding , Gce mediates JH-dependent gene expression and the hormone's vital role during development of the fly . Any one of three different single amino acid mutations in the ligand-binding pocket that prevent binding of JH to the protein block these functions . Only transgenic Gce capable of binding JH can restore sensitivity to JH mimics in D . melanogaster Met-null mutants and rescue viability in flies lacking both Gce and Met that would otherwise die at pupation . Similarly , the absence of Gce and Met can be compensated by expression of wild-type but not mutated transgenic D . melanogaster Met protein . This genetic evidence definitively establishes Gce/Met in a JH receptor role , thus resolving a long-standing question in arthropod biology .
Arthropods possess unique sesquiterpenoid hormones , represented by the juvenile hormones ( JHs ) of insects [1] and their non-epoxidized precursor , methyl farnesoate ( MF ) in crustaceans [2 , 3] . JHs regulate insect metamorphosis , polymorphism and social caste determination , and adult reproductive physiology [1 , 4–6] . Although the sesquiterpenoid structure of JH was determined nearly five decades ago [7] , a receptor for these important hormones has been notoriously difficult to identify . Non-peptide lipophilic hormones usually exert genomic effects by activating nuclear receptor proteins [8–10] . One insect member of the nuclear receptor family , Ultraspiracle ( USP ) , has been proposed as a mediator of sesquiterpenoid action , initially of JH itself [11] and currently of MF [6 , 12–14] . USP is an appealing JH receptor candidate given its homology to the vertebrate retinoid X receptor ( RXR ) and an apparent level of similarity between JH and the RXR ligand , 9-cis-retinoic acid [15] . Moreover , USP is a subunit of the insect ecdysone receptor complex [10 , 16 , 17] , thus providing a potential point where the steroid and JH signaling pathways might converge . Whether or not the putative hormone-binding pocket of USP is capable of biologically significant ligand binding is still debated [13 , 14 , 18 , 19] . Discovery of the Methoprene-tolerant ( Met ) gene that confers resistance to the JH analog insecticide methoprene in the fruit fly , Drosophila melanogaster , has provided an alternative JH receptor candidate [20] . Nonetheless , absence of obvious effects of Met mutations on D . melanogaster development argued against the JH receptor function of Met until knockdown of Met in the flour beetle , Tribolium castaneum , produced precocious metamorphosis phenotypes consistent with disrupted JH signaling [21] . Later , it was shown in D . melanogaster that simultaneous mutation of Met and deletion of its paralog , the germ cell-expressed ( gce ) gene , resulted in non-conditional lethality during the larva-pupa transition [22] , corresponding to the lethal phase associated with deficiency of JH [22 , 23] . The Met and gce paralogs in D . melanogaster arose via gene duplication during "higher fly" evolution , whereas mosquitoes or beetles possess only a single gene [24] . Based mainly on evidence related to the position of introns , gce is ancestral to Met and , in spite of the nomenclature , D . melanogaster gce is more similar to the single Met genes found in other insects [24] . Met and Gce belong to the basic helix-loop-helix ( bHLH ) /Per-Arnt-Sim ( PAS ) family of transcription factors [25] that are distinctly different from nuclear receptor proteins . Although no bHLH-PAS protein has previously been proven to be a receptor for an authentic hormone , the vertebrate aryl hydrocarbon receptor ( AhR ) is a transcription factor activated by xenobiotics ( e . g . , dioxin ) , or by endogenous ligands such as tryptophan metabolites , binding to its PAS-B domain [26 , 27] . Like JH , Gce/Met is unique to arthropods , and thus may have evolved to mediate JH signaling in insects , crustaceans , and other related taxa . In vitro , the Met proteins from D . melanogaster [28 , 29] , T . castaneum [29] , and the Aedes aegypti mosquito [30] bind native JH ( JH III ) with nanomolar affinities . Specific mutations within the PAS-B domain of T . castaneum or A . aegypti Met preclude this JH binding [29 , 30] . JH induces Met to bind to another bHLH-PAS protein Taiman ( Tai ) , also known as FISC or SRC [29 , 31–33] . The resulting complex binds JH-response DNA motifs and activates target gene transcription [30–35] . Similarly to Met , Tai has also been shown to mediate effects of JH on metamorphosis [36] and reproduction [37 , 38] in some insects . The D . melanogaster Met and Gce proteins interact with a chaperone Hsp83 , which facilitates nuclear import of Met and expression of JH-induced genes such as Krüppel homolog 1 ( Kr-h1 ) [34] . Most recently , Met and Gce were shown to mediate the effect of the JH precursor MF , which has been established as a circulating hormone in D . melanogaster [39] . Taken together , the above results favor Gce/Met as a JH receptor candidate . However , to establish conclusively that Gce/Met is a JH receptor , it must also be demonstrated that binding of the hormone is a necessary condition for functioning of the candidate receptor in vivo , during normal insect development . This study employs the power of Drosophila genetics to provide this critical missing evidence . It shows that transgenic Gce or Met proteins restore the natural sensitivity to JH mimics in the Methoprene-tolerant mutants and rescue the non-conditionally lethal Met gce double-mutant flies as long as their JH-binding pocket is intact .
The D . melanogaster S2 cell line expresses endogenous mRNAs encoding both Met and Gce paralogs and their single partner protein Tai ( S1 Fig ) . We initially tested whether Met , Gce and Tai mediated ligand-dependent transcriptional activation in the S2 cells . A luciferase reporter JHRE-luc was constructed using eight tandem copies of a JH-response element ( JHRE ) from the A . aegypti early trypsin gene [30 , 31] ( Fig 1A ) . JHRE-luc was activated by a native JH ( JH III ) , the JH mimic methoprene , and by MF in a dose-dependent manner ( Fig 1B ) . Mutation of the JHRE inhibited the response to JH III ( Fig 1B ) . RNAi-mediated knockdown of either tai or gce but not of Met prevented JH III or MF from inducing JHRE-luc ( Fig 1C ) . Expression of additional Tai enhanced this hormone-dependent activation , again in a manner dependent primarily on gce and tai ( Fig 1D ) . Similar results were obtained utilizing pyriproxyfen , a potent JH mimic of distinct , pyridine-based chemical structure [40] ( S2 Fig ) . The observation that Gce and Tai were required for activation of JHRE-luc by MF ( Fig 1C and 1D ) , is consistent with a previous finding that MF activated transcription through an ortholog of Gce/Met from the silkworm , Bombyx mori [29 , 31–33] and the recent finding that this natural JH precursor is a circulating hormone in D . melanogaster [39] . As the effect of JH in the S2 cell-based assay was essentially mediated by Tai and Gce , we examined the ability of the Gce protein in vitro to bind the activating ligands . [3H]JH III bound to Gce with a Kd of 19 . 3 ± 4 . 5 nM ( Fig 2A ) , an affinity within the physiological hormone range [13] . Following on from the reporter gene activation data ( Fig 1B ) , methoprene , pyriproxyfen , and MF all effectively competed with [3H]JH III for binding to Gce ( Fig 2B ) , consistent with both JH mimics and MF acting as JH receptor agonists . Similarly to binding affinities previously determined for the PAS-B domain of T . castaneum Met [29] , pyriproxyfen was the strongest competitor for binding to Gce , followed by methoprene and MF ( Fig 2B ) . The higher potency of methoprene to activate JHRE-luc , relative to JH III ( Fig 1B ) may be explained by the fact that the synthetic insecticide is chemically and biologically more stable than JH III . Due to marginal levels of total [3H]JH III bound to the in-vitro translated D . melanogaster Met protein [29] , we were unable to determine the ligand-binding affinities for Met . The activation and binding of Gce by MF is significant , as this circulating JH precursor prevails over JH III in D . melanogaster larvae [13 , 39] and exerts its own hormonal function [39] . Interestingly , MF has been reported to bind D . melanogaster USP with a high affinity ( Kd = 40 nM ) [12] , comparable to the Ki of 87 . 9 nM we observed for MF binding to Gce ( Fig 2B ) . USP has therefore been proposed as an intracellular MF receptor [6 , 13 , 14] . However , in agreement with genetic evidence [39] , our RNAi data ( Fig 1C and 1D ) clearly show that Gce and Tai are essential for MF to induce expression of the JHRE-dependent reporter and thus act as a MF receptor . MF is a "juvenile hormone" of crustaceans , where it promotes reproductive maturation and specific developmental events [41 , 42] . Interestingly , similar to JH in insects [29 , 31–33] , MF has been shown to stimulate interaction between Met and Tai/SRC orthologs from the cladoceran crustaceans , Daphnia pulex and D . magna [43] . Moreover , when a threonine residue in the PAS-B domain of Daphnia Met was replaced with valine that occurs in the corresponding position critical for JH III binding in insects , namely V315 in D . melanogaster Gce ( Fig 3A and 3B ) or V297 in T . castaneum Met [29] , the Daphnia Met protein became more responsive to JH III , without losing its responsiveness to MF [43] . Together with our current findings , this recent evidence suggests that Gce/Met has evolved as a receptor for sesquiterpenoid hormones in a common ancestor of crustaceans and insects . To determine whether Gce required direct binding of JH for its function , we individually mutated three amino acids ( T272Y , V315F , and C366M ) in the ligand-binding pocket of Gce PAS-B ( Fig 3A ) . The same substitutions of the corresponding residues have been shown to abolish binding to JH III in the Met proteins from T . castaneum [29] and A . aegypti [30] . As expected , all three mutated Gce proteins lost the ability to bind [3H]JH III in vitro ( S3 Fig ) , indicating that these conserved T , V and C residues are critical for hormone binding by D . melanogaster Gce . To test whether binding of JH was necessary for Gce to activate transcription , FLAG epitope-tagged wild-type ( FLAG-GceWT ) or mutated ( FLAG-GceT272Y , -GceV315F , and-GceC366M ) proteins were expressed in S2 cells , in which the endogenous Gce and Met were suppressed by RNAi . Clearly , only FLAG-GceWT responded to JH III to activate the JHRE-luc reporter , whereas the three Gce variants incapacitated for hormone binding did not ( Fig 3B ) . The wild-type and mutated Gce proteins all appeared to be stable in the S2 cells ( Fig 3B , inset ) . Although the endogenous Met protein did not appear to play a major role in the JH-dependent activation of JHRE-luc in the S2 cell line ( Fig 1C and 1D ) , D . melanogaster Met could in fact substitute for Gce in this reporter assay when transfected to the cells ( Fig 3C ) . The S2 cells were again subjected to RNAi-mediated depletion of the endogenous Met and Gce proteins but not of the added Met protein that was expressed from a synthetic DNA construct . Like Gce , Met but not its mutated versions , mediated induction of JHRE-luc by JH III ( Fig 3C ) . Importantly , the functional JHRE-luc reporter was not activated by Met that had been mutated in its PAS-B domain with individual substitutions T406Y , V449F , and C500M that correspond to the T272Y , V315F , and C366M mutations in Gce ( Fig 3C ) . These mutations did not lead to degradation of Met ( Fig 3C , inset ) . Although we have been unable to directly confirm the effect of these mutations on the ligand-binding activity of D . melanogaster Met , it is most likely that they prevent JH III binding just as equivalent substitutions of these highly conserved residues do in Gce ( S3 Fig ) and in the Met proteins from T . castaneum [29] and A . aegypti [30] . These results strongly suggest that the JH-binding capacity is required for the normal function of D . melanogaster Met . In D . melanogaster the Met and gce genes reside on the X chromosome and their simultaneous loss in females that are homozygous or males that are hemizygous for the Met27 and gce2 . 5k null alleles is lethal at the onset of pupation [22] . The Met27 gce2 . 5k double mutants are known to express reduced mRNA levels of the Kr-h1 gene , which is a direct target of Gce/Met [22 , 34 , 39] . In order to demonstrate that the JH-binding capacity of Gce is important in vivo for transcription of this relevant JH-response gene , we expressed the wild-type and mutated forms of Gce using the ubiquitous armadillo-Gal4 driver ( arm-Gal4 ) in the Met27 gce2 . 5k background . To avoid the lethal phase in this strain , we examined Kr-h1 expression in Met27 gce2 . 5k/Y male larvae that were selected and genotyped during mid-third instar . The equal performance of the Gce variants was ensured by inserting all transgenic UAS-gce constructs into the same genetic locus [44] . Consistent with previous reports [22 , 34 , 39] , we observed reduced Kr-h1 levels in Met27 gce2 . 5k mutants ( Fig 4 ) albeit the difference was less dramatic , likely due to the earlier stage of our animals at which Kr-h1 expression is lower and less dependent on JH [45] . Addition of the transgenic GceWT protein to the Met27 gce2 . 5k background significantly augmented Kr-h1 expression near levels occurring in Met+ gce+ sibling male larvae ( Fig 4 ) . In contrast , the Kr-h1 transcript remained low when any of the three mutated forms of Gce were expressed ( Fig 4 ) . Therefore , only when capable of binding its hormonal ligand ( S3 Fig ) , Gce could compensate for the missing endogenous receptor proteins in restoring the normal expression of their target gene . As the function of Kr-h1 is essential for D . melanogaster to complete the prepupal stage [45] , compromised Kr-h1 expression may be contributing to the lethality resulting from the absence of Gce and Met . To further investigate the receptor function of Gce in vivo , we tested the relationship of JH binding to the phenomenon of "methoprene tolerance"–the insecticide resistance phenotype for which the D . melanogaster Met mutants were originally isolated and named [20] . Strains singly mutant either for Met or , to a lesser extent gce , resist doses of JH mimics that kill flies possessing both wild-type genes [20 , 22 , 46 , 47] . It has been shown that ubiquitous expression of a gce+ transgene using the Gal4/UAS system is sufficient to reinstate sensitivity to methoprene in the Met27 null mutants [46] . We took this approach with our GceWT , GceT272Y , GceV315F , and GceC366M transgenic constructs . When expressed under the arm-Gal4 driver , only GceWT restored sensitivity to dietary methoprene in Met27 homozygotes ( Fig 5A ) . In fact , these Met27 animals expressing GceWT became more sensitive to methoprene than Met+ controls , reflecting a dominant effect of the additional GceWT protein . In contrast , Met27 males and females expressing any of the three mutated Gce variants remained resistant and emerged as adults after feeding on methoprene ( Fig 5A ) . Similar results were obtained with pyriproxyfen ( S4 Fig ) . Thus , the lethal action of the insecticidal JH mimics relies on the ligand-binding capacity of the transgenic Gce protein . To obtain similar information for Met , we repeated this experiment with fly strains expressing the D . melanogaster wild-type Met protein or its mutated versions MetV449F and MetC500M ( our initial attempt to transform flies with UAS-MetT406Y failed ) . Although the flies expressing MetWT under the arm-Gal4 driver did not become more sensitive than control Met+ flies , their response to dietary methoprene significantly increased relative to the original Met27 mutants or the same mutants carrying the MetV449F and MetC500M transgenes ( Fig 5B ) . Our data thus demonstrate that Gce and Met are mutual substitutes in rendering flies sensitive to exogenous JH mimics as long their ligand-binding pockets are unaffected by specific mutations . The non-conditional lethality of the Met27 gce2 . 5k double-mutants can be rescued with transgenic constructs providing either Met+ or gce+ function , thus reflecting partial redundancy between Met and Gce [22] . This genetic rescue offers an ideal system to answer the ultimate question as to whether Gce/Met requires its JH-binding capacity to sustain normal development of the animal . Using two ubiquitous drivers , arm-Gal4 and α-tubulin ( tub-Gal4 ) , and transgenic UAS-gce and UAS-Met constructs uniformly inserted to the attP2 chromosomal site [44] , we expressed the functional or mutated proteins in the Met27 gce2 . 5k background . Indeed , expression of GceWT or MetWT under both drivers rescued a major proportion of Met27 gce2 . 5k/Y hemizygous males to adulthood ( Fig 5C and 5D ) . In striking contrast , the mutated GceT272Y , GceV315F , GceC366M , MetV449F or MetC500M proteins did not allow any Met27 gce2 . 5k/Y adults to emerge ( Fig 5C and 5D ) . Therefore , only Gce/Met with intact JH-binding function can substitute for the absence of both genes during normal development . To examine whether Gce incapacitated for JH binding was stable in vivo , we expressed FLAG-tagged versions of GceWT , GceT272Y , and GceV315F in transgenic D . melanogaster . Again , only the functional but not the JH binding-deficient tagged protein provided a clear rescue of the Met27 gce2 . 5k mutants ( Fig 5C ) . Interestingly , marginal rescue of 7 . 5% of emerging adults was observed , albeit only with the arm-Gal4 driver , with FLAG-GceV315F ( Fig 5C ) , suggesting that this mutated protein might retain some residual functionality . The discrepancy between this weak effect and the total absence of rescue by untagged GceV315F ( Fig 5C ) might result from variable expression level of the FLAG-tagged construct that , unlike the untagged constructs , had been integrated to random loci rather than to the specific attP2 site . Indeed , from three independent FLAG-GceV315F transgenic lines , only one showed the partial genetic rescue . Importantly , all three FLAG-tagged Gce variants were detected on immunoblots from whole transgenic flies ( Fig 5E ) , and all were observed primarily in the nuclei of larval fat body cells ( Fig 5F ) , regardless of whether or not Gce was mutated to prevent binding of JH . Thus , the failure of mutated Gce to compensate for the loss of the endogenous Met and Gce proteins was more likely caused by inability to bind JH rather than by degradation or mislocalization of the mutated protein . In conclusion , our study shows that the capacity of Gce , and most likely also of Met , to promote gene expression and sustain normal development requires direct hormone binding to the protein in vivo . The case that Gce/Met acts as a JH receptor in insects is now unequivocal . Establishment of the nature of this receptor resolves a central problem in arthropod endocrinology . The ability of Gce to respond to methyl farnesoate , the crustacean JH , suggests that the role of Gce/Met in sesquiterpenoid signaling predates the evolutionary separation of the hexapod and crustacean lineages . Furthermore , it is of interest that Gce/Met provides the first clear example of a bHLH-PAS protein acting as a receptor for a genuine animal hormone .
DNA sequences corresponding to the D . melanogaster Gce ( amino acids 1–689; NCBI Reference Sequence NP_511160 . 1 ) and Met ( amino acids 1–716; NCBI Reference Sequence NP_511126 . 2 ) were synthesized for optimal D . melanogaster codon usage to encode the Gce and Met wild-type ( WT ) and mutated ( T272Y , V315F , C366M , T406Y , V449F , C500M ) variants . For transcription/translation in vitro , these DNA fragments were cloned using the Eco RI and Kpn I restriction sites behind the T7 promoter in the pK-Myc-C2 plasmid [48] . The same gce and Met DNA sequences were inserted under the UAST promoter in two different vectors for D . melanogaster transformation: pTFW ( Drosophila Genomics Resource Center ) , in which Gce was N-terminally tagged with a FLAG epitope , and pUASTattB [49] that permitted integration of the gce/Met transgenes into the specific attP2 chromosomal landing site [44] . Racemic ( RS ) tritiated JH III ( 10–20 Ci mMol-1 ) was purchased from Perkin Elmer . Racemic JH III , pyriproxyfen , trans , trans-farnesol and methoprene were from Sigma-Aldrich , and ( E , E ) -methyl farnesoate ( MF ) from Echelon Biosciences . The WT , T272Y , V315F , and C366M variants of Gce were produced with the rabbit reticulocyte lysate TnT Quick Coupled transcription/translation system ( Promega ) using 400 ng of template plasmid per 50-μl reaction . Each reaction was divided into 15-μl aliquots that were assessed for binding of [3H]JH III using the dextran-coated charcoal ( DCC ) method as described previously [29] . The dissociation constant ( Kd ) was determined for GceWT binding to [3H]JH III , and the Ki values for methoprene , pyriproxyfen , MF and farnesol were calculated from competition assays with the unlabeled compounds using GraphPad Prism 5 . 00 ( GraphPad Software ) as described [29] . A JH-responsive luciferase reporter ( JHRE-luc ) was generated using a JH response element ( JHRE ) ( 5'- CCATCCCACACGCGAAGACGATAAAACCA- 3' ) identified upstream of the Aedes aegypti early trypsin ( AaET ) gene [31] . A mutated version of this element ( 5'- CCATCCCAGTGCGCAAGACGATAAAACCA -3' ) was used to generate a negative-control mutJHRE-luc . DNA sequences were synthesized to include eight copies of either JHRE or mutJHRE , followed by a 140-bp minimal promoter of the AaET gene ( nucleotides -77 to +63 ) . These sequences were cloned to the pGL4 . 17 vector containing the firefly ( Photinus pyralis ) luc2 gene ( Promega ) . D . melanogaster Schneider 2 ( S2 ) cells were cultured in Shields and Sang M3 Insect Medium ( Sigma-Aldrich ) containing 8% of heat-inactivated fetal bovine serum ( Life Technologies ) at 25°C . For luciferase reporter assays , S2 cells were seeded in a 12-well plate containing 900 μl of medium per well , and cultured for 24 h . The JHRE-luc ( or mutJHRE-luc ) reporter plasmid ( 0 . 25 μg per well ) was co-transfected with a pCopia plasmid ( 0 . 1 μg per well ) encoding Renilla luciferase using the X-tremeGENE HP DNA Transfection Reagent ( Roche ) . Where appropriate , the D . melanogaster Tai protein was expressed from a pCMA plasmid ( 0 . 25 μg per well ) containing tai cDNA [31 , 50] . Expression of either the wild-type or mutated FLAG-tagged Gce and Met variants was achieved by co-transfecting 0 . 25 μg of a pTFW vector carrying the respective gce or Met DNA sequence under the UAST promoter with 0 . 1 μg of a plasmid expressing the Gal4 transcription factor under a D . melanogaster actin promoter . The total DNA load per well was kept constant at 1 μg by inclusion of non-specific plasmid DNA . Following transfection , cells were incubated for 48 h and treated for another 12 h with JH III , methoprene , pyriproxyfen , MF or farnesol ( all dissolved in ethanol ) . The cells were then processed with the Dual-Luciferase reporter assay system ( Promega ) . Relative luciferase activity was measured using the Orion II microplate luminometer ( Berthold Detection Systems ) and data were normalized against Renilla luciferase activity . Met , gce and tai cDNAs were obtained by reverse transcription of total D . melanogaster embryonic RNA , followed by PCR amplification with specific primer sets ( S1 Table ) . The cDNA fragments were flanked with T7 promoter sequences to enable synthesis of double-stranded RNA ( dsRNA ) using T7 RNA polymerase ( MEGAscript , Ambion ) . A 720-bp dsRNA derived from the egfp gene served as a control . To knock down Met , gce , and tai genes in S2 cells , 3 μg of dsRNA per well of a 12-well plate were added together with plasmid DNA in the transfection mixture . The dsRNA sequences targeting endogenous gce and Met did not interfere with expression of the Gce and Met ( WT or mutated ) proteins transfected with the pTFW vector , as those were encoded by synthetic DNA divergent from the endogenous DNA sequences . Moreover , gce dsRNA targeted an upstream region of the native gce transcript that did not overlap with the synthetic sequence included in the pTFW-gce constructs . Total RNA isolated from whole mid-third instar D . melanogaster larvae or S2 cells using the Trizol reagent ( Life Technologies ) was treated with TURBO DNase ( Ambion ) , and 1 . 5 μg of RNA was reverse transcribed to cDNA ( Superscript II , Life Technologies ) . Relative transcript levels were measured in a C1000 Thermal Cycler ( Bio-Rad ) using the iQ SYBR Green Supermix kit ( Bio-Rad ) using specific primer sets ( S2 Table ) and normalized against levels of the ribosomal protein 49 ( rp49 ) mRNA . Targeted insertion of gce transgenes into the attP2 landing site on the third chromosome ( cytological position 68A4 ) was achieved using the bacteriophage ϕC31 integrase method [44] . The pUASTattB constructs containing the WT and mutated gce or Met sequences were injected into embryos of the y w P{nos-ϕC31\int . NLS}X; P{CaryP}attP2 host strain ( Genetic Services , Inc . or BestGene , Inc . ) . Several independent transgenic lines for expression of the FLAG-tagged GceWT , GceT272Y , and GceV315F proteins were generated through conventional P-element mediated transformation by injecting embryos of the w1118 host strain with the pTFW-based vectors ( Genetic Services , Inc . ) . In all cases , expression of the transgenic proteins was induced using the Gal4/UAS system [51] with the ubiquitous armadillo ( arm-Gal4 ) or α-tubulin ( tub-Gal4 ) drivers ( Bloomington Drosophila Stock Center , Indiana ) . D . melanogaster with unaffected Met+ function are sensitive to exogenous JH or its mimics as early prepupae [52] , whereas flies deficient for Met tolerate exposure to these compounds [20 , 47] . To test for restoration of methoprene sensitivity to Met27 mutants , homozygous Met27; arm-Gal4 females were mated with Met27/Y; UAS-gce or Met27/Y; UAS-Met males carrying the wild-type or mutated transgenes , all inserted into the same attP2 landing site [44] . The uniform Met27; arm-Gal4/+; UAS-gce ( or UAS-Met ) /+ F1 progeny was exposed to methoprene , pyriproxyfen or ethanol alone from the outset of larval feeding , and numbers of emerged adults were scored relative to all animals forming pupae . To test for rescue of viability in the non-conditionally lethal Met27 gce2 . 5k double mutants , balanced Met27 gce2 . 5k/FM7c; arm-Gal4 or Met27 gce2 . 5k/FM7c; tub-Gal4 females were crossed with w1118/Y; UAS-gce or w1118/Y; UAS-Met males harboring the wild-type or mutated gce or Met transgenes in the attP2 landing site . To detect the transgenic Gce proteins , we used males with UAS-FLAG-gce transgenes carried on the pTFW vector and inserted into random genomic loci . Immunoblots were prepared from total D . melanogaster S2 cell lysates or from entire adult transgenic flies and processed with an anti-FLAG antibody ( Sigma-Aldrich; 1:4000 ) and with anti-Mbf1 or anti-Cheerio antibodies as previously described [53] . Clones overexpressing WT or mutant Gce proteins were induced using the heat-shock-FLPout technique [54] , whereby y w hs-flp; act>y+>Gal4 , UAS-GFP females were mated to UAS-FLAG-gce transgenic males . Fat bodies dissected from larval progeny one day after heat shock ( 37°C , 30 min ) were stained with anti-FLAG ( Sigma-Aldrich; 1:1000 ) and Cy3-conjugated ( Cell Signaling ) antibodies , and images were captured with the Olympus FV1000 confocal microscope .
|
Juvenile hormones ( JHs ) play critical roles in the development of arthropods , comprising half the animal biomass of the oceans and over a million insect species , which have an enormous impact on ecosystems , agriculture ( pollinators and pests ) and health of mankind ( disease vectors ) . Despite decades of research , a receptor for these unique sesquiterpenoid hormones has remained elusive . Here , we provide definitive genetic evidence establishing that the essential biological function of the Gce/Met protein during insect development is critically dependent on its ability to bind JH , in effect functionally defining a JH receptor . Unequivocal identification of a JH receptor has profound implications for our understanding of arthropod biology . It also defines a molecular target for development of environmentally friendly , safer insecticides .
|
[
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Genetic Evidence for Function of the bHLH-PAS Protein Gce/Met As a Juvenile Hormone Receptor
|
Structurally segregated and functionally specialized regions of the human cerebral cortex are interconnected by a dense network of cortico-cortical axonal pathways . By using diffusion spectrum imaging , we noninvasively mapped these pathways within and across cortical hemispheres in individual human participants . An analysis of the resulting large-scale structural brain networks reveals a structural core within posterior medial and parietal cerebral cortex , as well as several distinct temporal and frontal modules . Brain regions within the structural core share high degree , strength , and betweenness centrality , and they constitute connector hubs that link all major structural modules . The structural core contains brain regions that form the posterior components of the human default network . Looking both within and outside of core regions , we observed a substantial correspondence between structural connectivity and resting-state functional connectivity measured in the same participants . The spatial and topological centrality of the core within cortex suggests an important role in functional integration .
Human cerebral cortex consists of approximately 1010 neurons that are organized into a complex network of local circuits and long-range fiber pathways . This complex network forms the structural substrate for distributed interactions among specialized brain systems [1–3] . Computational network analysis [4] has provided insight into the organization of large-scale cortical connectivity in several species , including rat , cat , and macaque monkey [4–7] . In human cortex , the topology of functional connectivity patterns has recently been investigated [8–11] , and key attributes of these patterns have been characterized across different conditions of rest or cognitive load . A major feature of cortical functional connectivity is the default network [12–18] , a set of dynamically coupled brain regions that are found to be more highly activated at rest than during the performance of cognitively demanding tasks . Spontaneous functional connectivity resembling that of the human default network was reported in the anaesthetized macaque monkey , and functional connectivity patterns in the oculomotor system were found to correspond to known structural connectivity [19] . Computational modeling of spontaneous neural activity in large-scale cortical networks of the macaque monkey has indicated that anti-correlated activity of regional clusters may reflect structural modules present within the network [20] . These studies suggest that , within cerebral cortex , structural modules shape large-scale functional connectivity . Understanding the structural basis of functional connectivity patterns requires a comprehensive map of structural connection patterns of the human brain ( the human connectome [1] ) . Recent advances in diffusion imaging and tractography methods permit the noninvasive mapping of white matter cortico-cortical projections at high spatial resolution [21–25] , yielding a connection matrix of inter-regional cortical connectivity [26–29] . Previous studies have demonstrated small-world attributes and exponential degree distributions within such structural human brain networks [26 , 27] . In the present study , using diffusion spectrum imaging ( DSI ) we derived high-resolution cortical connection matrices and applied network analysis techniques to identify structural modules . Several techniques reveal the existence of a set of posterior medial and parietal cortical regions that form a densely interconnected and topologically central core . The structural core contains numerous connector hubs , and these areas link the core with modules in temporal and frontal cortex . A comparison of diffusion imaging and resting state functional MRI ( fMRI ) data reveals a close relationship between structural and functional connections , including for regions that form the structural core . We finally discuss anatomical and functional imaging data , suggesting an important role for the core in cerebral information integration .
Network analyses were carried out for high-resolution connection matrices ( n = 998 regions of interest [ROIs] with an average size of 1 . 5 cm2 ) , as well as for regional connection matrices ( n = 66 anatomical subregions ) ( see Methods and Figure 1 ) . All networks covered the entire cortices of both hemispheres but excluded subcortical nodes and connections . When not indicated otherwise , the data shown in this paper are based on the analysis of individual high-resolution connection matrices , followed by averaging across five human participants . Network measures included degree , strength , betweenness centrality , and efficiency ( see Methods ) . Briefly , degree and strength of a given node measure the extent to which the node is connected to the rest of the network , while centrality and efficiency capture how many short paths between other parts of the network pass through the node . A node with high degree makes many connections ( where each connection is counted once ) , while a node with high strength makes strong connections ( where strength is equal to the sum of connection density or weight ) . A node with high betweenness centrality lies on many of the shortest paths that link other nodes in the network to one another . A node with high efficiency is itself found to be , on average , at a short distance from other nodes in the network . We found binary , high-resolution brain networks to be sparsely connected , with connection densities varying between 2 . 8% and 3 . 0% . Between 9% and 14% of all binary connections were interhemispheric . 54% of the total edge mass ( the sum of all fiber densities ) was accounted for by connections linking ROIs belonging to the same anatomical subregion , 42% was made between ROIs belonging to different anatomical subregions located in the same cortical hemisphere , and 4% was interhemispheric ( homotopic or heterotopic ) . Confirming earlier reports [25] , we found that cumulative distributions of node degree and node strength ( Figure S1 ) were exponential rather than scale-free . While not scale-free , node degrees and node strengths for single ROIs can vary over a significant range ( approximately 10-fold ) , indicating that fiber densities are not uniformly distributed across the cortical surface . Figure 2A and 2B shows the distribution of average node degree and node strength rank-ordered by anatomical subregion . A large number of ROIs with high degree and high strength are localized within subregions of medial cortex ( e . g . , cuneus and precuneus , posterior and anterior cingulate cortex ) and temporal cortex ( e . g . , bank of the superior temporal sulcus ) . A plot of the distribution of node strengths on the cortical surface across all participants ( Figure 2C ) shows consistently high values in posterior medial cortex , in medial frontal cortex , and in superior temporal cortex . In addition , we found evidence for positive assortativity ( Text S1 ) and small-world attributes ( Text S2 ) . A representative example of a high-resolution structural connection matrix of an individual human brain is shown in Figure 3A . Entries of the matrix represent fiber densities between pairs of single ROIs . The matrix shown in the example displays a total of 14 , 865 symmetric connections ( connection density 3 . 0% ) . To visualize structural patterns within this connection matrix , we extracted the connectivity backbone ( [30] , see Methods ) , which is displayed in Figure 3B with a layout derived from the Kamada-Kawai force-spring algorithm [31] implemented in Pajek [32] . The algorithm generates a spatial arrangement of ROIs along clearly defined anterior-posterior and medial-lateral axes and reveals clusters of dense connectivity within posterior , temporal , and frontal cortex . Figure 3C shows the connectivity backbone plotted in anatomical coordinates . The dorsal view shows groupings of highly interconnected clusters of ROIs arranged along the medial cortical surface , extending from the precuneus via posterior and anterior cingulate cortex to the medial orbitofrontal cortex . Dorsal and lateral views additionally show clusters of temporal and frontal ROIs in both hemispheres . Major structural patterns become more evident when considering the average regional connection matrix ( Figure 4A ) . The matrix is constructed by calculating mean fiber densities over individual pairs of ROIs comprising each subregion , followed by the averaging of densities over all five participants . Regional connection matrices for each individual participant are shown in Figure S2 . Figure 4B displays the connectivity backbone constructed from the average regional connection matrix , revealing groupings of anatomical regions largely corresponding to those shown for the high-resolution backbone in Figure 3B . A dominant feature of the regional connection matrix is a single , callosally interconnected cluster of regions extending from the cuneus and precuneus via cingulate cortex to medial frontal cortex . In addition , each hemisphere contains a single , relatively distinct cluster of temporal cortical regions , as well as a less-densely interconnected frontal cluster comprising periorbital cortex , pars opercularis , pars triangularis , and other regions . While network visualization provides strong hints of connectional relationships , objective methods are needed to map structural cores , to delineate network modules , and to identify hub regions that link distinct clusters . We quantified these phenomena using k-core decomposition [33] , spectral community detection [34] , and nodal participation indices [35] , respectively . Intuitively , a network core is a set of nodes that are highly and mutually interconnected . For a binary network , the k-core is the largest subgraph comprising nodes of degree at least k , and is derived by recursively peeling off nodes with degree lower than k until none remain [33] . Each node is then assigned a core number , which is defined as the largest k such that the node is still contained in the k-core . We performed k-core decomposition on binary , high-resolution connection matrices from all five participants and derived the core number for each ROI , as well as the average core number for each anatomical subregion ( Figure 5 ) . A large core number indicates that an ROI or region is resistant to this erosive procedure and participates in high-k structural cores of the network . In all participants , full erosion occurs at a core number of ∼20 . The most consistent members of the highest degree k-core for each network ( Figure 5A and 5B ) were the precuneus , the posterior cingulate , the isthmus of the cingulate , and the paracentral lobule in both hemispheres . In all participants , the structural core was located within posterior medial cortex , and often extended laterally into parietal and temporal cortices , especially in the left hemisphere . A rank-ordered distribution of average core numbers per anatomical subregion ( Figure 5C ) identifies the posterior cingulate cortex , the isthmus of the cingulate cortex , the precuneus , the cuneus , and the paracentral lobule as regions with a high core number . Several temporal and parietal structures , including the superior and inferior parietal cortex , the bank of the superior temporal gyrus , and transverse temporal cortex all have high core rankings as well . k-Core decomposition , as applied in our study , largely discards edge weights . To test if the inclusion of edge weight information would alter our conclusions , we designed a procedure that operates on the weighted fiber density matrix and erodes vertices according to their strength ( “s-core decomposition” ) . s-Core decomposition ( Figure S3 ) identified the posterior cingulate cortex , the precuneus , the cuneus , the paracentral lobule , as well as the superior and inferior parietal cortex , all in both hemispheres , as members of the structural core . We used spectral graph partitioning [34] to identify modules within the weighted high-resolution ( n = 998 ) network as well as within the weighted average regional ( n = 66 ) network . The spectral algorithm provides a means of grouping regions in a way that optimally matches the intrinsic modularity of the network . Optimal modularity for the average regional connectivity matrix was achieved with six clusters ( Figure 6A and Table S1 ) . Four contralaterally matched modules were localized to frontal and temporo-parietal areas of a single hemisphere . The two remaining modules comprised regions of bilateral medial cortex , one centered on the posterior cingulate cortex and another centered on the precuneus and pericalcarine cortex . Recovering the modularity structure using high-resolution connection matrices produced similar results ( unpublished data ) . Knowledge of the distribution of connections within and between modules enabled us to identify provincial hubs ( hub regions that are highly connected within one module ) and connector hubs ( hub regions that link multiple modules ) [35] . Without exception , connector hubs are located within the anterior-posterior medial axis of the cortex ( Figure 6A ) , including bilaterally the rostral and caudal anterior cingulate , the paracentral lobule , and the precuneus . Examination of high-resolution connection matrices shows that the majority of connector hub ROIs is consistently found in posterior medial and parietal cortex ( Figure 6B ) . Provincial hubs are members of the frontal ( e . g . , medioorbitofrontal cortex ) , temporoparietal ( e . g . , bank of the superior temporal sulcus , superior temporal cortex ) or occipital modules ( e . g . , pericalcarine cortex ) . Most core regions , as identified by k-core or s-core decomposition , are members of the two medial modules . When combined into a single “core module , ” over 70% of the between-module edge mass is attached to the core . When modularity detection was applied to more restricted portions of the high-resolution connection datasets , for example the visual and frontal cortex , we were able to recover clusters that were consistent with those found in previous studies based on classical anatomical techniques , or orderings that were suggested based on functional subdivisions . For example , we found , in all five participants , a segregated dorsal and ventral cluster of visual ROIs , corresponding in location and extent to the dorsal and ventral stream of visual cortex [36] . Clustering of frontal cortical ROIs yielded distinct clusters centered on orbital , medial , and lateral frontal cortex ( Figure S4 ) . Regions with elevated betweenness centrality are positioned on a high proportion of short paths within the network [37] . The spatial distribution of ROIs with high betweenness centrality ( Figure 7A and 7B ) shows high centrality for regions of medial cortex such as the precuneus and posterior cingulate cortex , as well as for portions of medial orbitofrontal cortex , inferior and superior parietal cortex , as well as portions of frontal cortex . Figure 7B provides lateral views of the distribution of centrality across the two cerebral hemispheres showing that ROIs with high centrality are widely distributed . For example , ROIs with high centrality are found in the superior and middle frontal gyrus , in the inferior and superior parietal cortex , in addition to in regions of cingulate and medial posterior cortex ( Table S2 ) . Averaged over all ROIs belonging to the same anatomical subdivision and over all participants ( Figure 7C ) , centrality appears highest in the right and left posterior cingulate cortex , as well as other subdivisions of cingulate cortex , and the precuneus and cuneus . Efficiency is related to closeness centrality , in that regions with high efficiency maintain short average path lengths with other regions in the network . We find that the posterior cingulate cortex , the precuneus , and the paracentral lobule are most highly ranked in both cerebral hemispheres ( Figure 7D ) . Five lines of evidence support the robustness and validity of the diffusion imaging and tractography methodology applied in this paper ( see also Text S3 ) . First , within-participant interhemispheric differences in structural connections were modest , since the connection patterns between left and right cortical hemispheres were highly correlated ( r2 = 0 . 94 , p < 10−10 , Figure S2 ) . This indicates methodological consistency within individual scanning sessions . Second , two scans of participant A performed several days apart yielded highly consistent regional connection matrices ( r2 = 0 . 78 , p < 10−10 , Figure S2 ) . Third , we found that after introducing random perturbations of the structural connection matrix that fractionally degraded the connection pattern , our network measures were consistent with those reported for the intact connectivity , indicating that our main conclusions were insensitive to low levels of homogeneous noise potentially introduced in either scanning or tractography ( Figure S6 ) . Fourth , we collected diffusion imaging data from a single hemisphere of macaque cortex to compare connection data obtained by diffusion spectrum imaging to connection data obtained by anatomical tract tracing ( see Text S4 ) . An overlay of structural connectivity derived by DSI and a macaque anatomical connection matrix derived from Cocomac data [20] is shown in Figure S9 . We found that 78 . 9% of all DSI fibers were identified in positions where connections had been identified by tract tracing methods and recorded in Cocomac . A further 15 . 0% were placed in positions where the presence or absence of a pathway is currently unknown . The remaining 6 . 1% were placed in positions where connections had been reported to be absent . Fifth , we performed resting state fMRI in all five participants to derive networks of functional connections and to investigate the degree to which structural connections and functional connections are correlated . Figure 8A shows a map of the functional connections averaged over all five participants plotted for a group of five seed ROIs , all of which were within 10 mm of the Talairach coordinate [–5 −49 40] , which is located within the precuneus and posterior cingulate and was used in a previous study [17] to map the brain's default network ( see also the seed region ‘PCC' in Figure 1 of [17] ) . Consistent with earlier observations ( e . g . , [15 , 17 , 18] ) , we find that this seed region maintains positive functional connections with portions of posterior medial cortex , medial orbitofrontal cortex , and lateral parietal cortex . Figure 8B shows a scatter plot of structural connections and functional connections for the precuneus and the posterior cingulate cortex ( both hemispheres , all participants ) . The plot indicates that the strengths of structural connections as estimated from diffusion imaging are highly predictive of the strengths of functional connections ( r2 = 0 . 53 , p < 10−10 ) . Scatter plots of structural connections and functional connections for all anatomical subregions averaged over all five participants ( Figure 8C ) also reveal significant correlations between their strengths ( r2 = 0 . 62 , p < 10−10 ) . Figure 8B and 8C demonstrate that stronger DSI connections are quantitatively predictive of stronger functional connectivity . The results from this comparison of structural and functional connections support the validity of the DSI-derived structural connection patterns and suggest that structural connections identified by DSI do , in fact , participate in shaping the functional topology of the default network .
Cortical connectivity plays a crucial role in shaping spontaneous and evoked neural dynamics . We mapped structural cortico-cortical pathways in the human cerebral cortex at high spatial resolution and found evidence for the existence of a structural core composed of posterior medial and parietal cortical regions that are densely interconnected and topologically central . We characterize the structural core by mapping network indices , such as node degree , strength , and centrality , and by applying several network analysis methods: extracting a structural backbone , performing core decomposition , retrieving network modules , and classifying hub nodes . While several of these measures are known to be interrelated , each provides a different viewpoint from which to discern major features of the large-scale architecture . Based on their aggregated ranking scores across six network measures ( Table 1 ) , we identified eight anatomical subregions as members of the structural core . These are the posterior cingulate cortex , the precuneus , the cuneus , the paracentral lobule , the isthmus of the cingulate , the banks of the superior temporal sulcus , and the inferior and superior parietal cortex , all of them in both hemispheres . These regions are chosen because they exhibit elevated fiber counts and densities ( node degree and strength ) , they are most resistant to the erosive procedures of k-core and s-core decomposition and they have high topological centrality . The high degree of interhemispheric coupling within the core further suggests that it acts as a single integrated system from which processes in both cortical hemispheres are coordinated . The central structural embedding of posterior medial cortex in the human brain is consistent with a series of physiological findings including high levels of energy consumption and activation at rest [14] and significant deactivation during goal-directed tasks [13 , 14 , 17] . We found a significant positive correlation ( r2 = 0 . 49 , p < 0 . 01 , Figure S5 ) between centrality as reported in this paper and regional cerebral blood flow ( rCBF ) data from an earlier imaging study [14] . Studies of resting state functional networks have reported a high density of strong functional connections in posterior cortex [8] . In such networks , the precuneus was found to exhibit short path length , low clustering , and high centrality [8 , 11] . Activation of the precuneus [38] and of other cortical midline structures [39] has been linked to self-referential processing and consciousness . Reduced metabolic activation in the posterior cingulate cortex [40] , amyloid deposition , and atrophy [41] , as well as impaired task-dependent deactivation in posterior medial cortex , is associated with the onset of Alzheimer-type dementia [42 , 43] . The human default network comprises a set of interacting subsystems linked by hubs [44] . Key components of the default network are the posterior cingulate cortex , the precuneus , the lateral and medial parietal cortex , and the medial prefrontal cortex [12 , 13 , 15 , 17] . Of these areas , medial prefrontal cortex is the only component entirely excluded from the structural core . Our structural results suggest the hypothesis that default network activity may be driven from highly coupled areas of the posterior medial and parietal cortex , which in turn link to other highly connected and central regions , such as the medial orbitofrontal cortex . Consistent with this hypothesis , we found a close correspondence between the strengths of structural connections derived from DSI and functional connections derived from resting state fMRI in the same participants . Additional studies are needed to fully address the relationship between structural and functional connection patterns ( Honey CJ , Sporns O , Cammoun L , Gigandet X , Meuli R , Hagmann P; unpublished data ) . An important issue relates to the comparison of our present network analysis in human cortex to previous analyses carried out on anatomical connection matrices derived from tract-tracing studies in the macaque monkey . Direct comparison is made difficult by differences in spatial resolution ( 998 ROIs in human , 30–70 regions in macaque ) , the incomplete coverage of macaque cortex in most extant datasets , the lack of interhemispheric connections in the macaque , the lack of connection density data in the macaque , and the uncertainty of cross-species homologies between functionally defined brain regions [45] . A previous study focusing on the distribution of highly central hubs in macaque cortex had revealed the existence of connector hubs in some areas of prefrontal and parietal cortex [46] , but was lacking connectional data on significant portions of posterior medial and frontal cortex ( Figure S9 ) . Here , we report ROIs with high centrality in several human cortical subregions , including medial and superior frontal cortex , inferior and superior parietal cortex , as well as cingulate and posterior medial cortex . The structural embedding of core regions within the human brain is consistent with anatomical studies of the connections of the macaque posteromedial cortex , which includes posterior cingulate and medial parietal regions . These regions are reported to have high interconnectivity as well as widespread connection patterns with other parts of the brain [47] . Previous attempts to provide a map of structural connections of the human brain have utilized correlations in cortical gray-matter thickness [48] , as well as diffusion tensor imaging ( DTI ) [28 , 29] . Our approach to mapping human cortical structural connections was DSI followed by computational tractography [26 , 27] . DSI has been shown to be especially sensitive with regard to detecting fiber crossings . In macaque monkey [24] , this method has been shown to produce connection patterns that substantially agree with traditional anatomical tract tracing studies . By extending these results , we found significant overlap between macaque connectivity data derived from DSI and from tract tracing ( Text S4 and Figure S9 ) . A more detailed mapping of the structural core in macaque will require the analysis of high-resolution DSI data from macaque cortex ( Hagmann P , Gigandet X , Meuli R , Kötter R , Sporns O , Wedeen V; unpublished data ) . In human visual cortex , DSI connection patterns are in significant agreement with anatomical reports [27] . Furthermore , the high correlation of structural and functional connections patterns reported in this study , which holds for brain regions that are members of the structural core ( e . g . , the precuneus and posterior cingulate cortex , Figure 8B ) as well as across the entire brain ( Figure 8C ) , supports the validity of the DSI connectivity pattern . While these comparisons suggest that diffusion imaging can yield accurate connection maps , it must be noted that the method may be participant to scanning noise , errors in fiber reconstruction , and systematic detection biases . In particular , smaller fiber tracts and interhemispheric connections toward lateral cortices may be underrepresented given the limited resolution and complexity of the anatomy in the centrum semiovale . We note that our study focuses on a large-scale anatomical feature , the structural core , and that our main conclusions are insensitive to various degradations and manipulations of the original fiber density matrix ( Text S3 , Figure S6–S8 ) . Future improvements in diffusion imaging and tractography , as well as computational network analysis , will no doubt reveal additional features of the connectional anatomy of the human brain . It will be important to include major subcortical regions , such as the thalamus , into future network analyses . Another advance would be to parcellate cortex not on the basis of sulcal and gyral landmarks , but rather on the basis of regularities in functional connections that are observed in individual participants [49 , 50] . Our data provide evidence for the existence of a structural core in human cerebral cortex . This complex of densely connected regions in posterior medial cortex is both spatially and topologically central within the brain . Its anatomical correspondence with regions of high metabolic activity and with some elements of the human default network suggests that the core may be an important structural basis for shaping large-scale brain dynamics . The availability of single-participant structural and functional connection maps now provides the opportunity to investigate interparticipant connectional variability and to relate it to differences in individual functional connectivity and behavior .
|
In the human brain , neural activation patterns are shaped by the underlying structural connections that form a dense network of fiber pathways linking all regions of the cerebral cortex . Using diffusion imaging techniques , which allow the noninvasive mapping of fiber pathways , we constructed connection maps covering the entire cortical surface . Computational analyses of the resulting complex brain network reveal regions of cortex that are highly connected and highly central , forming a structural core of the human brain . Key components of the core are portions of posterior medial cortex that are known to be highly activated at rest , when the brain is not engaged in a cognitively demanding task . Because we were interested in how brain structure relates to brain function , we also recorded brain activation patterns from the same participant group . We found that structural connection patterns and functional interactions between regions of cortex were significantly correlated . Based on our findings , we suggest that the structural core of the brain may have a central role in integrating information across functionally segregated brain regions .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion"
] |
[
"neuroscience"
] |
2008
|
Mapping the Structural Core of Human Cerebral Cortex
|
The extraction of accurate self-motion information from the visual world is a difficult problem that has been solved very efficiently by biological organisms utilizing non-linear processing . Previous bio-inspired models for motion detection based on a correlation mechanism have been dogged by issues that arise from their sensitivity to undesired properties of the image , such as contrast , which vary widely between images . Here we present a model with multiple levels of non-linear dynamic adaptive components based directly on the known or suspected responses of neurons within the visual motion pathway of the fly brain . By testing the model under realistic high-dynamic range conditions we show that the addition of these elements makes the motion detection model robust across a large variety of images , velocities and accelerations . Furthermore the performance of the entire system is more than the incremental improvements offered by the individual components , indicating beneficial non-linear interactions between processing stages . The algorithms underlying the model can be implemented in either digital or analog hardware , including neuromorphic analog VLSI , but defy an analytical solution due to their dynamic non-linear operation . The successful application of this algorithm has applications in the development of miniature autonomous systems in defense and civilian roles , including robotics , miniature unmanned aerial vehicles and collision avoidance sensors .
There are four main classes of motion detection models , namely: ( 1 ) differential methods; ( 2 ) region-based matching; ( 3 ) phase-based and ( 4 ) energy-based techniques ( for review see [11] ) . All four consist of three basic components ( pre-filtering , local motion estimation and integration over the field of view ) but vary markedly in the approaches used to realize these steps . It has been shown that certain energy-based methods are equivalent to correlation-based methods [23] . Given the problems with this class of motion detection it is perhaps surprising that correlation-based models appear to be the ubiquitous form of motion detection in biology . The correlation motion detector model [24] has been used to explain direction selective motion detection in a wide variety of insects , birds and mammals , including humans [25]–[27] . This model involves a non-linear correlation of adjacent spatial samples , with an asymmetric delay filter giving rise to direction selective responses within a local elementary motion detector or EMD [24] , [28] . While the term “EMD” has been used in the context of numerous variant or alternative forms of local motion detector , in insects arrays of correlation-based EMDs are then summed by so-called lobula plate tangential cells ( LPTCs ) to provide measurements of wide-field optical flow or motion of specific targets [29] . By analogy to insect EMDs , our subsequent use of this term thus specifically refers to EMDs based on a local correlation operation . Two key questions arise from the observation that biological motion detectors are of the correlation class . Firstly , assuming biological vision has strong selective pressures to attain a robust and efficient system that is optimized for the task , what are the compelling advantages for this type of motion detector in the context for which they are used ? Secondly , how does the biological system overcome the intrinsic problems with this type of motion detector ? Detectors based on motion correlation have been shown to have significant advantages over gradient models [30] where detector noise is problematic [14] , [31] , e . g . at low contrasts or luminance . Certain features of the correlation EMD make it an extremely useful primitive for biological motion processing , particularly its robustness to both temporal and spatial noise [32] . However , such EMDs are also sensitive to non-motion-related parameters of visual stimuli , and do not by themselves give an unambiguous indication of angular velocity [33] , which is at odds with the apparent ease with which insects analyze this parameter [6] . This is due in large part to the inherent sensitivity of correlation-based EMDs to contrast and spatial structure of local features within moving scenes . This leads to ambiguity in the local response as a function of angular velocity , a phenomenon we term ‘pattern noise’ [33] . However previous work [34] has suggested that static and dynamic non-linearity associated with obvious components of physiological implementation of the model helps overcome some of the inherent limitations of the basic EMD . One contributing factor in the ability of correlation based motion models to accurately encode angular velocity is the relative consistency of the spatial statistics of natural scenes , in spite of structural difference [33] . Natural images tend to possess spatial power spectra with an approximate 1/f2+u characteristic , where f is spatial frequency and u is small ( i . e . a straight line on a log-log scale ) [35] . In addition to similarity between different scenes this characteristic implies a self-similarity in natural imagery at different spatial scales , although residual differences in structure remain . A recent electrophysiological breakthrough was made showing that unlike when using sinusoidal stimuli the LPTCs of insects shown natural images robustly encoded angular velocity independently of the contrast in the scene ( see Figure 3B from [36] ) , a characteristic not predicted by earlier models . This highlights the importance of testing biological motion detection , and models based upon it , under as ‘natural’ conditions as possible . In this paper we provide an explanation for a controversy that has plagued visual science . How is it that biological motion detecting neurons can reliably encode angular velocity across different scenes when electrophysiological evidence shows that they use correlation-based EMDs ? To do this we extend motion models , based directly on the well-studied LPTCs in the insect visual system [34] , [37] , by inclusion of additional dynamic non-linear components that combine to provide a robust estimate for global angular velocity and thus account for hitherto poorly understood properties of the fly LPTCs . The inclusion of these non-linearities , while overcoming many of the problems with motion energy models , is only slightly more complex computationally than the raw EMD model and far more efficient than most other motion detection algorithms . Furthermore , the model works on ‘real-world’ luminance levels , rather than the 8-bit normalized images captured by most current digital systems , making it more easily implemental on low power custom imagers .
Our primary purpose was to develop a model robust against the statistical variance between different scenes in nature , where luminance can vary by over 6 decades or more . In order to capture images for use as stimuli we therefore used a Nikon D-70 digital camera and panoramic tripod head attachment to obtain 14 panoramic images from a variety of urban and natural locations around Adelaide , South Australia in high dynamic range ( HDR ) format . Locations were selected to represent a range of luminance , contrast and spatial clutter conditions . Each panorama was obtained using a series of 12 overlapping panels saved in 16-bit NEF ( raw ) format ( 12-bits of actual dynamic range ) . Each panel was imaged at 3 different exposure levels ( −2 . 0 and +2 . 0EV bracketing ) in order to capture components of the scenes that exceeded the dynamic range of the camera sensor . We used PTGui ( New House Internet Services BV ) to stitch the 12 overlapping images together for each of the three different exposures into full 360 degree panoramas . For each panorama over-saturated pixels were discarded and local luminance was established using a linear gamma curve for the camera luminance values and cosine weightings depending on individual pixel values , i . e . low and high pixel values were assigned low weights while mid range pixels had high weights [38] . We combined the panoramas , with an offset depending on exposure , and converted them to floating point format ( IEEE single precision standard ) at 8000×1600 pixel resolution and full color using custom software written in LabView ( National Instruments ) . Such high resolution was not needed for the detail , as insect optics are too coarse to make use of it , but rather to permit accurate simulation of slow image speeds . The full color HDR images are available for use by interested parties by contacting the authors . Since the motion processing pathway of insects is known to be monochromatic [39] , [40] only the green channel was used as inputs to the motion detection model . All images used in this study , and the associated mean 1D row power spectra , and are shown in Figure 1 . There was a larger roll-off in the higher frequency components of the images than would be expected from the non-idealities of the lens used , caused by stitching artifacts in the generation of the HDR panoramic images . The inevitable time delay between taking each of the panels resulted in small movements of the fine details in the scene ( e . g . leaves ) thus producing a low-pass effect . Furthermore , spatial corrections for the lens distortions and software alignment of the panels to produce panoramas may have reduced the detail in the overlapping panel sections . However , the frequency region in the pass-band of the insect LPTCs modeled in this work ( <1 cycle/degree ) appeared unaffected by this smoothing . Table 1 shows the brightness and contrast for the 14 images illustrated in Figure 1 . Unlike in traditional imagery HDR images vary enormously in mean luminance . In order to compensate for this , and produce contrast metrics that were not dependant on image brightness , a crude global gain control was used ( divide by mean luminance ) . Because image normalization is a major role of the biological photoreceptors this step was omitted in subsequent modeling . Additionally , since defining image contrast is so difficult for natural scenes , we used several different measures to quantify it ( Table 1 ) , based either on the global image statistics , or taking into account the specific receptive field properties of local motion detection and the biological system it is intended to mimic [41] . RMS Contrast ( CRMS ) is the global standard deviation divided by global mean . As a global measure it gives a simple to calculate estimate of the contrast in the whole image and makes no assumptions about directionality . However it can produce large values simply by virtue of the fact many images contain large , yet uniform , bright ( e . g . sky ) and dark ( e . g . ground ) sections that do not necessarily produce strong local motion cues during horizontal ( yaw ) motion . Row Contrast ( CRow ) is the square root of the mean 1D row power spectra . Since the neurons we were mimicking are selective for horizontal ( yaw ) motion having an estimate bias in this direction was appropriate . However this measurement weighted all spatial frequencies equally , a situation that resulted in more influence being given to higher spatial frequencies ( fine detail ) than in either the biological system or our model of it . Effective Row Contrast ( CEffective ) is the square root of the y-intercept in the line of best fit for the mean 1D row power spectra between 0 . 01 and 0 . 5 cycles/degree ( on a log-log scale ) to match the observed spatial coding range for insect vision . Note that 0 . 5 cycles/degree is the Niquist limit for hoverfly spatial sampling , which is approximately 1 degree separation between pixels [42] , while field of view of 100 degrees or more are not uncommon in fly LPTCs [43] . This measure took advantage of the linear ( on a log scale ) relationship between image power and frequency in natural images and also the optical limitations ( spatial sampling ) of the system . While this is a more insect-biased contrast measurement than either of the previous two metrics it was still essentially based on low order image statistics . EMD Contrast ( CEMD ) is the square root of the response of a basic motion correlator model . The images were blurred and optically sampled as for motion detection ( section 3 . 2 ) , then passed through a basic unelaborated EMD model at a single speed , below the velocity maximum of the system . The size of the response to this raw EMD model gave an estimate of image contrast that took into account the exact conditions experienced by the motion detection model . Since the images were high dynamic range , image normalization ( division by global mean ) was performed so this measure of contrast was only influenced by the structure within the environment and not the absolute luminance of the image . Comparison of the differences in contrast by these four measures confirms we achieved our objective in obtaining a set of images that should provide an enormous range in responses for a classical motion energy model tuned to similar spatial sampling . Also , while the different contrast metrics did show some differences they produced similar results , with the average correlation ( r2 ) between the CEMD measure and the other three approximately 0 . 7 . Note that recent electrophysiological work using a comparable set of images ( but low dynamic range ) did show that neurons in the brain of the fly were able to robustly detect angular velocity independent of the scene [36] . The row contrast measurement ( CRow ) gave the smallest range of estimates for image contrast . This was due to the fact it was more heavily biased towards high spatial frequencies than the other measures and frequencies above 5–6 cycles/degree were likely to be influenced by lens distortion and stitching artifacts , hence reducing the contrast of the images . This limitation was addressed when using effective row contrast ( CEffective ) by logarithmically weighting the spatial frequency ( i . e . more weight to lower frequencies ) and limiting it to details larger than 0 . 5 cycles/degree where distortions were minimal . The motion detector used in this paper , shown in Figure 2 , was , at its core , based on the Hassenstein-Reichardt Correlator [24] . However we added a number of elaborations ( Figure 2B ) to help overcome the limitations of this class of model . This more robust model took into account a number of the processing steps known , or presumed , to exist in the fly visual system and is described in the results . All stages of the model were simulated using Matlab ( MathWorks ) . We tested the model under a range of velocities ( 6 points per decade ) from 0 . 01 degrees/s to 1000 degrees/s by rotating the panoramic input images within the virtual environment . Although our modeling used discrete time we utilized a high sample rate relative to the time constants of biological vision in order to approximate continuous time processing . The sample rate of the simulation was 1 kHz for all rotation speeds below 200 degrees/s and 5 kHz for all rotations above 200 degrees/s . The working angular velocity range of the model was below 100 degrees/s , with faster rotations producing increasingly smaller responses . Thus all analysis was limited to the range 0 . 1 degrees/s to 100 degrees/s . We employed linear sub-pixel interpolation during the simulated yaw rotations to ensure an accurate simulation of smooth motion at low velocities . In order to avoid ‘neural after-images’ the initial conditions were set to the mean luminance of the image . Simulations were run for 1050 ms to allow sufficient time for the system to reach steady state . All analysis was based on the average response of the last 50ms . Two parameters were calculated to quantify the output of the model at each angular velocity in terms of image invariance . Coefficient of Variation ( CV ) was defined as the standard deviation of the response of all images at a given rotational speed divided by the mean of the responses and is shown in equation 1 . This parameter was used to show variation ( ambiguity ) in model responses to different images at a specific angular velocity . Lower coefficients of variation meant less variability and a more reproducible result across different images . However , having a low CV does not automatically make a system a good velocity discriminator . Overlapping horizontal lines will have a low CV but will produce the same output value for a range of velocities , making it impossible to distinguish between different image speeds . ( 1 ) Where CVi is the coefficient of variation at point i , σ is the standard deviation of the image responses , is the mean of the model responses to the images and i is the test velocity . CV is expressed as a percentage in the text . Z Score was defined as the difference in the means at the two consecutive velocities divided by the sum of the two consecutive standard deviations then scaled for the number of samples per decade ( i . e . local slope divided by local variability ) and is shown in equation 2 . Unlike CV this parameter represents the ability of the system to discriminate between velocities . A higher Z score meant that the ability to determine the difference between velocities was greater . ( 2 ) Where Zi is the Z score at point i , ppd is the number of test points per decade ( in this case 6 ) , is the mean of the model responses to the images , is the standard deviation of the image responses , i is the test velocity and i−1 is the previous test velocity . All results are given in the form mean±95% confidence interval unless otherwise stated . Global CV or Z score statistics were calculated as the average over the range 0 . 1 to 100 degrees/s . This range was chosen as the maximum closely matches the optimal point seen in biological motion detecting neurons [36] and the minimum is within the accuracy of the animation method used to simulate image motion ( linear interpolation ) . However the model parameters could be altered to create a different coding range if desired .
Each stage of the model depicted in Figure 2B was built up sequentially in order to investigate the contribution of each stage to reliable angular velocity encoding . The response of the model to each of the 14 images , and the effect of adding each of the processing stages into the chain , is shown in Figure 3 . As with all correlation-based EMD models the system produced ambiguous responses , with the same signal value for two different velocities either side of an optimum . However in practice this limitation could be overcome by using the system only within the coding range ( i . e . below the optimum ) . The response of the modeling showed that the inclusion of bioinspired processing components could , in tandem , produce reliable angular velocity coding of visual inputs . However it was important to determine the requirements of this approach from an optical sampling view-point . In order to test the robustness of angular velocity coding for different spatial sample rates , we ran the full model for a range of possible constant sampling optical configurations . The spatial baseline used was the hoverfly ( Eristalis tenax ) , where resolution ( Δφ ) is maximally about 1 degree but can drop off to almost 2 degrees in the periphery [42] . Other types of flies can have even less resolution , e . g . Land [75] reports 2 . 8 degree resolution in house flies ( Musca domestica ) and as low as 5 . 8 degrees in fruit flies ( Drosophila melanogaster ) . Furthermore acceptance angles ( Δρ ) , which can be approximated by a Gaussian blur with a full width at half maximum of 1 . 4 degrees ( standard deviation of 0 . 59 degrees ) in hoverflies [44] , can be as large as 2 . 6 degrees in bees [76] and even 4 degrees in dark adapted locusts [77] . The results of varying Δρ and Δφ are shown in Figure 4 and at no time did we attempt to mimic the variable resolution found to exist across the biological compound eye . In all cases the optimum condition ( producing the largest average Z score ) was a Δφ of 2 degrees with a Δρ of 2 . 8 degrees . When Δρ was kept constant at 2 . 8 degrees ( Figure 4a ) all tested values of Δφ resulted in significantly lower Z scores than the case of 2 degree sampling , except for 1 . 26 degrees ( p<0 . 05 ) . However this solution came at the expense of increased computational effort , with 2 . 5 times more samples ( and hence processing power ) required to realize it . At a fixed Δφ the location of the optimal angular velocity ( corresponding to the largest Z scores ) can be shifted to higher velocities by increasing Δρ . In this system it was found that there was no significant difference in the reliability of angular velocity coding when using Δρ of 2 . 8 , 3 . 53 and 4 . 44 degrees and with no difference in the number of calculations required to produce these results ( assuming the blur was not performed by software convolution of an over sampled system , in which case smaller blur would be less computationally expensive ) then the selection of blur would depend only on the application , with systems with larger blurs tuned for higher velocities . By keeping the Δρ/Δφ equal to 1 . 4 ( Figure 4c ) it was possible to show that the system performance was not significantly different over a range of spatial sampling values ( 1 . 26–2 . 52 degrees ) . As with the constant sampling case increasing the absolute blur moved the optimal point to higher velocities . However in our model computational time increased between these limits by a factor of 4 . In computation , as in biology , greater efficiency might thus make lower spatial sampling rates more desirable . Overall the optical model found to produce the most accurate angular velocity coding was achieved using a Δφ of 2 degrees and a Δρ of 2 . 8 degrees . While the spatial sampling rate is lower than that found in the majority of insects the blur to sampling rate ratio ( Δρ/Δφ ) of 1 . 4 is the same as that seen in bees [76] , [78] and flies [42] , [44]; a ratio that has been predicted as optimal in an information-theoretical sense [79] , [80] . In comparison experiments in primates have shown that the detection of high temporal frequency stimuli is governed by the relatively low resolution magnocellular pathway [81] . Furthermore , throughout the animal kingdom , ranging from invertebrates to vertebrates including humans , the mechanisms underlying motion detection can be attributed to correlational EMD-like processing [82] . Thus there is substantial evidence for a common strategy of low-resolution motion vision in many biological systems . The reason that the optimum spatial sampling rate is so low is because the system was tested under both natural and urban images . Natural scenes have a fractal pattern ( self-similarity at different scales ) that means , in general , more information can be gained by increasing the resolution of the image . In contrast urban scenes ( such as indoor locations ) have a high degree of spatial redundancy ( such as uniformly painted walls ) , where increasing the resolution provides little increase to the overall information gained . Since the EMD is a motion energy model it relies on information change between pixels , if there is little information change there is little energy and hence a small motion signal . Thus increasing the resolution had little or no effect on the velocity consistency of the natural scenes , as they all tended to scale together , but it did cause the urban scenes to produce relatively smaller responses . Hence the ideal spatial resolution of a system may be dependant on the mix of urban and natural environments it needs to operate in . This finding is in direct opposition to the current trends in cameras and computer vision towards support for systems with higher spatial resolution . Unlike in most traditional artificial systems the optimum condition for this system was not a sharply focused image . This is because , due to the low spatial resolution , the system needed to detect sub-pixel motion in order to reliably encode slow velocities . If there were no optical overlap between pixels this would not be possible . However with overlap it was possible to detect small motion changes both within a given pixel and also in the neighboring pixels . Conversely , too much optical blur made the differences between the pixels too small , hence reducing the independence of each sample and resulting in less accurate angular velocity detection . Although it is possible to elicit a motion response by stimulating only two adjacent receptors ( see Figure 2A ) integration over a larger area reduces phase dependant pattern noise [33] . The optimum integration size will be task-dependent . For the special case simulated in this paper of ‘pure yaw’ ( e . g . as needs to be compensated for by a hovering fly ) complete elimination of pattern noise in the time domain can be achieved by sampling across the full 360 degrees of the horizontal visual field . However , what additional spatial summation is required to reduce variability due to differences in spatiotemporal contrast over the vertical extent of the field of view ? To address this , we varied the number of vertical rows averaged in stage 5 to investigate the degree to which spatial integration across a larger receptive field influenced angular velocity coding . In this case , we used the fully elaborated model ( i . e . all stages ) , with 2 degree spatial sampling and 2 . 8 degree optical blur as previous experiments had suggested this to be an optimum optical design due to its compromise between Z score and computational efficiency ( section 3 . 2 ) . All conditions involved a central row around the horizon and an equal number of rows evenly spaced above and below the centre up to a maximum of 72 degrees ( 29 rows due to hexagonal sampling and the inability to use the outer most rows ) . The results are shown in Figure 5 . The maximal average Z score was obtained by using 25 rows , however due to the logarithmic shape of the curve using any number of rows greater than 7 produced results within 10% of the full resolution . The effect of varying the slope of the curve and point of maximal response is shown in Figure 6 . The slope of the model can be modified to fit the desired scope of velocities for a given application . Using a smaller range ( i . e . greater response gain as a function of angular velocity ) made the system more robust against noise , which is more likely to be a problem at low speeds or where the output bandwidth is limited . However in a noise free simulation there was little or no benefit in reducing the working range . The average Z score was 6 . 77±1 . 16 , 7 . 53±1 . 62 and 7 . 15±1 . 59 ( mean±95% confidence interval ) for slope parameters of 0 . 3 , 0 . 5 and 0 . 75 respectively . The maximum ( optimal ) angular velocity of the system could be changed depending on the requirements of the system ( Figure 6b ) . In all cases the variation between images was much greater outside the working range ( above the optimum angular velocity ) . This is because the variability within individual images ( pattern noise ) increased with angular velocity and with a decreasing response to the true angular velocity the signal became swamped with noise . In all of the test conditions described in the paper to date the image angular velocity was constant and the motion detection model was given sufficient time to reach steady-state before the results were taken . However this is not a realistic situation for a motion sensor that would typically be required to produce a reliable response under dynamic conditions . In order to test the model under conditions of variable angular velocity and acceleration a 20 second stimulus was constructed that consisted of variable width periods of constantly ( in the log domain ) increasing and decreasing angular velocity . The exact waveform , and the model response , is shown in Figure 7 . With the median of the coefficient of variation being 7 . 28% the model showed little variation in response to the different scenes , even under rapid accelerations ( ±844 degrees/s2 ) . In fact the median coefficient of variation under constant conditions over the same rotational velocities ( 100−0 . 5 degrees/s ) was 6 . 91% , indicating a decrease in performance reliability of less than 5 . 5% under dynamic conditions where the model was not permitted to reach stead-state . Despite the system only being tested under positive angular velocities there were situations when the model produced negative results indicating that the model got the direction of motion wrong . This aliasing occurred at low velocities following high velocities and at the point where the stimulus went from decreasing to increasing rotational velocities . Moreover , it was more prevalent under larger accelerations . Despite not being explicitly included in the motion model ( see section 3 . 1 stage 5 ) this result is somewhat analogous to the waterfall effect where after the rapid removal of a large motion stimulus motion detecting neurons tend to hyperpolarize . Although flies are capable of extreme angular accelerations during saccades , much larger than the 844 degrees/s2 tested here , [83] it is not clear that the visual system is used for coding under such situations . Some authors ( e . g . [69] ) have made the point that the visual motion response may be deliberately damped to avoid sensitivity to such events in order to avoid instability in the optomotor response , in lieu of a mechanism for saccadic suppression ( as in primate vision ) otherwise required . Other sensory systems likely play a role in encoding high-speed acceleration ( e . g . halteres ) and the visual motion pathway seems deliberately tuned to low speeds in flies ( see [84] ) . In all previously described results the full 360 degrees of horizontal visual space was integrated in order to remove the dependence of the result on the part of the image being analyzed ( pattern noise ) . Although it can be reduced by integrating over smaller areas [85] using a fully panoramic field of view has been show to be the only way to eliminate the periodic responses dependant on image statistics [86] . Behavioral experiments in the fly have shown that they are sensitive to the contrast and orientation of patterns at the level of individual receptor pairs ( i . e . single EMDs ) [87] or when they cover a larger ( non-panoramic ) area of space [88] . While not realistic for the output of a single neuron [89] the outputs of populations of motion sensitive LPTCs combine to give an almost complete panoramic view , as evidenced by the output of neck motor neurons [90] , thus minimising pattern noise by means of spatial integration [87] . In addition to integrating over wider fields of view incorporating saturation and other non-linear processing elements predicted to exist in the biological motion processing pathway can modify and reduce pattern noise when a limited field of view is used [33] , [91] . In order to investigate the role of horizontal field of view on the temporal response of the system we reduced the field of view of the model to 20 degrees of visual space in both the raw ( but normalized for image brightness ) and fully elaborated models . Under such conditions there were two important sources of variability to consider , that between images ( inter-image , as already reported ) and that within images ( intra-image , i . e . the time domain variability of the system in response to a constant input ) . The responses of the model to a constant velocity of 50 degree/s are shown in Figure 8 . With the limited spatial integration and most basic EMD model the average coefficient of variation within images over a full rotation was 66 . 3±27 . 1% ( mean±standard deviation ) and the variation between the 14 image means was 46 . 6% . Thus showing the response was not a constant indicator of individual image velocity , or a good inter-image velocity estimator . Increasing the field of view to 360 degrees dramatically reduced the intra-image variation to 0 . 82±0 . 44% , however as expected it had no effect on the intra-image variation . Using the full model with the limited field of view resulted in an intra-image variation of 23 . 9±5 . 1% , reduced compared to the case with the spatially limited raw model but not to the same extent as with the panoramic view . The inter-image variation in this case was 11 . 7% , much improved over both raw cases . Finally , the full model with full 360 degree field of view produced both the smallest intra-image ( 0 . 48±0 . 14% ) and inter-image ( 2 . 2% ) variations . The difference in the inter-image variations between the full and limited field of view tests was due to both the reduced saturation ( in the spatially limited case ) and the different weighting factors in the non-linear global summation stage . By constructing a model for motion detection based on elements known , or suspected , to be present in the biological system we have shown that accurate and robust detection of global motion can be achieved using a system with very low resolution based on relatively simple mathematical operations . The key to the operation of the model was the way multiple non-linear elements interacted to produce an estimate of angular velocity that was independent of the scene it was viewing . Moreover , the performance of the system as a whole was greater than the linear addition of the individual components taken in isolation . While we have based our model on parameters derived from physiological analysis of the fly motion pathway the model may also be applicable to data from other species . In previous work Ibbotson described ‘velocity tuned’ ( VT ) neurons in the honeybee that appear to differ from our model and fly neurons in having monotonic responses to very high speeds ( 1000 degrees/s ) and apparently less dependent on spatial period of square-wave patterns [92] . While a degree of pattern invariance may result from the adaptive nature of our model , the apparent lack of response roll-off in the Ibbotson data is more difficult to reconcile with the fly data . Interestingly , however , because the bee spatial optimum is much lower than in flies ( coarser spatial sampling ) and the temporal optima much higher ( shorter delay ) [84] , the useful “coding range” ( as referred to in our model description ) is predicted to be shifted to 10 times that in flies ( see [93] ) , where velocity optima for natural scenes are already 200 degrees/s [36] . Since the Ibbotson data set only explored velocities below 1000 degrees/s it is thus likely that patterns were not animated at high enough velocities to see the response roll-off predicted by a correlation-based model ( including our fully elaborated model ) . There is strong evidence that the fly motion pathway processes negative and positive contrasts separately [47] . However , the motion model described here does not incorporate any kind of ‘contrast asymmetry’ . Although several authors have explored whether the motion pathway is fed by separate ‘on’ and ‘off’ pathways ( e . g . [94] , [95] ) , no studies have yet provided conclusive results . Recently we have shown that the separation between ‘on’ and ‘off’ pathways can be a useful primitive in target detection [96] , [97] . While others have shown that contrast separation can be used as a pre-processing stage in a different type of EMD-based model [98] the current model shows it is not a necessity for the accurate detection of wide field angular velocity using correlation-based EMDs . There is a significant push to reduce the complexity of bio-inspired algorithms so they will run in real-time on modern computer platforms [99] . The complexity of the model described in this paper may be too much to realize in a real-time application based on a single serial CPU . However its highly parallel nature and low resolution make it an ideal candidate for implementation in either a FPGA or GPGPU [100] based platform . Furthermore reduced versions have already been produced in analog VLSI [101] and may be suitable for serial digital systems as well [102] where frame rates in excess of 100Hz have already been achieved using standard consumer-level computers . It is also important to note that the computational complexity of an EMD based system can be orders of magnitude less than alternative schemes for computing local velocity vectors in optic flow analysis ( e . g . [103] ) .
|
Building artificial vision systems that work robustly in a variety of environments has been difficult , with systems often only performing well under restricted conditions . In contrast , animal vision operates effectively under extremely variable situations . Many attempts to emulate biological vision have met with limited success , often because multiple seemingly appropriate approximations to neural coding resulted in a compromised system . We have constructed a full model for motion processing in the insect visual pathway incorporating known or suspected elements in as much detail as possible . We have found that it is only once all elements are present that the system performs robustly , with reduction or removal of elements dramatically limiting performance . The implementation of this new algorithm could provide a very useful and robust velocity estimator for artificial navigation systems .
|
[
"Abstract",
"Introduction",
"Methods",
"Results/Discussion"
] |
[
"computer",
"science/natural",
"and",
"synthetic",
"vision",
"neuroscience/natural",
"and",
"synthetic",
"vision",
"neuroscience/sensory",
"systems",
"computational",
"biology/computational",
"neuroscience"
] |
2009
|
Robust Models for Optic Flow Coding in Natural Scenes Inspired by Insect Biology
|
Type IV pili are dynamic cell surface appendages found throughout the bacteria . The ability of these structures to undergo repetitive cycles of extension and retraction underpins their crucial roles in adhesion , motility and natural competence for transformation . In the best-studied systems a dedicated retraction ATPase PilT powers pilus retraction . Curiously , a second presumed retraction ATPase PilU is often encoded immediately downstream of pilT . However , despite the presence of two potential retraction ATPases , pilT deletions lead to a total loss of pilus function , raising the question of why PilU fails to take over . Here , using the DNA-uptake pilus and mannose-sensitive haemagglutinin ( MSHA ) pilus of Vibrio cholerae as model systems , we show that inactivated PilT variants , defective for either ATP-binding or hydrolysis , have unexpected intermediate phenotypes that are PilU-dependent . In addition to demonstrating that PilU can function as a bona fide retraction ATPase , we go on to make the surprising discovery that PilU functions exclusively in a PilT-dependent manner and identify a naturally occurring pandemic V . cholerae PilT variant that renders PilU essential for pilus function . Finally , we show that Pseudomonas aeruginosa PilU also functions as a PilT-dependent retraction ATPase , providing evidence that the functional coupling between PilT and PilU could be a widespread mechanism for optimal pilus retraction .
Type IV pili ( T4P ) are a widespread class of cell surface polymers found throughout the bacteria and archaea [1–3] . In bacteria , they allow cells to physically sense and interact with the environment around them [4] . Consequently they play critical roles in environmental survival and pathogenesis . The type IVa pilus ( T4aP ) machinery responsible for pilus biogenesis is conserved and widely distributed [2 , 5] . Briefly , individual pilin subunits are extracted from the membrane , polymerised into a filament composed primarily of a single major pilin , and guided across the cell envelope layers , before in Gram-negative bacteria exiting the cell surface via an outer membrane pore [6–9] . A unique feature of T4P is their ability to undergo repeated cycles of extension and retraction [10] . This affords considerable functional versatility . For example , cycles of extension , transient attachment , and retraction , powers a form of flagellum-independent motility known as twitching motility [11 , 12] . Retraction also allows cells to sense and adhere to surfaces , to take up DNA during natural competence for transformation and is also exploited as an entry mechanism by some bacteriophages [2] . T4aP dynamics are orchestrated by dedicated extension ( e . g . PilB ) and retraction ( e . g . PilT ) ATPases [13 , 14] . These proteins belong to the Additional Strand Catalytic ‘E’ ( ASCE ) subfamily of AAA+ ATPases [15 , 16] . Both are cytoplasmic proteins that in vitro form oblong hexamers around a central pore [17–22] . They interface with the T4aP machinery via the platform protein PilC , which protrudes from the membrane at the base of the machine and likely sits inside the central pore , and are further clamped in place by PilM [7 , 23 , 24] . The observation that ATP-binding and subsequent hydrolysis leads to a series of conformational changes , which propagate around the pores of PilB and PilT hexamers in opposite directions [20–22 , 25 , 26] , has led to a model whereby the rotation of PilC transduces the action of the ATPase [7 , 22 , 25 , 26] . Accordingly , when PilB is engaged PilC rotates in a clockwise direction , leading to pilus extension . Conversely , when PilT is engaged , its pore rotates in a counter-clockwise direction , leading to pilus retraction . However , the exact details remain unclear and an alternative PilC-gating model has also recently been proposed [10] . Dedicated retraction ATPases are a common feature of T4aP systems [2 , 5] and are thought to be required for generating the force needed to achieve pilus function [10] . Indeed , PilT is the strongest molecular motor known and has been studied extensively [27 , 28] . In a range of different species ( e . g . Acinetobacter baylyi , Dichelobacter nodosus , Pseudomonas aeruginosa , Pseudomonas stutzeri , Neisseria gonorrhoeae , Neisseria meningitidis and Synechocystis sp . PCC6803 ) the deletion of pilT results in a total loss of pilus function [29–37] . Although cells lacking pilT remain piliated , they are unable to mediate twitching motility or DNA-uptake , often exhibit altered surface adherence and are typically hyper-piliated . Notably , a second putative retraction ATPase , PilU , is often encoded directly downstream of pilT within an operon , or elsewhere on the genome , and like PilT has ATPase activity in vitro [13] . Interestingly , Neisseria sp . possess an additional PilT paralogue ( PilT-2 ) , and in some species as many as four PilT paralogues can be present [35 , 38 , 39] . As noted previously by Brown et al . , it can be difficult to unify results across different organisms , in particular due to differences in assay conditions [35] . Nevertheless , compared to PilT , PilU often appears to be dispensable for T4aP function as its deletion produces either subtle phenotypes or defects only in specific functions [31 , 33 , 35 , 37 , 40–42] . One notable exception is that both PilT and PilU are required for twitching underneath agar in several species [32 , 33 , 40] . Likewise , both PilT and PilU are required for pathogenesis in a number of species [33 , 42 , 43] . Vibrio cholerae is an aquatic Gram-negative bacterium responsible for the pandemic human disease cholera . Strains representative of the on-going 7th cholera pandemic utilise two distinct T4aP systems . First , mannose-sensitive haemagglutinin ( MSHA ) pili , which are produced constitutively under laboratory conditions , are required for surface sensing and attachment , are important for the initiation of biofilm formation , and are also a receptor for a filamentous bacteriophage [44–48] . Second , DNA-uptake pili are produced during growth upon chitinous surfaces , which are abundant in the aquatic environment [46 , 49 , 50] . These pili are highly dynamic and retract to take up DNA during natural competence for transformation [50–52] . Moreover , DNA-uptake pili bind to chitinous surfaces and are required for chitin colonization under flow [52] . Notably , PilT function is shared between these two different T4aP systems , and , as in other species , deletion of pilT leads to a total loss of pilus function . This renders MSHA pili defective for surface sensing , attachment , and biofilm formation [47 , 53] . Similarly , DNA-uptake pili loose their rapid dynamics , become hyper-piliated and are unable to mediate transformation [50–52] . Furthermore , DNA-uptake pili can also interact with one another in a sequence specific manner , which in liquid culture results in the auto-aggregation of retraction-deficient cells [52] . In contrast , PilU , which is encoded directly downstream of pilT , has no apparent effect on the function of either system [47 , 50 , 52] . The dispensability of PilU for T4aP function is in itself not surprising , since PilT remains available to mediate retraction . However , an enduring question has been why PilU is unable to take over in the absence of PilT , especially given the similarity between the two proteins . Here , using the two distinct T4aP of V . cholerae as model systems we have used a genetic approach to investigate the relative contributions of PilT and PilU . Unexpectedly , PilT variants engineered to be non-functional had only intermediate phenotypes , revealing that when PilT is inactivated PilU is indeed capable of taking over and functioning as a bona fide retraction ATPase . However , we go to demonstrate that PilU itself is not a separate retraction motor but functions exclusively in a PilT-dependent manner . We provide evidence that this functional coupling is likely a conserved feature of PilU .
We first set out to establish baseline results in assays designed to test pilus function . In the case of the DNA-uptake pilus we have used chitin-independent competence induction to assay natural transformation and auto-aggregation , as these phenotypes provide reliable readouts on the state of pilus retraction [50 , 52 , 54] . Indeed , strains deleted for pilT exhibit a 1000-fold drop in transformation frequency and in liquid culture form large aggregates that rapidly sediment ( Fig 1A and 1B ) . Under natural induction conditions on chitin surfaces , the effect of ΔpilT on transformation was even more prominent , with a 10 , 000-fold drop in transformation frequency , and equivalent to the frequencies observed with cells unable to make pili ( S1 Fig ) . In contrast , deletion of pilU had no effect on transformation and did not promote aggregation ( Fig 1A and 1B ) , in agreement with previous work showing that , under laboratory conditions at least , PilU is not required for DNA-uptake pilus function [50 , 52] . To assay the functionality of the MSHA pilus we have used flagella-dependent swarming motility on soft agar . Indeed , Jones et al . previously demonstrated that the adhesive function of MSHA pili in surface attachment and near-surface motility restricts the ability of cells to swim on soft agar [47 , 55] . Thus , cells unable to make MSHA pili or those with a loss of MSHA function exhibit an enhanced swarming motility phenotype [47] . As shown in Fig 1C our results recapitulate this phenotype , with strains lacking mshA or pilT exhibiting a clear gain of motility phenotype . Again , ΔpilU does not have an obvious effect on MSHA pilus function , as previously described [47] . Finally , in all cases , strains bearing a double ΔpilTU deletion behaved in an identical manner to those carrying the single ΔpilT deletion ( Fig 1A–1C ) . Thus , these assays confirm that PilT is absolutely required for the normal function of both DNA-uptake pili and MSHA pili . By analogy to other T4aP systems , the loss of MSHA pilus function in ΔpilT cells has been presumed to result from the loss of pilus retraction . Indeed , in the initial description that PilT is required for MSHA pilus function , WT cells were reported as having single pili whereas ΔpilT cells had multiple pili , as might be expected [53] . However , Jones et al . , recently reported that ΔpilT cells actually assembled less MSHA pili on the cell surface and did not observe hyper-piliation of ΔpilT cells by electron microscopy [47] . Since this is in contrast to what has been observed with other T4aP systems we therefore sought to confirm that PilT affects MSHA function directly . To do so , we used a previously described MshA cysteine variant ( MshA[T70C] ) , which is the major pilin of the MSHA pilus , and that allows pilus visualisation by cysteine labelling [56 , 57] . As shown in Fig 2A using the surface motility assay to test MSHA pilus function , the MshA[T70C] variant behaves in a manner indistinguishable from that of the parental controls . Cells producing MshA[T70C] were near uniformly peritrichously piliated , with the majority of cells displaying 4–6 short ( c . a . 0 . 5 μm ) pili per cell ( Fig 2B–2D ) . In agreement with the functional data above , the appearance of MSHA pili was not affected by the deletion of pilU ( Fig 2B–2D ) . In sharp contrast , the piliation of ΔpilT cells dropped to less than 20% , with almost all of these cells displaying a single long pilus ( Fig 2B–2D ) . Curiously , the length of these long pili ( 4 . 5 ± 0 . 51 μm; n = 281 pili ) is approximately equivalent to the sum of those displayed on a WT cell . Thus , we speculate that these long pili might result from a ‘runaway’ extension event and that PilT might normally antagonise or otherwise cooperate with the extension ATPase MshE , to limit pilus length . In summary , these data show that PilT is required for proper MSHA pilus biogenesis , but that the MSHA pilus behaves in a manner distinct from that of the DNA-uptake pilus and other T4aP . Moreover , these data further support the idea that the enhanced motility phenotype of ΔpilT is due to the loss of MSHA pilus function [47] and reveal a useful additional readout for PilT function . ASCE ATPases such as PilT and PilU contain four characteristic and highly conserved motifs ( S2 Fig ) [13–16 , 58] . The Walker A motif ( WA; GX4GKS/T ) is required for ATP-binding , whereas the atypical Walker B motif ( WB; Dh4GE; h , hydrophobic ) is required for ATP-hydrolysis and provides a catalytic glutamate to polarise an attacking water molecule ( Fig 1D ) [13 , 14 , 20 , 21] . An Asp box containing acidic residues is involved in magnesium coordination and a His box containing a pair of histidines are both also required for function [13 , 20 , 21] . Thus , to investigate the role of ATPase activity in pilus retraction , we tested the functionality PilT and PilU variants bearing alanine substitutions in the invariant WA box lysine , predicted to disrupt ATP-binding , and in the invariant WB box glutamate , predicted to disrupt ATP-hydrolysis ( Fig 1D ) . As expected , PilU [WA] and [WB] variants did not affect either transformation , aggregation or motility ( Fig 1A–1C ) . In contrast , the PilT [WA] and [WB] variants had only a modest transformation defect , as compared to ΔpilT ( Fig 1A ) . This was unexpected , as both substitutions have previously been validated as abolishing ATPase activity in vitro and function in vivo [13 , 14 , 17 , 59] in a variety of T4aP systems from other species . Indeed , the PilT[WA] variant did not promote auto-aggregation and showed normal motility ( Fig 1B and 1C ) . Consistent with this , DNA-uptake pili and MSHA pili did not appear severely affected , although in the case of DNA-uptake pili we often observed cells with more than one pilus , and MSHA pili appeared to approximately double in length , but were otherwise unperturbed ( Fig 3 ) . In contrast , the PilT[WB] variant displayed additional and more severe phenotypes . Cells producing PilT[WB] were hyper-piliated for DNA-uptake pili and auto-aggregated , albeit at slightly reduced levels compared to ΔpilT ( Fig 1B and Fig 3 ) . Likewise , the PilT[WB] variant showed an enhanced motility phenotype similar to that of ΔpilT ( Fig 1C ) . However , in contrast to the deletion , almost all cells displayed a mixture of long and short MSHA pili ( Fig 3 ) . Taken together these results indicate that substitutions designed to inactivate PilT produce intermediate phenotypes . We hypothesised that partial functional redundancy of PilU could explain these unexpected intermediate phenotypes . If correct , then either deleting pilU or creating WA and WB variants of PilU in these backgrounds should lead to a total loss of pilus function . Indeed , cells producing PilT[WA] or [WB] variants that were co-deleted for pilU showed a total loss of DNA-uptake and MSHA pilus function and produced phenotypes equivalent to that of the ΔpilT mutant ( Fig 1A–1C and Fig 3 ) . Importantly , the combinations with PilU[WA] and [WB] variants also behaved similarly ( Fig 1A–1C ) , indicating that the enduring pilus functionality that we observe in these backgrounds requires PilU ATPase activity . In summary , these data show that PilU can act as a bona fide retraction ATPase in V . cholerae and maintain the functionality of two distinct T4aP systems . Surprisingly , however , the results above suggest that PilU can only support pilus function in the presence of PilT-even when the latter is inactivated . The data so far suggest that PilU is unable to function for pilus retraction in the absence of PilT . However , two distinct models could explain these observations . In the first model , PilU could actually be a separate retraction motor , but because pilU sits in an operon with pilT , the deletion of pilT might inadvertently disrupt PilU production . Indeed , certain pilT deletions have been reported to have such an effect on PilU production in N . gonorrhoeae and P . aeruginosa [41 , 60] . In the second model , PilU would not form a separate retraction motor , but rather would function via a direct interaction with PilT . To distinguish between these possibilities , we first inserted a sequence coding for a 3xFLAG epitope tag at the 3’ end of pilU to create a PilU-3xFLAG fusion . Importantly , as shown in Fig 4A , a band corresponding to PilU-3xFLAG was readily detectable in otherwise WT cells and its levels were not noticeably affected by the deletion of pilT . Furthermore , using the same epitope tag approach we determined that PilT and PilU are both produced at similar levels ( S3 Fig ) , indicating that the inability of PilU to function independently is not simply due to lower abundance relative to PilT . Second , we created arabinose-inducible versions of pilT and pilU ( i . e . TnpilT and TnpilU ) , which were integrated at an ectopic locus , and then tested for their ability to mediate pilus functionality in various backgrounds using swarming motility as a readout for MSHA pilus function and natural transformation on chitin as a readout for DNA-uptake pilus function . As shown in Fig 4B and 4C , ectopic production of PilT was sufficient to fully complement the enhanced motility phenotype and transformation defect of ΔpilT . Conversely , ectopic production of PilU showed no activity in the ΔpilT background ( Fig 4B and 4C ) . As a control , to verify that TnpilU was capable of generating PilU at levels sufficient to mediate pilus function , we re-tested it in backgrounds containing the inactivated PilT[WA] and [WB] variants . In contrast to the results above , ectopic production of PilU in these backgrounds was sufficient to fully complement both motility and transformation phenotypes ( Fig 4B and 4C ) . Notably , PilU production also led to partial rescue of the PilT[WB] motility phenotype ( Fig 4B ) , suggesting that it is likely being overproduced relative the WT situation . Indeed , Western blotting confirmed that TnpilU induction leads to significant PilU overproduction ( S4 Fig ) . Finally , previous work in N . meningitidis and P . aeruginosa has identified a network of interactions between the extension and retraction ATPases [61 , 62] . As shown in Fig 4D , using the same bacterial two-hybrid assay approach , we also detected a similar interaction network between PilB , PilT and PilU . Thus , taken together , these results are consistent with the second model , in which PilU functions via a direct interaction with PilT . The simplest interpretation of these results is that PilT is required to recruit PilU to the pilus machinery . If correct , then the inactivated PilT[WA] and [WB] variants would be expected to behave similarly . However , as detailed above , although both variants retain PilU-dependent function , the PilT[WB] variant has additional stronger defects than that of the [WA] variant . Given that ATP-binding and hydrolysis are linked to a series of conformational changes [20 , 21 , 25 , 26] , one explanation for this difference could be that the [WA] , which is defective in ATP-binding , retains a greater degree of conformational freedom , as compared to the [WB] variant , which is unable to hydrolyse its bound ATP . Thus , these observations suggest that PilT does not simply recruit PilU but that they work together . Given the ability of PilT and PilU to cross-interact , it has been hypothesised that they might intermix to form hetero-hexamers [35 , 38 , 61 , 63] . However , our observation that the PilU[WA] and [WB] variants show no discernable phenotypes , even though we show above that they are indeed non-functional , argues against this idea . To test this more directly , we overproduced PilT and PilU and their respective [WA] and [WB] variants and assayed their ability to interfere with normal PilT function . As shown in S5 Fig PilT[WA] overproduction leads to dominant negative phenotypes in both MSHA pilus and DNA-uptake pilus function . In contrast , none of the PilU variants had an effect , supporting the idea that they do not intermix with PilT . Interestingly , but for reasons that we do not yet understand , the PilT[WB] variant also had no effect . In summary , we propose that PilU is not a separate retraction motor but that it exerts its function through PilT , and is therefore a PilT-dependent retraction ATPase . Finally , previous work with domain swapped chimeras indicated that the N-terminal domains of PilT and PilU might be distinct [64] . Thus , to test if this could explain the PilT dependent behaviour of PilU we created similar chimeras in which the N- and C-terminal domains of PilT and PilU have been swapped i . e . PilTN-PilUC and PilUN-PilTC ( S6A Fig ) . Strikingly , the PilUN-PilTC chimera is able to support near-WT levels of transformation ( S6B Fig ) , suggesting that the C-terminal domain of PilT likely contains specific residues required for mediating pilus function . However , this chimera was unable to support MSHA pilus function ( S6C Fig ) indicating that it is only partially functional and thus further work will be needed in this area . We previously showed that representative 7th pandemic O1 El Tor strains are all equally capable of auto-aggregation via their DNA-uptake pili , but that in liquid culture this is manifest only when pilT is deleted [52] . To extend this work we have now also investigated strain MO10 , which is often used as a representative of the O139 serogroup subtype . In 1992 strains belonging to this serogroup caused a severe cholera epidemic that spread rapidly throughout the Indian subcontinent and , to date , remains the only serogroup other than O1 known to cause pandemic disease [65–69] . Unexpectedly , competence-induced cells of MO10 aggregated in an otherwise unmodified background ( Fig 5A and 5B ) . Indeed , deletion of pilT did not affect the level of aggregation ( Fig 5A ) . Since MO10 carries a defective variant of the quorum-sensing regulator HapR [70] , and transformation is quorum-sensing dependent , we first repaired it to that of the canonical A1552 HapR ( = hapRRep ) . Nonetheless , this phenotype was not dependent on the defective HapR variant since the aggregation phenotype was further enhanced when it was repaired ( Fig 5A and 5B ) . However , in agreement with our previous work , deletion of pilA abolished aggregation ( Fig 5A ) . Thus , these results suggest that MO10 has a defect in DNA-uptake pilus retraction . Inspection of the MO10 pilTU locus revealed a single bp mutation in pilT , resulting in the substitution of a normally invariant arginine ( R206S ) that sits directly adjacent to the Walker B motif ( S2 and S7 Figs ) . R206 forms part of a network of arginine residues that surround the ATP-binding site and that are predicted to be required for proper PilT function [20] . Indeed , compared to the situation in A1552 ( Fig 3 ) , competence-induced cells of MO10 were hyper-piliated for DNA-uptake pili , irrespective of the hapR repair ( Fig 5C ) . Likewise , MSHA pilus biogenesis was also clearly altered ( Fig 5D ) , with the majority of cells having one or more long pili . In both cases , the effects on pili appear similar to the PilT[WB] variant examined above . Taken together , these data suggest that MO10 PilT is defective . The evidence for this assertion is as follows . First , replacing pilT[MO10] with that of A1552 abolished the ability of MO10 to aggregate ( Fig 5A ) and restored the expected configuration of DNA-uptake and MSHA pili ( Fig 5C and 5D ) . Second , recreating the pilT[MO10] mutation in A1552 reduced transformation frequency , allowed competence-induced cells to aggregate , and produced an enhanced motility phenotype ( Fig 5E–5G ) . Third , similar to the PilT[WB] variant described above , natural transformation of cells carrying pilT[MO10] was rendered PilU-dependent ( Fig 5E ) . Finally , deleting pilU in strain MO10 led to a dramatic drop in MSHA pilus biogenesis , such that now only a subpopulation of cells had pili ( Fig 5H and 5I ) , similar to that observed above for retraction-deficient A1552 ( Fig 2B and 2C ) . Moreover , deleting pilU also abolished transformation , unless the MO10 PilT variant was first replaced with that of A1552 ( Fig 5J ) . In summary , MO10 carries a naturally occurring defective PilT variant , which affects multiple processes , and requires PilU to maintain pilus function . Importantly , although multiple O139 genomes have now been sequenced , the MO10 PilT sequence is unique . Moreover , with respect to this variant , the genomes of four contemporary O139 isolates all contain a canonical PilT identical to that of A1552 [71] . Consequently , our data suggest that caution should be used when interpreting pilus related phenotypes in MO10 ( e . g . surface colonisation via MSHA and subsequent biofilm formation ) , as this strain may not be a representative isolate . To test if the functional coupling between PilU and PilT is conserved , we examined the ability of P . aeruginosa PilT and PilU to function in V . cholerae . To do so , we again used ectopically integrated arabinose-inducible constructs , i . e . TnpilT[Pa] and TnpilU[Pa] , and tested their respective abilities to support the functionality of the MSHA pilus during surface motility and the DNA-uptake pilus during natural transformation on chitin . As shown in Fig 6A and 6B , the production of PilT[Pa] was sufficient to completely counteract both the enhanced motility phenotype and the transformation defect that are present in a ΔpilT background . In contrast , PilU[Pa] showed no such activity in a ΔpilT background ( Fig 6A and 6B ) . Strikingly , however , PilU[Pa] production was sufficient to fully counteract the enhanced motility phenotype of the PilT[WA] ΔpilU background ( Fig 6A ) . Likewise , PilU[Pa] also showed the ability to restore natural transformation on chitin in both inactivated PilT backgrounds ( Fig 6B ) , albeit at reduced levels compared to that of PilU[Vc] . In summary , these data demonstrate that PilT[Pa] and PilU[Pa] can support the functions of the MSHA and DNA-uptake pili in V . cholerae . Moreover , these data indicate that although PilT[Pa] can function as an independent retraction ATPase , PilU[Pa] does not , and therefore provide evidence that PilU[Pa] , like PilU[Vc] , also functions as a PilT-dependent retraction ATPase . Here we demonstrate that when PilT is inactivated , via substitutions in the Walker A and B boxes , the PilT paralogue PilU can maintain the near normal functionality of two distinct T4aP systems , namely the DNA-uptake pilus and the MSHA pilus . We show that the ability of PilU to function as a retraction motor is dependent on the presence of functional Walker A and B boxes . Moreover , we establish that PilU is unable to function in the absence of PilT . Previous work in a wide range of bacteria had concluded that PilU does not play a major role in retraction; based mainly on the observation that ΔpilT leads to a total loss of function [29 , 31 , 33 , 35–38 , 40 , 41 , 60] . Additionally , some groups had noticed that PilU was unable to complement ΔpilT , leading Brown et al . , to suggest that PilU might not be a separate retraction motor [31 , 35 , 40] . The demonstration here that PilU acts as a PilT-dependent retraction motor provides direct evidence for this idea and likely explains these initially puzzling observations . Indeed , given our discovery that P . aeruginosa PilU also acts in a PilT-dependent manner , our findings in V . cholerae likely extend to other bacteria , and suggest that the functional coupling between PilU and PilT could be a widespread mechanism for optimising pilus retraction . While this work was under completion [72] , a recent preprint by Chlebek et al . , has also examined PilU function and reached similar conclusions to those reported here [73] . Notably , these authors went on to show that despite the lack of a functional phenotype , the deletion of pilU reduces both the rate and force of DNA-uptake pilus retraction . Moreover , Chlebek et al . , also showed that PilU functions in a similar PilT-dependent manner in A . baylyii . This , taken together with our demonstration that P . aeruginosa PilU also functions in a PilT-dependent manner , adds to the idea that this is likely a conserved feature of PilU . Interestingly , Neisseria sp . carry an additional PilT paralogue , PilT-2 , which in N . gonorrhoeae was shown by Kurre et al . , to modulate T4aP retraction speed and enhance twitching motility [38] . Moreover , it was shown that in N . meningitidis pilT-2 , like pilU , was unable to cross-complement a ΔpilT strain [35] . Thus , an interesting hypothesis to test in future work is that PilT-2 also functions in a PilT-dependent manner . Compared to PilT , the requirement for PilU function appears to be context dependent . For example in P . aeruginosa , PilU is only required for twitching motility underneath agar , which represents a high-friction environment [74] . Similar specialised functions have been reported in other bacteria [31 , 33 , 35 , 37 , 40–42] . The functional coupling between PilU and PilT likely facilitates this versatility by ensuring that the action of the second motor is always coordinated with that of the first , and thus allows the system to deal with a range of loads without the need for separate retraction motors of varying strengths that could interfere with one another . An outstanding question that remains unclear is when PilU acts during the lifecycle of V . cholerae . Our data demonstrate that in strain MO10 , PilU is required to maintain pilus functionality due a naturally occurring defect in PilT . However , acting as a ‘backup’ in this way is unlikely to be a primary role . Indeed , PilU is more likely required under certain environmental scenarios that require increased power e . g . to retract pili bound either to very large fragments of DNA or within the complex gel-like environments found in biofilms . PilU may also be required to power pilus retraction upon surface attachment . Finally , a key question going forward will be to determine how PilU directs its function via PilT . In the simplest model PilU would form an independent hexamer that sits directly beneath PilT . Curiously , PilT contains a highly conserved C-terminal motif ( AIRNLIRE; S2 Fig ) that is absent in PilU [59] . One possibility is that this motif is required either for the interaction with , or else the functional coupling with PilU . In support of this idea , the proposed interaction of the PilT hexamer with PilC , which forms the base of the pilus machinery , orients the protein such that the AIRNLIRE motif faces the cytoplasm [25 , 26] .
The bacterial strains and plasmids used in this study are listed in S1 Table . The V . cholerae strain used throughout this work , A1552 [75] , is a fully-sequenced [76] toxigenic O1 El Tor Inaba strain representative of the on-going 7th cholera pandemic . Bacterial cultures were grown aerobically at 30˚C or 37˚C , as required . Liquid medium used for growing bacterial strains was Lysogeny Broth ( LB-Miller; 10 g/L NaCl , Carl Roth , Switzerland ) and solid medium was LB agar . Ampicillin ( Amp; 100 μg/mL ) , gentamicin ( Gent; 50 μg/mL ) , kanamycin ( Kan; 75 μg/mL ) , streptomycin ( Str; 100 μg/mL ) and rifampicin ( Rif; 100 μg/mL ) were used for selection , as required . To induce expression from the PBAD promoter , cultures were grown in media supplemented with 0 . 2% L-arabinose . Natural transformation of V . cholerae on chitin flakes was done in 0 . 5x DASW ( defined artificial seawater ) , supplemented with vitamins ( MEM , Gibco ) and 50 mM HEPES , as previously described [49] . Thiosulfate citrate bile salts sucrose ( TCBS; Sigma-Aldrich , Switzerland ) agar was used to counter-select for E . coli following bacterial mating . SacB-based counter-selection was done on NaCl-free medium containing 10% sucrose . Molecular cloning was performed using standard methods [77] . All constructs were verified by PCR and Sanger sequencing ( Microsynth AG , Switzerland ) . Genetic engineering of V . cholerae was done using either a combination of natural transformation and FLP-recombination ( TransFLP; [78–80] ) or allelic exchange using bi-parental mating and the counter-selectable plasmid pGP704-Sac28 [46] . Tri-parental mating was used to integrate the mini-Tn7 transposon carrying araC and various PBAD-driven genes into the large chromosome , as previously described [81] . Natural competence was induced in liquid culture using an established chitin-independent approach that relies on the integration of a mini-Tn7 transposon containing an arabinose-inducible copy of tfoX ( i . e . araC , PBAD-tfoX ) , which we refer to as TntfoX . [54] . In the presence of inducer , strains carrying TntfoX turn on the expression of the competence genes according to the known regulatory pathways and upon reaching high cell-density are transformable at levels similar to those seen on chitin [54] . In the absence of inducer strains are non-transformable [54] . A chitin-independent transformation assay was used to assess transformation in strains carrying TntfoX , as previously described [50 , 54] . Briefly , overnight cultures were back-diluted 1:100 and grown 3h at 30˚C with shaking ( 180 rpm ) . Genomic DNA ( derived from strain A1552-lacZ-Kan ) was added to a final concentration of 2 μg/mL and cultures incubated for 5h at 30˚C with shaking ( 180 rpm ) . Transformation frequency was calculated as the number of kanamycin resistant transformants divided by the total number of bacteria . Natural transformation , without and with enrichment , was done as previously described [50] . Otherwise , natural transformation on chitin was performed as previously described , with slight modifications [79] . Chitin flakes were submerged within 950 μL 0 . 5x DASW + HEPES + Vitamins in a 1 . 5 mL eppendorf tube . 50 μL of overnight culture was added , vortexed briefly to mix and incubated standing for 8h at 30˚C , at which point arabinose was added ( final 0 . 2% ) to induce the expression of the various transposon-encoded pilT and pilU . 24h after the initial inoculation , 2 μg genomic DNA ( derived from strain A1552-lacZ-Kan ) was added , mixed by inversion , and incubated at 30˚C for 7h . Bacteria were detached from chitin flakes by vortexing , before serial dilution and enumeration , as described above . DNA-uptake pilus mediated aggregation was quantified as previously described [52] . Briefly , overnight cultures were back-diluted 1:100 in LB + 0 . 2% arabinose and grown at 30˚C for 6h in 14 mL round bottom polystyrene test tubes ( Falcon , Corning ) on a carousel-style rotary wheel ( 40 rpm ) . Aggregates were allowed to settle by standing the tube at RT for 30 min . Aggregation was determined by measuring the optical density at 600 nm ( O . D . 600 ) before and after mechanical disruption ( vortex max speed; ~5 sec ) , which serves to disperse any settled aggregates , and is expressed as the ratio of the O . D . 600 Pre/Post-vortexing . Motility phenotypes were quantified by spotting 2 μL of an overnight culture onto soft LB agar ( 0 . 3% ) plates ( two technical replicates ) . Plates were incubated at RT for 24h prior to photography . The swarming diameter ( cm ) is expressed as the mean of three independent biological repeats . A flagellin-deficient ( ΔflaA ) non-motile strain served as a negative control . Cells were mounted on microscope slides coated with a thin agarose pad ( 1 . 2% w/v in PBS ) , observed using a Zeiss Axio Imager M2 epi-fluorescence microscope , and images analysed and prepared for publication using ImageJ ( http://rsb . info . nih . gov/ij ) , as previously described [52] . Overnight cultures were back-diluted 1:100 and grown at 30˚C for 3 . 5-4h on a rotary wheel , as above , in the absence ( MSHA pili ) and presence ( DNA-uptake pili ) of competence induction , as required . Pilus labelling was performed as previously described [52 , 56] . 100 μL of culture was mixed with AF-488-Mal ( Alexa Fluor 488 C5 Maleimide; Thermo Fisher Scientific; Cat# A10254 ) at a final concentration of 25 μg/mL and incubated at RT for 5 min in the dark . Labelled cells were harvested by centrifugation ( 5000 x g; 1 min ) , washed once with LB , re-suspended in 200 μL LB and imaged immediately . Overnight cultures were back-diluted 1:100 and grown in LB at 30˚C for 6h . Lysates were prepared by suspending harvested cells in an appropriate volume of 2x Laemmli buffer ( 100 μL buffer per O . D . unit ) and boiling at 95˚C for 15 min . Proteins were resolved by SDS-PAGE , blotted onto PVDF membranes using a wet-transfer apparatus and immuno-detection was performed as described previously [54] . Primary anti-FLAG antibodies ( Monoclonal ANTI-FLAG M2 , Sigma; Cat# F1804 ) were used at a dilution of 1:2000 . Anti-Mouse IgG HRP ( Sigma; Cat# A5278 ) diluted 1:5000 was used as a secondary antibody . Sample loading was verified with Direct-Blot HRP anti-E . coli RNA Sigma70 ( BioLegend; Cat# 663205 ) diluted 1:10 , 000 . To study potential interactions between T4aP ATPases a Bacterial Adenylate Cyclase-Based Two-Hybrid ( BACTH ) system was employed ( Euromedex , Cat# EUK001 ) [82] . PCR fragments carrying pilB , pilT and pilU were cloned into each of the BACTH vectors , which were routinely maintained in E . coli XL-10 Gold . Each vector pair was then introduced into chemically competent cells of E . coli BTH101 , in all possible combinations . Empty vectors were used as negative controls . 5 μL of each transformation reaction was spotted onto LB plates containing 50 μg/mL Kan , 100 μg/mL Amp , 0 . 5 mM IPTG ( isopropyl-β-D-thiogalactopyranoside ) , 40 μg/mL X-gal ( 5-Bromo-4-chloro-3-indolyl-β-D-galactopyranoside ) and incubated at 30˚C for 40h prior to photography . All data are representative of the results of three independent biological repeats . All replication attempts were successful . Bar graphs show the mean value , error bars specify the standard deviation .
|
Bacteria interact with their surroundings using micrometre scale polymers called type IV pili . They allow bacteria to physically sense , attach and move on surfaces , and even to take up DNA . Consequently they represent important mechanisms of environmental survival and pathogenesis . The versatility of type IV pili is made possible by dedicated motors that power repeated cycles of extension and retraction . Curiously , although the ATPase PilT is well established as the retraction motor , many species have an additional PilT-like protein called PilU . However , how PilU functions has remained unclear , especially since when PilT is absent it is unable to take over its function . In this work we took a different approach . Instead of deleting pilT , we made inactivated variants and studied the functionality of two distinct types of pili used by the human pathogen Vibrio cholerae to survive in its natural aquatic environment . This allowed us to make the unexpected discovery that PilU is capable of acting as a retraction ATPase , but that it is not an independent motor and instead exerts its function via PilT . Our results suggest this functional coupling between PilT and PilU may be common in other bacteria .
|
[
"Abstract",
"Introduction",
"Results",
"and",
"discussion",
"Materials",
"and",
"methods"
] |
[
"chitin",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"enzymes",
"pathogens",
"vibrio",
"enzymology",
"microbiology",
"pili",
"and",
"fimbriae",
"phosphatases",
"pseudomonas",
"aeruginosa",
"vibrio",
"cholerae",
"materials",
"science",
"genetic",
"elements",
"cellular",
"structures",
"and",
"organelles",
"bacteria",
"macromolecules",
"bacterial",
"pathogens",
"research",
"and",
"analysis",
"methods",
"pseudomonas",
"polymers",
"polymer",
"chemistry",
"sequence",
"analysis",
"genomics",
"sequence",
"alignment",
"bioinformatics",
"proteins",
"medical",
"microbiology",
"microbial",
"pathogens",
"chemistry",
"adenosine",
"triphosphatase",
"pathogen",
"motility",
"biochemistry",
"cell",
"biology",
"virulence",
"factors",
"database",
"and",
"informatics",
"methods",
"genetics",
"transposable",
"elements",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"materials",
"mobile",
"genetic",
"elements",
"organisms"
] |
2019
|
The type IV pilus protein PilU functions as a PilT-dependent retraction ATPase
|
Sporadic adrenocortical carcinomas ( ACC ) are rare endocrine neoplasms with a dismal prognosis . By contrast , benign tumors of the adrenal cortex are common in the general population . Whether benign tumors represent a separate entity or are in fact part of a process of tumor progression ultimately leading to an ACC is still an unresolved issue . To this end , we have developed a mouse model of tumor progression by successively transducing genes altered in adrenocortical tumors into normal adrenocortical cells . The introduction in different orders of the oncogenic allele of Ras ( H-RasG12V ) and the mutant p53DD that disrupts the p53 pathway yielded tumors displaying major differences in histological features , tumorigenicity , and metastatic behavior . Whereas the successive expression of RasG12V and p53DD led to highly malignant tumors with metastatic behavior , reminiscent of those formed after the simultaneous introduction of p53DD and RasG12V , the reverse sequence gave rise only to benign tumors . Microarray profiling revealed that 157 genes related to cancer development and progression were differentially expressed . Of these genes , 40 were up-regulated and 117 were down-regulated in malignant cell populations as compared with benign cell populations . This is the first evidence-based observation that ACC development follows a multistage progression and that the tumor phenotype is directly influenced by the order of acquisition of genetic alterations .
Cancer is an heterogeneous disease . Thus , tumors in different organs display markedly different clinical behaviors and tumors that arise in a single tissue can even exhibit an array of pathologies , ranging from benign adenomas to highly invasive malignancies [1] , [2] . The phenotypic diversity observed in neoplastic tumors has been generally ascribed to the deregulation of multiple signal transduction pathways . However , it remains unclear for most cancers which genetic alterations in a cell or group of cells play a causative role in tumor initiation and progression , and which ones represent bystanders with no selective advantage . Cellular transformation is a process where a normal cell accumulates mutations , as well as epigenetic changes , that activate oncogenes or down-regulate tumor suppressor genes to give rise to a clonal expansion of a subset of transformed cells , independently of both external and internal signals that normally control cell growth [3] . This general concept of multistage tumorigenesis has been demonstrated in the case of the colon cancer where distinct histological stages are directly correlated with genetic alterations in key tumor suppressors and oncogenes . Most adenomatous polyps of the colon , even though they are the precursors of invasive cancer , never actually progress to that stage [4] , [5] . Thus , clinically the occurrence of benign tumors is much more frequent than carcinomas . Sporadic adrenocortical carcinomas ( ACCs ) are rare endocrine neoplasms in humans , notorious for their aggressive behavior , metastatic potential and poor outcome [6] with an estimated worldwide annual incidence of 2 per million in adults [7] . By contrast , benign adrenocortical adenomas ( ACAs ) are rather common in the general population ( present in 2 . 3% of persons at autopsy [8] ) . Moreover , the availability of high-resolution imaging modalities has resulted in an increase of the detection of adrenal masses of which cortical adenomas represent 52% of the surgically resected incidental tumors [9] . Whether ACAs represent a separate entity or are in fact part of a process of tumor progression leading to the emergence of ACC is still an open question , however these numbers are consistent with the hypothesis that only a very small fraction of ACAs may progress to cancer upon the accumulation of additional changes . Although the direct progression of benign adrenocortical tumors to malignant carcinomas has not been clearly demonstrated , the best evidence for a multistage adrenal tumorigenesis comes from two clinical cases where a localized tumor was found to be composed of a benign part surrounded by a malignant area [10] , [11] . Progress into the elucidation of the genes and pathways involved in the pathogenesis of sporadic ACC has been slower than that for most other cancers , largely because of the rarity of this tumor [12] . Nevertheless it has been shown that TP53 somatic mutations are present in about 30% of sporadic adult ACCs and almost never in ACAs [12] whereas , activating mutations of N-Ras gene have been observed in both benign and malignant adrenal cortical neoplasms with an incidence of 12 . 5% [13] . In addition , a suitable animal model for unraveling the role of a given genetic alteration and its possible cooperation with other gene defects in the pathogenesis of the disease has also been lacking . We have previously determined that the sequential introductions of the catalytic subunit of the human telomerase , the simian virus 40 large T ( LT ) and an oncogenic allele of Ras ( H-RasG12V ) suffice to transform normal bovine adrenocortical ( BAC ) cells into tumorigenic cells , when transplanted beneath the kidney capsule of SCID mice [14] . Our data suggested that a limited number of genetic alterations cooperate to transform long-lived mammalian adrenocortical cells . We sought now to develop an in vivo system for the neoplastic transformation of primary BAC cells in order to reveal a minimal set of genes that had been recognized to be altered in human adrenocortical tumors ( ACT ) and to study the influence of each of these genetic alterations taken separately on the pathogenesis of the disease . Here , we report that the simultaneous disruption of the p53 pathway by using a truncated form of the protein , p53DD , which acts as a dominant-negative [15] and the Ras pathway through the stable expression of an active Ras protein ( H-RasG12V ) [16] is sufficient to transform normal BAC cells into a tumorigenic state . Strikingly , we show , using our in vivo tissue reconstruction model , that the order of acquisition of genetic mutations is a critical determinant in the outcome of tumor development and aggressiveness .
The primary BAC cells were infected simultaneously with two replication-defective amphotropic retroviruses based on Moloney murine leukemia virus ( MoMLV ) expressing either p53DD [17] or H-RasG12V [18] , each encoding a drug selection marker , hygromycin and neomycin , respectively . Following infection , the cells were selected for 7 days by supplementing the culture medium with both antibiotics . At the end of the selection process , we established a polyclonal population termed p53DD/RasG12V ( PR ) cells ( Figure 1A ) . Two parallel cultures of primary BAC cells were infected simultaneously either with a retrovirus expressing p53DD ( P ) and a control pLNCX2 ( pL ) retrovirus , or with a retrovirus expressing H-RasG12V ( R ) and a control pBabe-Hygro ( pB ) retrovirus . Thus , we generated two control populations termed P and R , respectively ( Figure 1A ) . We first confirmed that the three polyclonal BAC cell populations transduced with p53DD ( P ) , RasG12V ( R ) or both p53DD and RasG12V ( PR ) expressed the desired transgenes ( Figure 1B ) . Then the cells were assayed for the expression of the desired transgenes by immunoblot analysis . We found that the resulting polyclonal cell populations expressed similar amounts of RasG12V and p53DD proteins ( Figure 1C ) The replication of pL , R and P cells ceased at high density suggesting that these cells were still sensitive to contact inhibition ( Figure 1D ) , a regulatory mechanism through which cells enter a stage of reversible G1 arrest [19] . On the contrary , PR cells did not demonstrate any decrease in cell proliferation at high cell density ( Figure 1D ) and formed multilayered foci in culture ( data not shown ) , a phenomenon commonly associated with malignant transformation [19] . Thus , infection of adrenocortical cells with the combination of p53DD and RasG12V dramatically increased the proliferation rate in comparison to infection with either p53DD or RasG12V alone ( Figure 1D ) . We also studied the proliferation by determining the percentage of Ki-67 positive cells in each cell population . In serum-supplemented medium , each of these populations displayed a similar percentage of cells engaged in the cell cycle ( Figure 1E ) . However , in the absence of serum , only cells transduced with RasG12V and p53DD proliferated independently from extrinsic mitogens . Conversely , pL and P cells required mitogens for their proliferation , whereas R cells exhibited a reduced dependence to growth factors ( Figure 1E ) . Therefore , in cells with defective p53 signaling , oncogenic Ras is able to partially substitute for a mitogenic signal . Finally , the PR cell population and the two control cell populations P and R were seeded in soft agar to assay for anchorage-independent growth . Whereas expression of p53DD was unable to support anchorage-independent growth of adrenocortical cells , cells expressing RasG12V formed small abortive colonies characteristic of transit-amplifying cells ( Table 1 ) . Only the expression of both p53DD and RasG12V led to robust cell growth in soft agar ( Table 1 ) . We thus concluded from these experiments that PR cells were transformed since they displayed all the in vitro characteristics ascribed to tumor cells , i . e . loss of contact inhibition in culture , proliferation in the absence of extrinsic mitogens and in the absence of anchorage . Although mutation or overexpression of genes such as TP53 and Ras are detected in human adrenocortical tumors [20] , it is not known whether these genetic changes must occur in combination to induce tumor growth . We then wondered whether these two genetic changes are sufficient to endow BAC cells with the ability to form adrenocortical carcinoma in vivo . When the PR cells were transplanted beneath the subrenal capsule ( SRC ) of adrenalectomized SCID mice , the rate of tumor formation was 100% and no latency period was observable suggesting that the microenvironment provided by the SRC was favorable to tumor development . This was obvious on sections of the grafts taken on day 8 , 14 and 21 ( Figure S1 ) . The transplanted PR cells produced continuously expanding tumor masses , which first protruded from the site of transplantation and finally destroyed the kidney ( Figure S1A , S1B ) . Eight days after cell transplantation , the xenografts formed a solid tissue structure on the kidney surface ( Figure S1B ) . Invasive characteristics of the tumors were evidenced by day 14 , as they infiltrated the adjacent kidney parenchyma ( Figure S1B , S1C ) and by day 21 , the neighboring tissues , skeletal muscle and adipose tissue ( Figure S1D , S1E ) . Ultimately , by day 35 , the kidney was destroyed ( Figure 2A , 2B ) and the adjacent tissues and organs ( fat , muscle and pancreas ) were invaded ( Figure 2H–2J ) . The tumors were poorly differentiated carcinomas composed of eosinophilic cells with high nuclear grade , high mitotic activity and prominent nucleoli . Necrosis , a typical histopathological marker of malignancy was commonly observed in tumors at day 35 ( Figure 2C ) . Rare apoptotic cells were detected in these tumors ( Figure 2G ) . Examination of Ki-67 expression showed that the cells had a very high proliferation rate which was sustained over time ( Figure S1C; Figure 2F and Table 2 ) . Ras and p53 antibodies showed strong staining throughout the tumor , confirming the long term expression of the transgenes ( Figure S1D , S1E; Figure 2D , 2E ) . Metastases are responsible for 90% of deaths from solid tumors and arise following the spread of cancer cells from the primary site and the formation of new tumors in distant organs . The metastatic process comprises a series of steps including angiogenesis and lymphangiogenesis , which allow the tumor cells to escape the confines of the primary tumor [21] . Moreover , the formation of new blood vessels is an almost absolute requirement in the early development of tumors by providing oxygen and nutrients to the cells . We consistently observed a dense vascular network on the surface of the PR masses that was confirmed on tissue sections by immunofluorescence with an antibody against CD31 ( Figure 2K ) . At day 35 , it was clear from gross appearance that the primary tumor had spread to intraperitoneal organs . Metastatic sites included spleen , diaphragm , abdominal muscle and mesentery ( Figure 2M ) . Most of the metastases grew on the surface of the organs ( Figure 2M ) and the cells forming the metastases were issued from the PR primary tumors as demonstrated by Ras and p53 expression ( Figure 2N , 2O ) . As recent experimental studies and clinicopathological reports suggest that tumor lymphangiogenesis can promote tumor spread through the secretion of lymphangiogenic growth factors [21] , [22] , we investigated the presence of lymphatic vessels in primary tumors by immunofluorescence staining for LYVE-1 ( lymphatic vessel endothelial hyaluronan receptor-1 ) , a specific marker of lymphatic endothelial cells [23] . At the time when metastases were detected , numerous lymphatics were present in the tumor ( Figure 2K ) , suggesting that the spread of tumor cells might occur through de novo development of a lymphatic network . Adrenalectomized animals bearing transplanted cells lived well until they are sacrificed , demonstrating that the cells were functional and produced cortisol ( data not shown ) . The mouse glucocorticoid , corticosterone , was replaced in plasma by the bovine glucocorticoid , cortisol . Cortisol levels gave an unambiguous measure of the function of the transplanted cells because mice lack expression of the steroid-17α-hydroxylase in the adrenal cortex , thus resulting in the biosynthesis of corticosterone rather than cortisol . In addition to cortisol production , we wanted to ascertain that the tumors and metastases were formed from BAC cells transformed by the transduction of p53DD and RasG12V and not from cells such as fibroblasts or other stromal cells possibly contaminating the primary culture . Immunohistochemical analysis of the tumors for the expression of the steroid-converting enzyme 3-β-hydroxy-Δ5-steroid dehydrogenase/isomerase-1 type II ( 3βHSD ) involved in cortisol biosynthesis showed that most if not all tumor cells were positively stained ( Figure 2L ) , consistent with the steroidogenic origin of the initial cells . The formation of malignant tumors by cells expressing only RasG12V and p53DD was unexpected since it has been previously shown that adrenocortical cells require at least the ablation of two tumor suppressor genes ( p53 and Rb through the expression of LT antigen ) and the mutation of one oncogene ( expression of RasG12V ) to undergo transformation [24] . To rule out the possibility that mutations other than the introduction of RasG12V and p53DD occurred during the process of engineering the PR cells , we ensured to keep our population polyclonal , to reduce the period in culture for its generation as short as possible and to produce three independent other PR polyclonal populations from primary BAC cells . Following transplantation , tumors formed in 100% of the injected mice and all tumors were highly neoplastic , poorly differentiated and invaded the adjacent kidney and organs ( data not shown ) . Since it is unlikely that all three polyclonal cell populations have acquired the same mutation that was essential for tumorigenesis , we can conclude that the malignant potential of the cell population is a property of transduced PR cells in general rather than the result of an overgrowth of a minor subpopulation . So , the simultaneous alteration of both p53 and Ras pathways is sufficient to fully transform primary BAC cells and form metastatic ACC when transplanted beneath the SRC of mice . Next , we analyzed the phenotype of tissues resulting from implantation of cells that had been singly transduced with p53DD or RasG12V retroviruses . Following transplantation , P cells formed a small tissue spread between the kidney parenchyma and the capsule ( Figure 3A ) . P tissue presented a uniform structure of regular eosinophilic adrenocortical cells without invasion in the renal parenchyma ( Figure 3B ) . Examination of Ki-67 expression in tissue sections showed that the transplanted cells had a low proliferation rate ( 4 . 6% , Figure 3D; Table 2 ) . In contrast , R cells gave rise to a voluminous tissue with no sign of invasion ( Figure 3A ) . The R cells had formed an heterogeneous benign expanding tumor with an irregular architecture , cellular pleomorphism and some nuclear atypia ( Figure 3B ) . We confirmed by p53 and Ras expression that these phenotypes were due to the expression of the transgenes ( Figure 3C ) . The proliferation rate of the R cells in the transplants was intermediate between that measured in P tissues and in PR tissues ( 18 . 7% , Figure 4D; Table 2 ) . Both P and R tissues were vascularized and interestingly , no lymphatics were detected consistent with the absence of the occurrence of metastases in the recipient mice ( Figure 3E ) . Finally , the P and R tissues were functional as probed by cortisol production ( data not shown ) and by the detection of 3βHSD ( Figure 3F ) , consistent as described above with the adrenal origin of the initial cells . Simultaneous activation of the Ras oncogene and inactivation of the p53 tumor suppressor deregulate the transcriptional programs and confer PR adrenocortical cells a tumorigenic potential when transplanted into mice . However , the exact importance of the order of acquisition of these genetic events on the tumor phenotype has not yet been clearly established . It was proposed that multiple alternative genetic pathways may lead to the formation of a primary tumor and that the characteristics of a tumor may vary as a function of the activated pathway [25] . To address this important issue , primary cells were successively transduced with two retroviruses , allowing between 5 to 7 days of selection after each infection . Thus , singly transduced R cells were infected with p53DD retroviruses ( R+P ) and conversely , singly transduced P cells were infected with RasG12V retroviral particles ( P+R ) ( Figure 4A ) . The controls for each doubly transduced cell were prepared by using the empty vector used to clone the second transgene ( pL or pB ) , establishing two populations termed P+pL and R+pB ( Figure 4A ) . We also generated from primary adrenocortical cells a control cell population transduced with both empty vectors ( pL+pB ) . Following appropriate antibiotic selection , the five resultant polyclonal cell populations were confirmed to express the desired transgenes with the exception of the pL+pB cells which did not express either one ( Figure 4B ) . Importantly , the expression levels of the products of the introduced genes were comparable ( Figure 4C ) . The P+pL and R+pB control cell populations were indistinguishable from their singly infected P and R counterparts in terms of proliferative capacity and in their ability to form steroidogenic tissues once transplanted ( Figure 4D; data not shown ) . Interestingly , the longer time in culture required for efficient selection did not alter the phenotype expressed by these cells , suggesting that no additional genetic changes had appeared during the process of generating these cells . The expression of constitutively active Ras followed by inactivation of wild-type p53 conferred to the cells a proliferation capacity similar to the PR cells ( Figure 4D ) . On the contrary , the reverse order of gene transduction resulted in a marked reduction in the cell proliferation capacity ( Figure 4D ) . To further characterize these cells , cell proliferation was determined by the percentage of Ki-67+ cells in each cell population . The R+P cell population maintained in 0 . 1% FCS-medium displayed a percentage of cells engaged in the cell cycle close to the percentage measured in PR cells , suggesting an almost complete independence from extrinsic mitogens for proliferation ( Figure 4E ) . Conversely , the P+R cells exhibited a reduced proliferation index in the absence of mitogens and thus , a stronger dependence to growth factors ( Figure 4E ) . Therefore , in cells with defective p53 signaling , oncogenic Ras , as a second genetic hit , is not able to totally substitute for a mitogenic signal . When the respective abilities of the R+P and P+R cells to form colonies in soft agar were compared , R+P cells were as efficient as the PR cells whereas the P+R cells were unable to grow , just like the R cells ( Table 1 ) . We then concluded from these in vitro experiments that R+P cells were transformed similarly to the PR cells . The transplantation of R+P cells resulted in highly neoplastic ( Figure 5A , 5B ) , proliferative ( Figure 5C ) and poorly differentiated tumors that invaded the kidney parenchyma and adjacent organs such as muscle , pancreas and adipose tissue ( data not shown ) . At the time of necropsy , the primary tumors spread from the kidney to the spleen , abdominal muscle , intestinal mesentery and diaphragm ( Figure 5G–5J ) . Therefore , R+P cells were as potent as PR cells to induce the formation of a metastatic adrenocortical carcinoma . In contrast , the tissues formed following the transplantation of P+R cells were benign tumors with no signs of kidney parenchyma invasion ( Figure 5A , 5B ) or metastases . The number of Ki-67+ cells was lower than in the R+P tumors ( Figure 5C ) . This milder phenotype in P+R tissues could not be due to a difference in the level of expression of both transgenes as the levels of expression of p53DD and RasG12V detected by immunohistochemistry were very similar in PR , R+P and P+R ( Figure 5D , 5E ) . When the respective ability of R+P and P+R tissues to develop lymphatics was compared , a close correlation with the metastatic potential was found ( Figure 5F ) since P+R tumors were devoid of lymphatic vessels . We next sought to confirm the immunostaining and performed Western analyses of the levels of expression of p53DD and RasG12V proteins in three R+P and P+R transplants . We noticed that the RasG12V and p53DD protein levels appeared slightly lower in the R+P transplants than in the P+R transplants ( Figure 5K ) . The malignant tumors generated by the transplantation of R+P cells is always more heterogenous than is benign counterparts , this is partly due to the invasion of the mouse adjacent tissues ( kidney , pancreas and muscle ) and also by the recruitment of other mouse cells such as stromal cells and fibroblasts . To confirm this , we examined the RasG12V to p53DD ratio and found that the results between both groups yielded similar values ( Figure 5L ) . These results further support the notion that , in this model system , the phenotype difference is mainly due to the acquisition order of the genetic alterations rather than the absolute expression level of the transgenes . The tumor suppression function of p53 relies on its ability to act as a potent sequence-specific transcriptional activator , regulating a program of gene expression . In particular , p53 transactivates expression of the CDK inhibitor p21WAF1/Cip1 . Therefore the lack or decrease of p21 expression will be a marker of the p53 function loss . As shown on Figure S2 , pL and R cells expressing p53 demonstrated a strong p21 expression in almost all nuclei . In contrast , inactivation of p53 resulted in an absence or a marked decrease of p21 expression in P , P+R , R+P tissues and in an intestinal metastasis ( Figure S2 ) . To identify patterns of gene expression associated with fully malignant behavior , we performed a transcriptomic microarray analysis on two P+R and three R+P cell populations . The gene list obtained from a class comparison between benign and malignant populations was filtered to ensure a 1 . 3 fold or higher change in expression level between the two groups . A total of 468 genes ( 499 probe sets ) met this criterion . Among them , 157 were involved in cancer development and progression , 40 were over-expressed and 117 were under-expressed in the R+P cell populations compared to the P+R cell populations ( Table S1 ) . The other genes are involved in diverse cellular functions , including cell-to-cell signaling and interaction , transcription , cell death and metabolism . These results indicated that the distinct orders of genetic hits acquisition leads to distinct transcriptome signatures associated with a very distinctive in vivo phenotype . We further did real-time RT-PCR analysis of 5 genes known to be involved in tumor development and progression: secreted protein , acidic , cysteine-rich ( osteonectin , Sparc ) , leucine-rich repeats and immunoglobulin-like domains 1 ( LRIG1 ) , tumor protein D52 ( TPD52 ) , egl nine homolog 2 ( EglN2 ) and cyclin D1 ( CCND1 ) using RNA from cells used in the microarray analysis . The expected differential expression was confirmed in all genes tested . Sparc and LRIG1 were under-expressed in R+P cells compared with P+R cells ( Figure 6A ) . EglN2 , TPD52 and CCND1 were over-expressed ( Figure 6A ) . To validate these malignant to benign tumor transcriptome changes we performed immunohistochemical analysis on tissues formed after cell transplantation on the 5 genes already validated by QRT-PCR . For the 5 genes tested we confirmed the mRNA differential expression at the protein level: Sparc , LRIG1 proteins were under-expressed while TPD52 , EglN2 and cyclin D1 proteins were over-expressed in the R+P tumors compared to P+R tumors ( Figure 6B–6F ) .
Our present study establishes a new experimental in vivo system for understanding the pathogenesis of ACT . Starting with primary BAC cells , we have successfully transformed such cells into either benign or highly malignant and metastatic tumor-forming cells through the perturbation of one or two signaling pathways , respectively . Evidence has revealed that expression of both H-RasG12V and LT antigen was sufficient to convert primary BAC cells into fully malignant tumor cells [22] . However , the requirement of LT for transformation renders the analysis of these results complicated since LT viral oncoprotein is known to have several functions and to target a wide range of cellular proteins [26] . Moreover , the SV40 proteins are rarely involved in the etiology of human cancers [27] . However , ablation of both mammalian pRB and p53 tumor suppressor pathways has been recently shown to be sufficient to replace the function of LT oncoprotein in the combination of genes to transform normal human cells [28] . The significant progress made in identifying genetic alterations involved in ACT has been used as a platform to discriminate those that we believe might reasonably be involved in multistage tumorigenesis [11] , [12] . The proto-oncogene N-Ras has been found to be mutated in 12 . 5% of ACA and ACC [13] . The involvement of this alteration in ACT development might appear relatively weak; however , it is noteworthy that overexpression of the epidermal growth factor receptor ( EGFR/c-erbB1 ) is present in 3 to 43% of ACA and 76 to 100% of ACC [29]–[31] , and is frequently associated with an overexpression of TGF-α , a natural ligand of EGFR in ACC [30] . As the signal transduced by EGFR tyrosine kinase activity involves Ras proteins among others , it is conceivable that Ras is activated in a larger proportion of ACT . Chronically active wild type Ras might promote tumorigenesis through activation of multiple Ras effectors that contribute to deregulated cell growth , dedifferentiation , and increased survival , migration and invasion . Somatic mutations in the TP53 tumor suppressor gene occur in 25% to 33% of ACC but not in benign tumors [32] , [33] , suggesting that mutations in TP53 participate in tumor progression rather than in initiation . Moreover , TP53 inactivating mutations have been shown to identify a sub-group of ACC patients developing an aggressive tumor associated with a poor outcome [34] . Patients with TP53 mutation also showed a trend towards a shorter survival duration [32] , [33] . Several models of human or swine cell transformation using distinct combinations of mammalian genetic elements have been published in recent years [35]–[38] . Each model addressed the tumorigenic conversion of mammalian cells using various sequences of transgene introduction into target cells . Among these studies , the number of genetic events necessary for full cell transformation appeared to vary from 4 to 6 . It is worth reminding that implantation of transformed cells in the subcutaneous space , as used in these studies , does not allow the survival of cells unless they are fully tumorigenic and , as a consequence , is not adapted for the study of the premalignant stages . In contrast , the SRC site is an advantageous niche for survival and growth of cells , due probably to the immediate access to oxygen and nutrients and to the rapid angiogenic response developed from the dense developed renal vasculature . Indeed , our previous studies on adrenocortical cells showed that injection of normal primary BAC cells under the kidney capsule was an important feature for the successful reconstruction of a functional and vascularized tissue , whereas these same cells did not survive when placed subcutaneously [39] . Moreover , although the kidney represents an ectopic site for adrenocortical cells , the tissues formed beneath the SRC recapitulate histological features characteristic of normal or pathological adrenal cortex [39]–[41] . The ability to form functional tissues from three independent R polyclonal populations contrasted with the report showing that high levels of ectopically activated Ras protein may result in premature senescence [42] . Therefore , there may exist a selection against cells overexpressing RasG12V , leaving a population with moderate expression of the activated oncogene . The level of Ras expression would be then sufficient to activate one or more downstream signaling pathways controlled by Ras , such as the MAP kinase or the PI3 kinase pathways to a level that is essential to disrupt the fine balance between differentiation and proliferation , and to trigger some irreversible changes towards a benign phenotype . To our knowledge , no other study has derived cell cultures from the same batch of initial cells and , after transduction with defined genetic elements in different orders , evaluated their effects on the expressed phenotype in an in vivo experimental model . In this tissue reconstruction model , each singly infected cell population produced a distinctive phenotype which might be thoroughly examined . Significantly , these findings show that malignant progression in ACT might be controlled not only by the acquisition of specific genetic changes but also , and more importantly , with their order of acquisition as we found that only BAC cells expressing Ras and p53DD -in that order- could form carcinomas . This supports the prediction that overexpression or mutation in Ras signaling pathway mediate important early events underlying later tumorigenesis . Sun et al . noted that the order of introduction of RasG12V and LT did not affect the outcome of the transplantation; both cell populations formed very aggressive tumors [24] . One possible explanation is that , LT being such a powerful viral oncoprotein , the requirement for cooperation with RasG12V might be minimal whatever the order of acquisition is . Currently , we do not know how the order of acquisition of genetic alterations impacts the underlying mechanism of cooperation leading to different tumor phenotypes but this is the focus of our in progress investigations . Mouse models have been developed to dissect the interplay between mutant p53 and oncogenic Ras in human cancer and have demonstrated that the presence of both genetic alterations give rise to highly invasive and metastatic tumors associated with a decrease in survival [43]–[45] . According to our results , a model of pancreatic tumor progression involving initiation through K-Ras oncogenic mutation and progression through acquisition of p53 point mutation has been suggested , however the reverse combination was not studied [43] . Two p53 target genes , BTG2 and ATF3 , have been identified as mediators of the ability of wild-type p53 to resist Ras oncogenic transformation through reduced growth rate , anchorage independent growth and tumor formation in mice [46] , [47] . The establishment of malignant phenotype for a transformed cell resides in the acquisition of new biological properties such as cellular motility , which makes possible invasion and metastasis . RhoA , a small GTPase involved in the cell motility process has been found to be negatively regulated by functional p53 and positively regulated by H-RasV12; both signals resulting in a basal level of activated RhoA Upon loss of p53 function , RhoA activation increases , which in turn induces cell motility and disease progression [48] . Recently , microarray analysis of immortalized human fibroblasts transformed by the expression of H-RasG12V and inactivation of p53 identified a NFκB-dependent pro-inflammatory gene signature endowing these cells with an increased tumorigenicity [49] , [50] . In our experimental model of adrenocortical tumorigenesis , BAC cells from the same genetic background acquire two alterations that in turn deregulate major cellular signaling pathways . If the order of the introduced transforming genes was irrelevant to the phenotype , then the tumor formed following transplantation should be identical . This was not the case , however . A transcriptomic analysis using cDNA microarrays has been used to identify the molecular signature that might explain the distinctive in vivo phenotypes . The analysis of P+R and R+P cell populations identified 468 differential genes and among those 157 genes directly involved in cancer development and progression that were differentially expressed between partially and fully transformed cells . Moreover , histochemical validation done on a subset of 5 gene products further confirmed their differential expression in malignant versus benign tumors formed after transplantation of these cells . The 5 genes were chosen on their apparent importance in tumor development and cancer progression . Sparc is an extracellular matrix-associated glycoprotein and a lower Sparc expression is correlated with increased growth , metastatic behavior and reduced apoptosis in multiple cancers [51] . LRIG1 is a transmembrane protein acting as a negative feedback regulator of EGF signaling [52] . Its expression is downregulated in a variety of human cancer supporting the hypothesis that decreased expression of LRIG1 unleashes EGFR signaling , which might contribute to tumorigenesis . To date no data are available on LRIG1 status in ACC where EGFR expression is markedly elevated [29]–[31] . However , LRIG1 expression has also been shown to be up-regulated in prostate cancer and leukemia which highlighted that LRIG1 might act as an oncogene depending on the cellular contexts [53] . Increased TPD52 expression and gene copy number have been reported in breast , prostate and ovarian cancer increasing cell proliferation in vitro and tumorigenicity in mice [54]–[57] . Cyclin D1 overexpression driven by genomic alterations , post-transcriptional regulation , or post-translational protein stabilization is implicated as driving feature in various human tumors [58] . Finally , level of EglN2 , a prolyl hydroxylase , has been shown to be significantly higher in human renal clear cell carcinoma than in normal kidneys [59] . Moreover , inactivation of EglN2 down-regulated Cyclin D1 and cell proliferation in several cancer cell lines [60] . Increased EglN2 expression in R+P cells compared to P+R cells might participate to higher cell proliferation through Cyclin D1 regulation . Thus far , we have employed a rational modeling approach to improve our understanding of the genetic changes leading to the initiation and progression of adrenocortical cancer and to shed some light on the critical importance of the order of genetic alterations for the tumor development . We have focused on Ras and p53 genes because modifications in their expression and/or in their genomic sequence are commonly observed in human ACT . Other genes such as IGF-2 , β-catenin , H19 , p57 , EGFR have also been shown to play a role in adrenal pathogenesis and need to be tested in future studies . Hence , the system that we established will enable us to test the oncogenic potential of these genes singly or in combination , in order to identify those that might truly contribute to adrenocortical tumor development and those that might only be bystanders . We are confident that other gene combinations will lead to ACT development with some specific clinical and histopathological features and it will be then possible to link the genotypes with the tumor phenotype . Finally , the first identification of the minimal combination of two master pathways sufficient to trigger ACC development will help to design new therapeutic options targeting these specific gene products or the downstream targets of their signaling pathways .
Animal use was conducted according to the institutional guidelines and those formulated by the European Community for the Use of Experimental Animals . The animal protocol was approved by the Institutional Animal Care and Use Committee at the Commisariat à l'Energie Atomique . The H-RasG12V cDNA previously inserted into pBabe-Hygro ( a gift from Pr . J . W . Shay; [61] ) , was subcloned into MoMLV derived vector pL ( Clontech ) , downstream of the immediate early cytomegalovirus promoter . A dominant negative p53 fragment , p53DD cloned into pBabe-Hygro was purchased from Addgene ( plasmid 9058 ) [14] . pL-RasG12V ( resistant to neomycin ) and pBabe-Hygro-p53DD ( resistant to hygromycin ) constructs and the corresponding empty retroviral vectors were used to transfect the amphotropic packaging cell line PT67 ( Clontech ) using the Effecten Transfection Reagent ( Life Technologies Invitrogen ) . The cells underwent selection with 400 µg/ml neomycin for 10 days or 50 µg/ml hygromycin for 6 days . Then , the viral supernatant was collected and filtered through a 0 . 45 µm syringe filter to obtain cell-free viruses for adrenocortical cell infection . Primary adrenocortical cells were prepared by dissection and enzymatic digestion of adrenal glands from 2-yr-old steers [62] . They were grown at 37°C under a 5% CO2-95% air atmosphere in DMEM/Ham's F-12 1∶1 supplemented with 10% FCS , 10% horse serum and 1% ( v/v ) UltroSer G ( BioSepra ) ( complete medium ) . When reaching 40–50% confluence , BAC cells were infected by a mix of two retroviral suspensions for 24 hours ( pBabe-Hygro-p53DD/pL-RasG12V or pBabe-Hygro-p53DD/pL or pBabe-Hygro/pL-RasG12V ) . Infected cells were selected with 400 µg/ml G418 for 7 days and 50 µg/ml hygromycin for 5 days to obtain stable cell lines . In different experiments , primary cells were transduced with a single retrovirus , pL-RasG12V or pBabe-Hygro-p53DD and selected with G418 for 7 days or hygromycin for 5 days , respectively . Stably infected cultures were then infected with either pBabe-Hygro or pBabe-Hygro-p53DD , or pL or pL-RasG12V , respectively , and selected with G418 for 7 days and hygromycin for 5 days to obtain stable cell lines . Cells were not grown extensively between the two infections . Primary BAC cells transduced only with the empty vectors pL were used as control cells for the effect of the genes of interest . Three separate adrenocortical primary cell preparations have been used to generate the stably infected cells described above . Gene expression analysis was assessed by RT-PCR . One microgram of total RNA of cultured cells , prepared using the RNAgents Total RNA Isolation System ( Promega ) , were reverse transcribed using the ImProm-II Reverse Transcriptase ( Promega ) with random primers PdN6 ( Life Technologies Invitrogen ) ; after which 2 µL of each reaction were PCR amplified using following primers: 5-ATGACGGAATATAAGCTGGTGGT and 5-TCAGGAGAGCACACACTTGC ( RasG12V ) , 5-AAAGGATGCCCATGCTACAG and 5- TTGCCGGGAAGCTAGAGTAA ( p53DD ) , and 5-GCGGCTATCGTGAAGAACATTG and 5-CCTTGCGTTTGAGAGCAGGG ( RP-L27; Ribosomal Protein-L27 ) . Proliferation was determined by assessing cell number after growth in complete medium , for 7 days ( 5×103 cells of each cell line were initially plated ) . Each day , cells were counted in triplicate using a Coulter Z1 ( Coultronics ) . Proliferation was also assessed by the percentage of Ki-67 positive cells in each cell population . For each cell line , 104 cells were plated in complete medium for 24 hours in 4 well LabTek chamber slide ( Fisher Scientific ) . After deprivation in DMEM/0 . 1% FCS for 48 hours , cells from two wells were transferred in complete medium whereas cells from the two other wells were maintained in 0 . 1% FCS-medium for 24 hours . Cells were then fixed in 4% paraformaldehyde and immunostained with the anti Ki-67 monoclonal antibody ( clone MIB-1; Dako ) . An average of 300 nuclei were counted on each well , n = 4 for each experimental condition . 5×103 cells of individual cell lines were seeded in triplicate in soft agar and the resulting colonies were scored three weeks later . Each experiment was repeated at least once . Both male and female SCID mice , originally purchased from Taconic , were used at an age greater than 6 weeks ( ∼25 g body weight ) in these experiments . Under tribromoethanol anesthesia , mice were adrenalectomized and 2×106 genetically modified adrenocortical cells were transplanted under the kidney capsule [39] , [63] . Six mice were used per polyclonal cell population generated . Post-operative care for the animals consisted of the administration of analgesics and antibiotics in drinking water for 4 days [39] . Animals were killed at various times , from 8 to 35 days after transplantation and subjected to necropsy . All tissues ( adrenocortical transplants and metastases ) were fixed in 4% paraformaldehyde and embedded in paraffin . Microtome sections ( 5 µm thick ) were stained with H&E for histological analysis . Expression of the transduced genes was analyzed by standard immunohistochemistry using the anti-Ras mouse monoclonal antibody ( clone 18; BD Transduction Laboratories ) , the anti-p53 mouse monoclonal antibody ( clone pAb421; Calbiochem ) , and the anti-p21 mouse monoclonal antibody ( clone EA10; Calbiochem ) detected with a biotin-conjugated anti-mouse IgG antibody and an avidin-biotin-peroxidase complex ( Vector Laboratories ) . Sections were counterstained with hematoxylin . The differentiation status and proliferation index of tissues were determined using a rabbit polyclonal anti-3βHSD antibody ( produced in our laboratory ) and the MIB-1 antibody that recognizes the proliferation-associated Ki-67 antigen , respectively . The number of Ki-67 positive cells per 100 BAC cells was designated as the proliferation index . Counting was performed using two non-consecutive tissue sections per tissue sample , selected at random in each group . DNA fragmentation associated with apoptosis was detected by nick end labeling of sections using the TdT-FragEL kit ( TUNEL ) ( Calbiochem ) . Vascular endothelial cells were labeled with a rat monoclonal anti-CD31 antibody ( PECAM-1; BD Biosciences ) and lymphatic endothelial cells were labeled with a goat polyclonal anti-LYVE 1 antibody ( R&D Systems ) , on paraffin-embedded sections ( 5 µm thick ) of tissues fixed in Accustain formalin free fixative ( Sigma-Aldrich ) . Secondary antibodies were Cy3- or FITC-labeled donkey anti-rat IgG or anti-goat IgG respectively . Sections were counterstained with DAPI . Cells were rinsed in PBS and lysed in RIPA buffer ( 10 mM Tris-HCl , pH 7 . 4 , 150 mM NaCl , 1% TritonX-100 , 0 . 5% Na deoxycholate , 0 . 1% SDS ) supplemented with protease inhibitor cocktail ( #P8340; Sigma ) for 10 minutes on ice , scrapped from the culture dish , and cleared with centrifugation in a microfuge tube for 20 min at 4°C . Extracts were analyzed for protein concentration by Bradford assay . Equal amount ( 25 µg ) of total cell protein was separated by 15% SDS-PAGE gel and transferred to nitrocellulose membrane ( Bio-Rad Laboratories ) . Filters were blocked for 1 hr at room temperature in 5% dry milk in Tris Buffered Saline , and incubated with primary antibodies in 5% dry milk in TBS at 4°C overnight . The following primary antibodies were used: anti-Ras mouse monoclonal antibody ( clone 18; BD Transduction Laboratories ) , anti-p53 mouse monoclonal antibody ( clone pAb421; Calbiochem ) , anti-actin mouse monoclonal antibody ( clone AC-15; Sigma-Aldrich ) . After several washes , secondary peroxidase conjugated antibodies ( Thermo Scientific ) were used at a 1∶10000 dilution . The membrane was washed in TBS-5% Tween 20 and the proteins were detected using an enhanced chemiluminescence ( ECL ) detection system ( Amersham ) . For protein isolation from tissues , the xenograft is carefully dissected out from the kidney and the adjacent tissues when possible and finely cut with a razor blade into a mortar with approximately 500 µl of RIPA lysis buffer containing a protease inhibitor cocktail . The tissue was ground well with a pestle , transferred into a microfuge tube and centrifuged for 20 min at 4°C . Protein extracts were then submitted to the same protocol as described above . Quantification was done with Image J image software ( National Institute of Health ) . One microgram of total RNA prepared for the microarray hybridisation was used to generate cDNAs by reverse transcription using the iScript system ( Bio-Rad ) as recommended by the manufacturer . Real-time PCR was performed using Bio-rad CFX96 apparatus and qPCR Master Mix ( Promega ) . The values for the specific genes were normalized to the RP-L27 . Specific primers sequences are provided in Table S2 .
|
A sequential acquisition of genetic events is critical in tumorigenesis , and a dysregulation of a limited set of pathways has been demonstrated as sufficient to progressively transform normal cells into tumor cells in several human tissues . However , in the case of adrenocortical tumorigenesis , whether benign tumors represent a separate entity or are in fact part of a process of tumor progression leading ultimately to an adrenal carcinoma is still an unresolved issue . Moreover , the importance of the order in which these genetic events must occur to transform a cell has not been established . Here , we developed a tissue reconstruction model in mice that allows direct comparison of cells modified with sequential introduction of two genetic events . This revealed that adrenocortical tumor development follows a multistage progression and that the tumor phenotype , including histopathology and metastatic behavior , is directly influenced by the order of acquisition of genetic alterations .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"biology"
] |
2012
|
Acquisition Order of Ras and p53 Gene Alterations Defines Distinct Adrenocortical Tumor Phenotypes
|
Neural crest cells ( NCC ) are a transient migratory cell population that generates diverse cell types such as neurons and glia of the enteric nervous system ( ENS ) . Via an insertional mutation screen for loci affecting NCC development in mice , we identified one line—named TashT—that displays a partially penetrant aganglionic megacolon phenotype in a strong male-biased manner . Interestingly , this phenotype is highly reminiscent of human Hirschsprung’s disease , a neurocristopathy with a still unexplained male sex bias . In contrast to the megacolon phenotype , colonic aganglionosis is almost fully penetrant in homozygous TashT animals . The sex bias in megacolon expressivity can be explained by the fact that the male ENS ends , on average , around a “tipping point” of minimal colonic ganglionosis while the female ENS ends , on average , just beyond it . Detailed analysis of embryonic intestines revealed that aganglionosis in homozygous TashT animals is due to slower migration of enteric NCC . The TashT insertional mutation is localized in a gene desert containing multiple highly conserved elements that exhibit repressive activity in reporter assays . RNAseq analyses and 3C assays revealed that the TashT insertion results , at least in part , in NCC-specific relief of repression of the uncharacterized gene Fam162b; an outcome independently confirmed via transient transgenesis . The transcriptional signature of enteric NCC from homozygous TashT embryos is also characterized by the deregulation of genes encoding members of the most important signaling pathways for ENS formation—Gdnf/Ret and Edn3/Ednrb—and , intriguingly , the downregulation of specific subsets of X-linked genes . In conclusion , this study not only allowed the identification of Fam162b coding and regulatory sequences as novel candidate loci for Hirschsprung’s disease but also provides important new insights into its male sex bias .
The enteric nervous system ( ENS ) is the intrinsic neural network of the gastrointestinal tract . One of its essential roles is to regulate intestinal motility . The ENS is made up of interconnected neural ganglia , themselves composed of neurons and supporting glial cells , forming two main parallel networks: the submucosal plexus and the myenteric plexus . The muscles of the bowel wall that ensure peristaltic movements are controlled by the myenteric plexus of the ENS . The ENS is constructed during embryo development by derivatives of migrating neural crest cells ( NCC ) [1] . These multipotent cells originate from the dorsal part of the neural tube , undergo an epithelial-mesenchymal transition , and migrate extensively to contribute to numerous embryonic structures . Among several different cell types , NCC generate melanocytes as well as all enteric neurons and glia . The developing bowel is mainly colonized by NCC derivatives originating from the vagal region of the neural tube . Such colonization proceeds as a rostro-caudal wave lasting more than 5 days in the mouse ( from embryonic day ( e ) 9 . 0 to 14 . 5 ) , with NCC derivatives first entering the foregut , passing through the midgut ( prospective small intestine ) and finally populating the hindgut ( prospective colon ) either by migrating through the intestinal mesenchyme [2] or by taking a shortcut via the mesentery [3] . The hindgut is the last part of the intestines to be colonized and , therefore , the most susceptible to enteric NCC ( eNCC ) developmental defects . Sacral NCC also contribute to the ENS , but this later contribution is minor and cannot compensate for a lack of vagal NCC [4] . Defects in hindgut colonization by eNCC result in a lack of neural ganglia in the colon , leading to intestinal blockage due to absence of peristalsis . This phenotype is generally described as “aganglionic megacolon” because of the subsequent massive accumulation of fecal material and severe distention of the colon . In humans , this condition is called Hirschsprung's disease ( HSCR ) and , depending on the length of aganglionosis , is clinically subdivided in short-segment ( i . e . restricted to the rectosigmoid colon ) and long-segment forms . Short-segment HSCR represents the vast majority of cases and is more common in males than females , with an overall ratio of ~4:1 [5] . In patients displaying longer segments of aganglionosis , the sex bias is much less pronounced or absent altogether . Although mutations in at least 15 genes have been implicated in HSCR , heritability is unexplained for the majority of cases [6] . HSCR is thus a classic example of a complex disease involving multiple genes , incomplete penetrance , variable expressivity and an intriguing male bias . Most known HSCR-associated genes encode players from two signaling pathways: the GDNF ligand/ RET receptor and EDN3 ligand/ EDNRB receptor pathways . In fact , RET is the main gene associated with HSCR [5] . For both pathways , the receptor is found at the surface of eNCC while the ligand is dynamically secreted from the surrounding mesenchyme during the colonization phase . The role of GDNF/RET and EDN3/EDNRB signaling in ENS formation has been well conserved evolutionarily and studies in animal models have revealed that both pathways profoundly influence every key aspect of eNCC development such as proliferation , survival , differentiation and , most especially , migration [7] . Mouse models have been particularly informative in this regard and multiple lines bearing mutation—either spontaneous or targeted—of genes encoding members of Gdnf/Ret and Edn3/Ednrb pathways have been studied [8–13] . However , the incomplete penetrance and , above all , the male bias observed in human HSCR have been poorly replicated in current animal models [14] . Here , we report the creation of a new insertional mutant mouse model for HSCR that displays , for the first time , incomplete penetrance of the aganglionic megacolon phenotype with a very strong male bias . Extensive characterization of this mouse line and independent validation via transient transgenesis indicate that this outcome is , at least in part , initiated by the specific upregulation of Fam162b in NCC .
The TashT mouse line was obtained from an insertional mutagenesis screen for genes involved in NCC development . This screen was based on the random insertion of a Tyrosinase ( Tyr ) minigene in the FVB/n genetic background . Owing to its specific expression in melanocytes , the Tyr minigene rescues the albino phenotype of FVB/n mice and thus provides a visible—and generally uniform—pigmentation marker for transgenesis [15] . Since melanocytes are derived from NCC , this genetic tool also proved to be a potent indicator of abnormal NCC development via identification of non-uniform pigmentation patterns . This approach yielded several transgenic mutant lines ( to be described elsewhere ) among which TashT ( Tachetée , in French ) was identified due to its variegated pigmentation ( Fig . 1a ) . In addition to the pigmentary anomalies that are similar in both heterozygous and homozygous mutants , a subset of TashTTg/Tg animals suffer from aganglionic megacolon around weaning age . These animals are smaller than unaffected TashTTg/Tg siblings ( reaching about 74% of littermate weight ) and exhibit bowel obstruction concomitant with lack of myenteric ganglia in the distal colon ( Fig . 1a , c ) . The most striking and interesting feature of this lethal phenotype is the fact that the vast majority of affected animals are male ( Fig . 1b ) . In addition , we found via histological analyses that colonic aganglionosis can not only be detected in megacolon-suffering but also in non-affected TashTTg/Tg animals of both sexes ( Fig . 1c ) . To determine whether megacolon expressivity could be correlated with extent of aganglionosis , we undertook a systematic analysis of the length of the colonic ENS for a random group of TashTTg/Tg animals of weaning age via staining of acetylcholinesterase activity ( Fig . 1d ) . In accordance with such correlation , quantification results first revealed that—regardless of the sex of the animal—a minimal length of the colon ( ~ 80%; critical region in Fig . 1e ) has to be properly innervated in FVB/n mice in order to avoid blockage . No intermediate phenotype was noted in the course of these analyses as none of the non-affected animals showed signs of abnormal accumulation of feces in the colon . Furthermore , these results confirmed that most TashTTg/Tg animals exhibit aganglionosis in the distal colon and that males ( ganglionated on average over 79% of total colon length ) are more affected than females ( ganglionated on average over 89% of total colon length ) ( Fig . 1e ) . Importantly , a similar statistically significant difference between male and female animals ( 74% vs 90% ) was also observed in neonatal colons , thus confirming the developmental origin of this male-biased defect . To analyze ENS formation in TashT embryos , we took advantage of the fact that this line bears a second co-injected transgene ( pSRYp[1 . 6kb]-YFP ) that labels migrating NCC derivatives ( including eNCC of vagal and sacral origin ) with YFP fluorescence [16] ( see also S1 Fig ) . Whole-mount detection of fluorescence in dissected stage-matched embryonic intestines revealed that , in comparison to TashTTg/+ or G4-GFP control embryos [17] ( S2 Fig ) , a colonization delay by eNCC of vagal origin is clearly observed for TashTTg/Tg embryos starting around e11 . 0 ( Fig . 2a ) . Although ectopic fluorescence in the cecum and proximal hindgut regions impeded precise visualization of the migration front between e11 . 5 and e14 . 5 , we found that this delay persists through the time at which normal colonization of the digestive tract is overtly completed ( e15 . 5 ) . The presence of scattered fluorescent cells beyond the chains of vagal-derived eNCC at this stage also suggests that the contribution of sacral-derived eNCC to the distal hindgut is not abrogated in TashTTg/Tg embryos . Closer inspection of the migration front at e11 . 0 showed that cell protrusions in the form of filopodia were not overtly affected in TashTTg/Tg eNCC ( S3 Fig ) , suggesting that cells could still investigate , and had the capacity to respond to , their environment . To further characterize the colonization defect , leader cells at the tip of eNCC chains were then visualized during several hours in ex vivo cultures of e11 . 0 intestines from littermate control ( TashTTg/+ ) and mutant ( TashTTg/Tg ) embryos ( Figs . 2b , S4 and S1–S2 Videos ) . It is noteworthy that TashTTg/Tg intestines were selected for these analyses on the basis of the severity of their colonization defect in order to increase the odds of detecting differences between control and mutant eNCC . Using these conditions , we found that the average migration speed , and therefore travel distance , is almost halved in TashTTg/Tg leader eNCC while directionality of migration is not noticeably affected ( Fig . 2c ) . Given the selection bias towards more affected embryos , it is important to bear in mind that this severe effect is most likely not representative of the true average migration speed of mutant eNCC . It should also be noted that the small difference in eNCC location along the intestines—before cecum ( TashTTg/Tg ) vs entry of cecum ( TashTTg/+ ) —cannot account for the observed difference in speed since the eNCC migration front normally displays a fairly stable net speed of ~35 micron/hour ( ~0 . 58 micron/min ) between e10 . 5 and e12 . 5 [18] . Migration of eNCC is dependent on multiple signaling pathways among which GDNF/RET and EDN3/EDNRB are recognized as the most critical regulators [19–22] . To evaluate the status of these signaling pathways in TashT embryos , we made use of a recently described quantitative migration assay using e12 . 5 midgut explants [23] . With control TashTTg/+ or G4-GFP tissues , collagen gels containing either GDNF or EDN3 increased the number of cells coming out of the explants ( Fig . 2d ) . Interestingly , when these ligands were used in combination , a synergistic increase in eNCC numbers invading the collagen was observed . However , little reaction to these extracellular ligands was detected in eNCC derived from TashTTg/Tg embryos ( Fig . 2d ) . In fact , eNCC from homozygous embryos migrated out of intestinal explants even in the absence of exogenous ligands , suggesting they lost some sensitivity to their endogenous microenvironment and distinguished poorly between intestinal tissue and collagen gel . Premature differentiation or a scarcity of progenitor cells can disturb eNCC colonization and lead to incomplete ENS formation [19 , 22 , 24 , 25] . To verify whether these processes might contribute to the TashT phenotype , we performed a detailed marker analysis of embryonic intestines in order to quantify neuronal and glial differentiation as well as proliferation and cell death of eNCC . Quantification of proliferation and cell death in e12 . 5 stage-matched bowel tissues failed to reveal any significant difference between TashTTg/Tg and control TashTTg/+ embryos ( S5 Fig ) . Assessment of neuronal differentiation at the same stage also failed to reveal any significant difference ( S6a-S6b Fig ) . On the other hand , glial differentiation at e15 . 5 was found to be less prevalent in TashTTg/Tg distal bowel tissues , to the benefit of undifferentiated progenitors ( S6c-S6d Fig ) . This , however , is most likely a consequence of the delay in rostro-caudal colonization , and goes contrary to the idea that premature differentiation is the cause of the TashT migration defect . Taken together , these results thus highlight the eNCC migration defect ( concomitant with insensitivity towards GDNF and EDN3 ) as the principal cause of the TashT aganglionosis phenotype . Moreover , since no clear sex bias was observed in these analyses , the relatively modest male bias in phenotype severity is most likely the result of an accumulation of subtle differences during the whole ENS developmental time window . Breeding of the TashT line revealed systematic co-segregation of pigmentation with YFP fluorescence , meaning co-integration of both transgenes into a single autosomal locus which is frequent when an equimolar mixture of each transgene is micro-injected [15] . FISH analysis first allowed a rough estimate of the localization of the transgene insertion site on chromosome 10 at bands B2–B3 ( S7a Fig ) . To obtain a more precise localization , we sequenced the whole genome of a TashTTg/Tg mouse . Mapping of high-throughput paired sequencing reads allowed us to localize the transgenic insertion around the middle of a 3 . 3Mb gene desert between Hace1 and Grik2 ( Fig . 3a , b ) . Twice as many reads were observed in a 26kb non-coding region of chromosome 10B2 , indicating a duplication . Flanking this duplicated region were paired reads with one end mapping to chromosome 10 and the other end mapping to sequences corresponding to either one or the other transgene ( Fig . 3a ) . A schematic representation of the inferred organization of the TashT transgene insertion site is shown at the bottom of Fig . 3a . The number of transgene copies was estimated from the mapping data and the total size of the insertion calculated to be about 700kb . The TashT locus contains several blocks of evolutionary conserved non-coding sequences ( Fig . 3b ) . To evaluate their regulatory potential , we cloned seven ~1kb fragments containing most of these constrained elements ( CE ) —named CE1 to CE7—in a luciferase expression vector bearing a minimal thymidine kinase promoter . Transcriptional activity was then assessed in various cell lines ( Neuro-2a , P19 , Cos7 and NIH 3T3 ) via luciferase assays ( Figs . 3c and S9 ) . Overall , this analysis revealed very strong repression activity in a cell type-independent manner for two of the cloned regions ( CE5 and 6 ) ( Figs . 3c and S9 ) . These luciferase assays thus suggest that the TashT transgenic insertion has disrupted at least one important long-range regulatory element that normally represses expression of a surrounding gene . However , expression of the two most proximal neighboring genes on each side of the gene desert ( Lin28b and Hace1 as well as Grik2 and Ascc3 ) ( Fig . 3b ) was found to be similar in TashTTg/Tg and control G4-GFP e12 . 5 eNCC recovered by FACS ( S8 Fig ) . To cast a wider net and detect transcript variation in an unbiased manner , we sequenced the rRNA-depleted transcriptome of FACS-recovered eNCC from anterior intestinal tissues of stage-matched control ( G4-GFP ) and TashTTg/Tg e12 . 5 embryos . This stage was chosen because it combines ease of intestine dissection with clear presence of the eNCC colonization defect in TashTTg/Tg tissues . It is also important to note that this analysis was restricted to anterior intestinal tissues only ( prospective oesophagus , stomach and small intestine ) because endogenous YFP labelling in the TashT line is strictly specific to eNCC in these regions ( S1b Fig ) . As mentioned above ( see Fig . 2a ) , the TashT line exhibit ectopic YFP fluorescence in non-NCC derivatives in more posterior regions ( prospective cecum and colon ) and , therefore , these regions were excluded from both control and mutant cell preparations . Analysis of RNAseq data revealed that over 1200 coding and non-coding genes are differentially expressed in a significant manner ( ≥2-fold and p <0 . 001 ) in TashTTg/Tg eNCC , among which upregulated and downregulated genes are equally represented ( S1 Dataset and S10c Fig ) . The most deregulated genes ( ≥4-fold ) are listed in Table 1 and this shorter list now indicates a strong enrichment for upregulated genes in TashTTg/Tg eNCC ( 41 downregulated vs 188 upregulated ) . Each of these 229 genes was manually assigned to a category based on function and/or localization of their gene product . Assembling these categories in “super-categories” reveals that TashT affected genes are mostly involved in the control of cell signaling ( categories: Ligand-receptor and Signal transduction ) and gene expression ( category: Transcription factor ) as well as in the composition of , and interaction with , the cell microenvironment ( categories: Extracellular matrix , Cell adhesion as well as Channel and transmembrane transport ) . Especially notable examples within each of these super-categories include , respectively , genes encoding Gdnf and Edn3 ligands , many Hox transcription factors as well as various Collagen members . Another category worth mentioning is the Metabolic pathway which notably contains many players of retinoid signaling ( Aldh1a1 , Aldh1a2 , Aldh1a7 and Rdh10 ) . We verified the expression level of selected genes by semi-quantitative RT-PCR . Our selection criteria included genes known as playing a major role in HSCR ( Table 2 and S10a Fig ) as well as genes located on a sex chromosome whose change in expression could explain the observed male bias ( i . e . upregulated on chromosome Y or downregulated on chromosome X ) ( Table 3 and S10b Fig ) . All genes tested followed the trend set by the RNAseq data . Given that intra-chromosomal regulatory chromatin contacts are much more prevalent than inter-chromosomal ones [26] , the selection criteria for identification of the TashT causative gene was its location relative to the repressive elements disrupted by the transgene insertion . We therefore focused on the handful of upregulated genes located on chromosome 10 ( Table 4 ) in our search for a gene directly regulated by the conserved elements . The closest candidate , Fam162b , is ~3 . 6 Mb telomeric to the conserved elements ( Fig . 3b ) and its overexpression in TashTTg/Tg eNCC was validated via semi-quantitative RT-PCR ( S11a Fig ) . Note that Fam162b mRNA can also be detected in eNCC FACS-sorted from control embryos , indicating its low level expression in normal eNCC populations ( see S1 Dataset ) . Analyses of Fam162b open-reading frame sequences using ExPASy ( www . expasy . org ) and Uniprot ( www . uniprot . org ) resources suggest that this uncharacterized gene encodes a single pass transmembrane protein localized to mitochondria . We subsequently verified that the conserved silencer elements near the transgene insertion site normally interact with the Fam162b locus using chromosome conformation capture ( 3C ) techniques [27] . These analyses first revealed that such interaction can be detected in wild-type whole embryonic intestines using different primer pairs ( Fig . 4a , b ) . We also found that this interaction is cell type-specific as it is detected in NCC-derived Neuro-2a cells but not in undifferentiated P19 cells ( Fig . 4c ) , the same murine embryonic cell lines used in our luciferase assays ( Fig . 3c ) . Importantly , we further found that this interaction is lost in TashTTg/Tg embryonic gut tissues ( Fig . 4b ) . Therefore , these results strongly suggest that the TashT transgenic insertion disrupts intra-chromosomal contacts that normally repress Fam162b expression in NCC ( Fig . 4d ) . To independently validate the candidacy of Fam162b as being involved in the TashT ENS defect , we generated transgenic e15 . 5 embryos specifically overexpressing Fam162b in NCC and analyzed the impact on ENS formation using specific markers . The transgenic construct consisted of a bicistronic cassette containing Fam162b and eGFP coding sequences driven by a previously described Sox10 enhancer ( U3 , also known as MCS4 ) fused to the Hsp68 minimal promoter [28 , 29] . Western blotting confirmed that the cloned Fam162b sequences express a protein of expected size ( S11b Fig ) . Using fluorescence from eGFP as a surrogate marker for transgene expression , we obtained a total of three Fam162b transgenic embryos ( all females ) exhibiting variable levels of transgene expression . In contrast to littermate controls for which bowel tissues were fully colonized , all three transgenic embryos displayed incomplete colonization of the distal hindgut by vagal-derived eNCC and , as a result , this region was found to only contain scattered Sox10-positive eNCC of presumably sacral origin in a way similar to what is observed in TashTTg/Tg tissues ( compare Fig . 4e with Fig . 2a ) . In addition to having used a different genetic background for these experiments ( B6C3 ) , we believe that the fact that all Fam162b overexpressers were female likely explains , at least in part , the more modest effect observed in comparison to TashTTg/Tg embryos . Although we cannot currently exclude the possibility that other gene ( s ) might also be primary target ( s ) of the TashT mutation and might thus also contribute to the TashTTg/Tg phenotype , these transgenesis data support a role for Fam162b overexpression in TashTTg/Tg pathogenesis .
To the best of our knowledge , this study is the first to report a transcriptome analysis of sorted eNCC . Previous screens for genes expressed in eNCC were performed on whole embryonic intestines using DNA microarrays and based on the differential expression between normal and Ret-null aneural tissues [30 , 31] . In addition to analyzing eNCC directly , we took full advantage of the RNAseq technology and included non-coding RNA in our analyses . This resulted in a much more extensive list of genes known to be expressed in eNCC , from a few hundred to several thousand ( S1 and S2 Datasets ) . Importantly , the transcriptional signature of TashTTg/Tg eNCC not only highlighted Fam162b as a potential causative gene , but also provided mechanistic insights into the identified cell migration defect . In this regard , it is noteworthy that several modulated transcripts in TashTTg/Tg eNCC encode components of the extracellular matrix ( ECM ) ( Table 1 and S10c Fig ) . NCC have been suggested to modify their extracellular environment during , or perhaps as a requirement for , migration [32 , 33] . One possibility is thus that the modulated ECM in TashTTg/Tg embryonic intestines is less permissive to cell migration [34 , 35] . Another , not mutually exclusive possibility is that the reduced sensitivity of TashTTg/Tg eNCC to growth factors/chemoattractants normally present in the intestinal ECM underlies the migration defect . In agreement with this , we have found that TashTTg/Tg eNCC have lost their ability to respond to exogenous GDNF and EDN3 in explant assays , with eNCC appearing unable to distinguish between bowel tissue and collagen gel ( Fig . 2d ) . This latter outcome is most likely due to the surprising robust overexpression of both Gdnf and Edn3 by TashTTg/Tg eNCC ( Table 2 and S10a Fig ) . Indeed , given that Gdnf and Edn3 are both normally heavily secreted from the surrounding mesenchyme , additional oversecretion from eNCC is expected to disturb the dosage of ligands these cells normally encounter and/or to disrupt any gradient that might be present . Our RNAseq data also suggest that the lack of responsiveness to GDNF and EDN3 might be due to an overabundance-induced negative feedback on their cognate receptor . This hypothesis is supported by the fact that expression of both Ret and Ednrb is reduced in TashTTg/Tg eNCC ( Table 2 and S10a Fig ) . Interestingly , in the case of Gdnf/Ret signaling , this hypothesis is further supported by the observed overexpression of Lrig1 and Lrig3 in TashTTg/Tg eNCC ( S1 Dataset ) . These genes encode functionally-redundant transmembrane proteins [36] which , as specifically demonstrated for Lrig1 in neuronal cells , can be induced by Gdnf at the transcriptional level and then physically interact with Ret in order to reduce Gdnf binding and tyrosine kinase activity [37] . Regardless of the exact underlying mechanism , a lack of responsiveness to EDN3 might well be responsible for the slower migration of TashTTg/Tg eNCC ( Figs2b-c and S4 ) , as inhibition of Ednrb signaling has been recently shown to primarily affect the speed of eNCC migration [38] . Most of the known and characterized long-range acting regulatory elements do not interact with their nearest promoter but bypass several intervening genes in order to reach their target promoter [39] . Long-range enhancer-promoter interactions are also thought to be more commonly involved in the regulation of tissue-specific genes [40] . Moreover , most studies of intra-chromosomal long-range interactions involve loci up to several hundred kb away from each other , though ultra-long-range events ( several Mb ) between enhancer and promoter are not uncommon [27 , 40 , 41] . Our 3C data are in accordance with these observations and indicate that the interaction between the conserved elements near the TashT transgene insertion site and the Fam162b gene ~3 . 6 Mb away ( Fig . 4b ) falls in the ultra-long-range category . Little is known of the spatiotemporal pattern of Fam162b expression . The evidence to date is in agreement with an expression in neural derivatives: it is weakly expressed in the mouse olfactory bulb ( Allen Brain Atlas , RP_051012_01_G06 ) and expressed in the frontonasal prominence of mouse embryos , proximal to the oral cavity [42]—a tissue heavily populated by cranial NCC . These observations are consistent with the idea that Fam162b is poised for active transcription in neural and/or neural crest cells but kept in check through a repressive mechanism . Using luciferase assays , we demonstrated that a subset of the highly conserved elements near the TashT transgene insertion site has a robust negative regulatory function on transcription in a cell type-independent manner ( Fig . 3c and S9 Fig ) . Given the forced juxtaposition of regulatory elements with the proximal promoter in such assays , the absence of a cell type-specific activity in our analysis thus points to a chromatin conformation-dependent mechanism conferring specificity in vivo . As supported by our 3C data ( Fig . 4c ) , we suggest that an ultra-long-range chromatin loop maintains Fam162b expression at a basal level in a subset of neural-derived cells , including eNCC . Further investigations into the regulation of Fam162b expression will be necessary to confirm and expand this hypothesis . The biological function of the Fam162b gene product is also currently unknown . Characterization of this function will clearly be facilitated by the wealth of information obtained from the RNAseq data as well as by the observed cell migration defect . Robust correlation between extent of aganglionosis and expressivity of the megacolon phenotype in TashTTg/Tg animals allowed us to identify the minimal distance of myenteric innervation necessary for successful movement of luminal content across the colon in FVB/n mice . This critical region ( ~80% of colon length; Fig . 1e ) represents a threshold level beneath which intestinal blockage occurs systematically , and in a sex-independent manner . The fact that the mean length of the ganglionated region of TashTTg/Tg males ends in this critical region while females typically show a more extensive ENS explains the apparent contradiction between the male bias in megacolon expressivity and the near complete penetrance of distal aganglionosis in both sexes . A common defect of eNCC colonization , slightly exaggerated in males , is thus the source of the observed sex bias of the megacolon phenotype in TashTTg/Tg animals . In this regard , it is noteworthy that a similar link between extent of aganglionosis and megacolon expressivity has been previously described in mice bearing the Ednrbs-l allele [43] as well as in Ret+/-::Ednrbs/s compound mutants [14] . Interestingly , although only a very modest male sex bias in megacolon expressivity was reported in this latter case ( ~1 . 5:1 ) , the correlations made with extent of aganglionosis are in agreement with the threshold level revealed by our study . Indeed , full penetrance of megacolon in Ret+/-::Ednrbs/s males was correlated with a mean length of the ganglionated region clearly beneath the threshold ( 59% of colon length ) whereas partial penetrance of megacolon in Ret+/-::Ednrbs/s females was correlated with a mean length of the ganglionated region much closer to the threshold ( 72% ) [14] . However , as evidenced by the fact that Sox10Dom mutants on a C57BL/6J—C3HeB/FeJ mixed background display a shorter aganglionic zone leading to megacolon ( ~10% ) , it should also be noted that position of the threshold level may vary as a function of the genetic background [44] . The aganglionic megacolon of TashTTg/Tg animals share striking similarities with both the variable penetrance and male sex bias of short-segment HSCR , the most common form of the disease ( ~80% of cases ) . The threshold level identified with the TashT line is also in accordance with the fact that virtually no sex bias is observed in long-segment HSCR . Our analysis of the TashT line thus provides useful insights into the ontogeny of aganglionic colon and the origin of the sex bias , and shows that , though perhaps suffering from chronic constipation , TashTTg/Tg mice are nevertheless able to pass intestinal material when more than 4/5 of their colon is innervated . Apart from Ret+/-::Ednrbs/s compound mutants , it is interesting to note that a male bias in the extent of aganglionosis—but not megacolon expressivity—has also been reported in other mouse and/or rat models and in each case implicated a mutation in either Ret or Ednrb [13 , 45 , 46] . Taken together with our data showing deregulated Ret and Ednrb signaling in TashTTg/Tg animals as well as with the previous description of a RET non-coding mutation that is twice as frequently transmitted in boys than in girls [47] , these observations strongly suggest that both pathways are involved in the regulation of expression of a still undefined sex chromosome-linked gene with critical function in the developing ENS . We reasoned that exaggerated defects in males could arise from overexpression of male-specific genetic material ( upregulated genes on Y chromosome ) or from a deficiency in the expression of genes present as single alleles ( downregulated genes on X chromosome ) . Another interpretation of this second possibility is that females are protected through biallelic expression of some X chromosome genes , provided they escape X-inactivation [48–51] . Analysis of our transcriptome dataset revealed that expression of Y-linked genes in eNCC is limited to a group of four genes ( Kdm5d , Eif2s3y , Uty and Ddx3y ) ( S2 Dataset ) . Of these clustered genes—also known to be expressed in the developing brain [52 , 53]– , Ddx3y is the only one that approaches significant upregulation in TashTTg/Tg eNCC ( 1 . 5-fold; edgeR p = 0 . 0011; DESeq p = 0 . 0055 ) . In marked contrast , multiple X-linked genes were found to be significantly downregulated in TashTTg/Tg eNCC , including Dcx—a previously suggested potential HSCR susceptibility locus ( Table 3 ) [30] . As X-inactivation escapees tend to be found in clusters [48] , other interesting candidates include a group of genes—in the vicinity of Dcx—that contains Bex1 and the Plp1-Rab9b gene pair ( Table 3 ) . It is noteworthy that the candidacy of these genes is also supported by human cases with Xq22 microdeletions encompassing them . Indeed , while such cases are assumed to be embryonic lethal in males , female patients suffer from Pelizaeus-Merzbacher-like disease with symptoms of gastrointestinal motility problems including constipation [54] .
Work with mice was performed in accordance with the guidelines of the Canadian Council on Animal Care ( CCAC ) and approved by the relevant institutional committee ( Comité institutionnel de protection des animaux; CIPA reference #650 ) of University of Quebec at Montreal ( UQAM ) . Mice were euthanized by gradual-fill carbon dioxide ( CO2 ) gas preceded by isoflurane anesthesia . TashT transgenic mice and Fam162b transgenic embryos were generated via standard pronuclear microinjection [55] , using embryos derived from FVB/n albino and B6C3 mice , respectively . For the TashT line , two transgenes were co-injected at equimolar ratio: a Tyrosinase minigene to allow visual identification of transgenic animals via rescue of pigmentation [15] and a pSRYp[1 . 6kb]-YFP construct that provides fluorescent marking of migrating NCC in the developing embryo [16] . The previously described Gata4p[5kb]-GFP line ( G4-GFP ) was used as wild-type control [17] . Mice were mated overnight and noon on the day a vaginal plug was observed was designated as embryonic day ( e ) 0 . 5 . For Fam162b transient transgenics , a transgene carrying the PCR-amplified Fam162b open reading frame under the control of the Hsp68 minimal promoter [56] and a NCC-specific Sox10 enhancer ( U3 , also known as MCS4 ) [28 , 29] was used . In order to provide a positive control for transgene expression , an IRES-GFP cassette ( pIRES2-EGFP , Clontech ) was also included immediately downstream of the Fam162b ORF , creating a bicistronic message . Fifteen days after microinjection , foster mothers were sacrificed , embryos were collected , individually analyzed for GFP and immunostained for Sox10 and βIII-Tubulin ( see immunofluorescence section below ) . Antisense and sense digoxigenin-labelled RNA probes were synthesized using a DIG transcription kit ( Roche ) . Mouse spleen lymphocytes were collected and metaphase slides prepared using standard cytogenetic protocols [57 , 58] . Slides were aged at room temperature for 7 days , and then GTG banded , again using standard protocols [55] . Slides were scanned at 100X using a Nikon Eclipse E800 microscope , and representative metaphases were photographed at 1000X using a Nikon DXM1200 digital camera and SimplePCI software . Slides were de-stained using CitriSolv ( Fisher ) for 10–15 minutes , fixed for 1 minute in a 1% paraformaldehyde ( PFA ) solution in 2X SSC then dehydrated in graded ethanol washes and stored for future use in 100% ethanol . Slides were air dried then incubated in a denaturation solution for 5 minutes at 73°C , then again dehydrated in graded ethanol washes and stored in 100% ethanol . Probe was generated using a digoxigenin-labeled Tyr minigene DNA fragment [57] . Just prior to adding the denatured FISH probe , slides were air dried . Five to 10 μl of denatured probe was added per slide , which was then incubated overnight at 37° under a coverslip in a humidified dark atmosphere . Detection was performed using a rhodamine-conjugated anti-DIG antibody ( Roche ) , following the manufacturer’s instructions . Slides were counterstained with DAPI ( Sigma ) and antifade ( P-phenylenediamine; Sigma ) . Previously photographed G-banded metaphases were re-located using epifluorescence , rephotographed , and compared to FISH images . Intestines were dissected from e12 . 5 G4-GFP and TashT embryos without their posterior end ( prospective cecum and colon ) because , in contrast to more anterior regions in which only eNCC are fluorescently-labelled ( see S1b Fig ) , TashT intestines also contain high amounts of YFP-positive mesenchymal cells that are not derived from NCC in these regions ( see Figs . 2a and S4 ) . A MoFlo XDP ( Beckman Coulter ) cell sorter was used to collect GFP- or YFP-positive single viable cells from the remaining dissociated intestinal tissue . Dissociation was carried out at 37°C in EMEM containing collagenase ( 0 . 4 mg/ml; Sigma C2674 ) , dispase II ( 1 . 3 mg/ml; Life Technologies 17105–041 ) and DNAse I ( 0 . 5 mg/ml; Sigma DN25 ) . Whole genome and transcriptome library generation and sequencing was performed by McGill University and Génome Québec Innovation Centre using the HiSeq 2000 platform ( Illumina ) . Fifty to 150 million paired-end sequences ( 100 bp length ) were obtained from 300–500 bp library inserts , resulting in an overall 30x coverage for genome reads . Sequences were filtered based on quality and mapped onto the Mus musculus reference genome ( mm9 for genomic DNA , mm10 for RNA ) . For the transcriptome , total RNA was rRNA-depleted before making the libraries . Three libraries were generated for each cell population , though only two were of sufficient quality for subsequent bioinformatic analysis . DESeq and edgeR differential gene expression analyses ( adjusted p-value < 0 . 001 ) , as well as a minimum 2-fold expression difference , were taken into account to determine significantly deregulated messages between FACS-recovered eNCC from control ( G4-GFP ) and TashTTg/Tg embryos . Gene Ontology analyses were performed using GOToolBox , and at least 2-fold enriched/depleted categories were selected from a hypergeometric test with a Benjamini-Hochberg corrected p-value threshold of 0 . 01 ( http://genome . crg . es/GOToolBox ) . Dissected postnatal intestines were fixed in 4% PFA overnight at 4°C and embedded in paraffin . Sections ( 7 μm ) were stained by immunohistochemistry according to standard techniques . Briefly , the paraffin was removed from slide mounted sections by washing with xylene and ethanol , and sections were treated for antigen retrieval with boiling sodium citrate pH 6 . 0 for 10 min . Slides were blocked for 1 h in blocking solution ( 1% bovine serum albumin , 1% milk , in Tris-buffered saline pH 7 . 5 ) , then incubated overnight at 4°C with mouse anti-βIII-Tubulin ( Abcam ab78078 , 1:200 ) primary antibody . Following several Tris-buffered saline ( TBS ) pH7 . 5 washes , sections were incubated for 2h at room temperature ( RT ) in alkaline phosphate-conjugated anti-mouse secondary antibody ( Abcam ab97043 , 1:200 ) . After two washes in TBS pH 7 . 5 and one at pH 9 . 5 , the staining was revealed with a nitro blue tetrazolium ( NBT , 500μg/ml ) and 5-bromo-4-chloro-3-indolyl-phosphate ( BCIP , 187 . 5μg/ml ) solution ( Roche Applied Science ) for 10 to 20 minutes . The reaction was stopped with a solution of TBS pH7 . 5 containing EDTA ( 20mM ) and the slides were mounted with glycerol mounting medium ( DAKO ) . For embryonic tissues , freshly dissected intestines were fixed 1 hour at RT with 4% PFA in PBS . Alternatively , whole embryos were fixed overnight at 4°C and the intestines dissected afterwards . For adult tissues , whole intestines were fixed in 4% PFA overnight at 4°C , cut longitudinally along the mesentery and washed in PBS . The outermost muscle layers were then stripped from the mucosa/submucosa . Fixed tissues were dehydrated in methanol . After rehydration , the intestines were incubated 2 hours at RT in blocking solution ( 10% fetal bovine serum , 0 . 1% Triton-X100 in PBS ) . Tissues were incubated with primary antibodies overnight at 4°C . The antibodies used were: mouse anti-βIII-Tubulin ( 1:200; Abcam ab78078 ) , rabbit anti-S100β ( 1:500; Dako Z0311 ) , goat anti-Sox10 ( 1:100 , Santa Cruz Biotech . sc-17342 ) and rabbit anti-Ki67 ( 1:1000; Abcam ab15580 ) . Secondary antibodies Alexa Fluor 594- or Alexa Fluor 647-conjugated anti-goat , -mouse or -rabbit ( 1:500 , Jackson Immunoresearch ) were incubated for 2 hours at RT and counterstained with DAPI . All antibodies were diluted with blocking solution . For the TUNEL assay , tissues were permeabilized 20 min at 37°C in 0 . 3% Triton-X100 , 0 . 1% sodium citrate in PBS 1x , then stained in a 1:9 mix of enzyme solution:label solution from the in situ cell death detection kit , TMR red ( Roche Applied Science 12156792910 ) 1h at 37°C . The whole colon ( from cecum to anus ) was dissected from adults and neonates . It was cut longitudinally along the mesentary ( adult tissues only ) , rinsed , pinned flat and fixed in 4% PFA O/N at 4°C . Staining was performed on tissues as previously described [59] . Following the staining procedure , muscle strips were prepared as described above ( adult tissues only ) . Live imaging of eNCC was performed using a suspended culture technique adapted from Nishiyama et al . , 2012 [3] . The abdomen of TashT e11 . 0 embryos was opened and the surrounding tissues trimmed just enough to expose the developing intestine . The embryo was placed on a small nitrocellulose filter ( Millipore GSWP01300 ) soaked in PBS , and the extra tissues surrounding the intestine were slightly pressed onto the membrane . The PBS was then blotted off before being replaced with DMEM/F12 media ( containing 10% FBS and antibiotics ) . The filter was flipped on top of a DMEM/F12-filled 2 mm-wide trough in a 1% agarose film covering the round glass bottom of a 35 mm culture dish ( Greiner Bio One 627860 ) , so that the intestine would float in media without touching either the agarose or the glass . The dish was incubated at 37°C , 5% CO2 during 6 hours , while 250μm-thick stacks were acquired with a 10x objective and a Nikon A1R confocal unit . Cell morphology viewed by YFP fluorescence allowed us to label the center of the cell located at the tip of a chain of migrating eNCC at each timeframe . Five to 6 chain tip cells were tracked from at least 3 intestines of each genotype . Speed and directionality were calculated from this dataset , with the orientation of the mesentery as a reference angle . We cannot totally exclude the possibility that more than one cell was included per chain tip measurement as individual eNCC at the wavefront sometimes exchange places with one another by a leapfrogging process [38] . Ex vivo cell migration assays were performed as previously described [23] . Collagen gels containing or not GDNF ( 10 ng/ml , Cedarlane CLCYT305–2 ) and/or EDN3 ( 250 ng/ml , Sigma E9137 ) were prepared at least 1 hour before use and kept at 37°C in a CO2 ( 5% ) incubator . Vibratome transverse sections ( 200 μm ) of embryonic small intestines were put down on the collagen gels for 3 days at 37°C with 5% CO2 to let cells migrate into the gel . Bowel sections were removed and cells within the gels were fixed with 4% PFA for 1 hour at RT before being stained with DAPI ( Sigma-Aldrich ) to detect cell nuclei . All images were taken at 70X magnification with a Leica M205FA fluorescence stereomicroscope as described below . YFP expression in TashT tissues or in migrating cells was visualized using a Leica M205FA fluorescence stereomicroscope . IHC slides were observed using a DM2000 Leica upright microscope . Images were acquired with a Leica DFC495 digital camera and Leica Application Suite ( LAS ) software ( Leica microsystems ) . IF of stained intestines were examined using an inverted Nikon TI microscope . Images were acquired with a Nikon A1 confocal unit and NIS-Element AR4 software , using standard excitation and emission filters for visualizing DAPI , YFP , Alexa Fluor 594 and 647 , as well as spectral imaging coupled with linear unmixing in order to distinguish between YFP and GFP fluorescence . All images were processed with ImageJ software [60] . Image J was also used for cell counting with the analyze particles function , or with the cell counter manual function . Constrained elements around the TashT insertion site were identified using Ensembl’s 35 eutherian mammals multiple alignment track ( EPO_LOW_COVERAGE ) on the NCBIm37 mouse genome version ( www . ensembl . org ) . Seven regions named CE1 to CE7 ( ranging between 600 to 1400bp ) and containing one or multiple constrained elements were amplified by PCR ( oligo sequences available upon request ) , cloned in the pGEM-T vector ( Promega ) and validated by sequencing . Luciferase reporter constructs were generated by subcloning each PCR fragment , in both the sense and antisense orientation , into a modified pXP2 vector containing the 109bp Thymidine kinase minimal promoter [61] . Neuro-2a neuroblastoma cells were propagated in EMEM supplemented with 10% FBS whereas P19 embryocarcinoma cells were propagated in alpha-MEM supplemented with 2 . 5% FBS and 7 . 5% CBS . Cos7 and NIH 3T3 cells were propagated in DMEM supplemented with 10% FBS . Transfections in 24-well plates and luciferase assays were performed in triplicate at least three times as previously described [62] . Western blots using whole cell extracts of transfected Cos7 cells were performed as previously described [63] . Cells were transfected with a CMVp-driven expression vector for the same Fam162b-IRES-eGFP bicistronic cassette used to produce transgenic embryos and protein expression was assessed using the following primary antibodies: rabbit anti-Fam162b ( 1:1000; Abcam ab122309 ) , rabbit anti-GFP ( 1:5000; Abcam ab290 ) and rabbit anti-Gapdh ( 1:2500 , Santa Cruz Biotech . sc-25778 ) . Total RNA was extracted using the RNAeasy Plus purification mini kit ( QIAGEN ) on FACS-sorted eNCC . The OneStep RT-PCR kit ( QIAGEN ) was then used on 100 ng of RNA with primers specific to the desired target ( sequences available upon request ) . PCR consisted of 25–30–35 or 30–35–40 cycles of: 20 seconds at 95°C , 30 seconds at 62°C and 30 seconds at 72°C . Amplicons were resolved on a 2% agarose gel and quantified using the densitometry tools of ImageJ . The expression level of the housekeeping gene Gapdh was used for normalization . 3C was performed as previously described [27] with minor modifications . Starting material was either 1x108 cells ( for Neuro-2a and P19 cell lines ) , as recommended , or ~1x106 cells ( for e12 . 5 intestinal material ) , in which case reaction volumes were divided by a factor of 20 . Only one cycle of phenol , then phenol/chloroform extraction was performed . Glycogen ( 0 . 05 mg/ml ) was added as a co-precipitant prior to ethanol precipitation . For each PCR reaction , 20 ng ( BAC library ) , 50 ng ( embryonic intestines libraries ) or 100 ng ( cell lines libraries ) of library DNA was used . Sequencing of the amplicon confirmed the identity of the chimeric fragment amplified . Data are presented as mean ± standard deviation , with the number of experiments ( n ) included in the figure and/or legend . Quantification data were subjected to Student's t-test for statistical significance , except for directionality circular data which were compared through an ANOVA . Differences were considered statistically significant when the p value was less than 0 . 05 .
|
Hirschsprung’s disease ( also known as aganglionic megacolon ) is a severe congenital defect of the enteric nervous system ( ENS ) resulting in complete failure to pass stools . It is characterized by the absence of neural ganglia ( aganglionosis ) in the distal gut due to incomplete colonization of the embryonic intestines by neural crest cells ( NCC ) , the ENS precursors . Hirschsprung’s disease has an incidence of 1 in 5000 newborns and a 4:1 male sex bias . Although many genes have been associated with this complex genetic disease , most of its heritability as well as its male sex bias remain unexplained . Here , we describe an insertional mutant mouse line ( “TashT” ) in which virtually all homozygotes display colonic aganglionosis due to defective migration of enteric NCC , but in which only a subset of homozygotes develops megacolon . Surprisingly , this group is almost exclusively male . The TashT ENS defect stems , at least in part , from the disruption of long-range interactions between evolutionarily conserved elements with silencer activity and Fam162b , resulting in NCC-specific upregulation of this uncharacterized protein coding gene . Global analysis of gene expression further revealed that several hundreds of genes are significantly deregulated in TashT enteric NCC . Interestingly , this dataset includes multiple X-linked candidate genes potentially underlying the male sex bias . Taken together , our data pave the way for a clearer understanding of the intriguing male sex bias of Hirschsprung’s disease .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Male-Biased Aganglionic Megacolon in the TashT Mouse Line Due to Perturbation of Silencer Elements in a Large Gene Desert of Chromosome 10
|
The prospects for the success of malaria control depend , in part , on the basic reproductive number for malaria , R0 . Here , we estimate R0 in a novel way for 121 African populations , and thereby increase the number of R0 estimates for malaria by an order of magnitude . The estimates range from around one to more than 3 , 000 . We also consider malaria transmission and control in finite human populations , of size H . We show that classic formulas approximate the expected number of mosquitoes that could trace infection back to one mosquito after one parasite generation , Z0 ( H ) , but they overestimate the expected number of infected humans per infected human , R0 ( H ) . Heterogeneous biting increases R0 and , as we show , Z0 ( H ) , but we also show that it sometimes reduces R0 ( H ) ; those who are bitten most both infect many vectors and absorb infectious bites . The large range of R0 estimates strongly supports the long-held notion that malaria control presents variable challenges across its transmission spectrum . In populations where R0 is highest , malaria control will require multiple , integrated methods that target those who are bitten most . Therefore , strategic planning for malaria control should consider R0 , the spatial scale of transmission , human population density , and heterogeneous biting .
Each year , Plasmodium falciparum causes approximately 515 million clinical malaria cases [1] and over one million deaths [2 , 3] . Most malaria-related mortality and a large fraction of malaria cases occur in sub-Saharan Africa , where transmission can be very intense [4] . Strategic planning for malaria control should consider the transmission intensity of malaria , which is described by several parasitological and entomological indices ( Table 1 ) . The intensity of malaria transmission affects most aspects of malaria epidemiology and control , including the age at first infection , the fraction of a population that is infected ( i . e . , the parasite rate [PR] ) , the frequency and type of disease syndromes , the incidence of severe disease , the development and loss of functional immunity ( i . e . , immunity that reduces the frequency and severity of clinical symptoms ) , total malaria mortality , and the expected outcome of malaria control [4–8] . Good estimates of malaria transmission intensity are therefore necessary to compare and interpret malaria interventions conducted in different places and times and to objectively evaluate options for malaria control . The basic reproductive number , R0 , has played a central role in epidemiological theory for malaria and other infectious diseases because it provides an index of transmission intensity and establishes threshold criteria . R0 is generally defined as the expected number of hosts who would be infected after one generation of the parasite by a single infectious person who had been introduced into an otherwise naïve population [9 , 10] . If R0 is greater than one , the number of people infected by the parasite increases , and if R0 is less than one , that number declines . Thus , if sustained disease control reduces transmission intensity by a factor that exceeds R0 , the parasite will eventually be eliminated . Alternatively , the fraction of a population that would need to be protected to confer “herd immunity” and interrupt transmission is 1 − 1/R0 . The classic formula for R0 is based on a quantitative description of the P . falciparum life cycle [11 , 12] ( Figure 1 ) . It assumes that human populations are effectively infinite and that all humans are bitten at the same rate , but human populations are finite and some people are bitten by vectors more than others [13 , 14] . In infinite human populations , heterogeneous biting increases R0 because those humans who are bitten most are also most likely to become infected and then , by infecting a large number of mosquitoes , to amplify transmission [15 , 16] . Thus , in infinite human populations , the classic formulas underestimate R0 . Classic and neoclassic ( i . e . , with heterogeneous biting ) formulas for R0 describe idealized populations , where each infectious bite lands on a different host . In reality , some infectious bites land on previously infected hosts because malaria transmission is local . The spatial scale of malaria transmission is affected by vector ecology , especially the distribution of larval habitat and host-seeking behavior , human population density and distribution , and human movement [17 , 18] . Therefore , we reconsider R0 in finite human populations with heterogeneous biting , where some bites reinfect humans or mosquitoes . When the number of humans is not effectively infinite , what is the expected number of infected hosts or vectors after one complete generation of the parasite ? How are these expectations changed when biting is heterogeneous , and what do these ideas imply about malaria control ? Because R0 is both an index of how well malaria spreads and a measure of the effort required to eliminate malaria , it would be the ideal index for strategic malaria control planning , but it has not been routinely recorded . Previous estimates of R0 were made with a variety of methods , and they have a limited spatial coverage . Since each method introduces different sources of potential error and bias , the estimates are not directly comparable [10] . One method estimates each parameter in the classic and neoclassic formulas [19 , 20]; this is rarely done because it is technically and logistically difficult . A second method is based on the rate of increase in the number of human cases during an epidemic in an uninfected and immunologically naïve population [21 , 22] . Obviously , this method has limited application in most African populations , where a substantial fraction of people harbor malaria infections . Equilibrium methods , originally suggested by Macdonald and colleagues [23] ( see Dietz [10] for a review ) , rely on mathematical models that describe the relations between R0 and the population at the steady state . The terms of R0 are rearranged into a set of indices that can be measured in populations where malaria is endemic , so they are most broadly applicable . Here , we introduce new equilibrium methods for estimating R0 that consider heterogeneous biting and factors that introduce a bias , such as sampling issues and immunity . We have used these new ideas to estimate R0 for 121 African populations . These estimates are based on a common methodology and have a continental spatial coverage , so they provide a more useful index of malaria transmission than previous attempts , and one that is suitable for strategic planning for malaria control .
Our estimates of R0 are based on two more commonly measured indices called the entomological inoculation rate ( EIR ) ( E in equations ) , which is the average number of infectious bites received by a person in a year , and the PR ( also called the parasite ratio ) ( X in equations ) , which is the prevalence of malaria infection in humans . Like other equilibrium methods , our method relies on mathematical models that define the steady state relationships between indices and parameters; these are the EIR , the PR , the vectorial capacity , V , which measures the number of infectious bites that arise from all the mosquitoes that are infected by a single infectious person on a single day [24] , the infectivity of humans to mosquitoes , c , and the stability index , S , which measures the number of human bites taken by a vector during its lifetime [25] . The classical parameters and several malaria transmission indices are described in Tables 1 and 2 . At the equilibrium , the relationship between these indices is given by a simple formula ( Methods ) : A simple relationship exists between R0 and vectorial capacity . R0 sums vectorial capacity , discounted for imperfect transmission efficiency , over the average infectious period [26 , 27] . In a population with heterogeneous biting , where the squared coefficient of variation in biting rates is α , R0 is larger by the factor 1 + α , because the humans who are bitten most amplify transmission [15 , 16]; we call α the index of biting disparity . The relationship between R0 , vectorial capacity , and the other indices is given by the formula These formulas are based on the classic assumptions: mosquito lifespan and the duration of human infections are assumed to be exponentially distributed , and R0 is computed for a single parasite type ( for a longer discussion of the assumptions , see the Methods ) . Using equation 2 , estimates of annual EIR and PR from studies of 121 African populations [3] , and parameter estimates from other studies , we generated 121 estimates of R0 ( Figure 2 ) . Parameter estimates for b/r and α were taken from 91 of these studies that included only children less than 15 y old [14] . Published estimates of the stability index range from less than one up to five [9 , 28]; we use the estimate S ≈ 1 , at the low end of published studies . For the infectivity , we use the value c = 0 . 5 , a number that agrees with estimates from direct-feeding experiments [29] . The R0 estimates range from near one to more than 3 , 000 . The median was 115 and the interquartile range was 30−815 . These values are consistent with previous estimates , including one estimate of 1 , 600 [20] in Mngeza , in northwest Tanzania , and another of 2 , 000–5 , 000 [19] in Lira township , in central Uganda . Had these studies considered heterogeneous biting , they would have exceeded our highest estimates . In an area around Madang , Papua New Guinea , where entomological surveys have shown that annual EIR is approximately 150 [30] , and where our methods would suggest that R0 is larger than 500 , an estimate based on age seroprevalence was R0 ≈ 7 . The biological basis for the large discrepancy remains unresolved; one possibility is the strain theory of transmission [31] . Equilibrium methods for estimating R0 are based on the simple assumptions of mathematical models; the difference between these simple assumptions and variance in real populations can introduce a large bias . When biting rates are heterogeneous , for example , mosquitoes bite infected humans at a different frequency than when humans are sampled in a study . Thus , PR may be a biased measure of the probability a mosquito becomes infected after biting a human . In addition , the intensity of transmission at equilibrium may be lower than it would be in that same population without immunity; immunity would reduce the infectivity of humans to mosquitoes ( i . e . , transmission-blocking immunity ) [29 , 32] , or mosquitoes to humans ( i . e . , by clearing an infection before the stages that infect red blood cells develop ) . We have derived new formulas that consider these potential sources of bias , and we use them to modify the previous estimates of R0 . When biting is heterogeneous and when there is some transmission-blocking immunity , it is necessary to introduce a new term called the sampling bias index , σ , that estimates the bias introduced by assuming that the fraction of mosquitoes that would become infected after biting a human is proportional to PR . σ is the ratio of two proportions . The numerator is the proportion of mosquitoes that become infected after biting a human , in a population at equilibrium; it is determined by EIR , by the index of biting disparity ( Table 1 ) , and by the level of transmission-blocking immunity . The denominator is the estimated PR , the proportion of humans that test positive in a study ( Methods ) . Thus , the parameter σ encompasses several complex and poorly quantified processes , including differences in the way that human populations are “sampled” by mosquitoes and scientists , sporadic production of the infectious sexual stages during an infection ( PR is an estimate of the prevalence of the noninfectious asexual stages ) , the reduced infectivity of humans to mosquitoes following the development of transmission-blocking immunity , and the sensitivity of the method used to detect parasites in humans . When infectivity is estimated in a population where malaria is endemic and where there is some degree of immunity , the average infectivity of mosquitoes and humans , denoted bE and cE , respectively , may vary with EIR . The relevant parameters in the formula for R0 are taken from populations without immunity , so infectivity estimates would be from naïve populations , b0 and c0 . The bias introduced by transmission-blocking immunity is included in σ . A correction for infections that are cleared before patency ( i . e . , before the stages that infect red blood cells are detected ) is found by multiplying the formulas for R0 by the term BE = b0/bE , which we call the susceptibility bias index . Thus , we have a new formula for R0: For the same 121 estimates of annual EIR and PR , we generated new estimates of R0 based on different assumptions about σ and BE ( Figure 2 ) . The original estimates effectively assumed that PR is a constant and unbiased index of infectivity ( i . e . , σ = 1 ) and that our estimates of susceptibility were not biased ( i . e . , BE = 1 ) . Our analysis suggests that σ is a complicated function of EIR ( Figure 3; Methods ) . At low EIR ( less than ten per year ) , mosquitoes sample infected individuals more efficiently than a stratified random sample of the population , so estimates of PR are biased by a factor that equals the product of infectivity and the amplification from heterogeneous biting , i . e . , σ ≈ c0 ( 1 + α ) . At moderate to high EIR ( 10–700 per year ) , transmission-blocking immunity reduces the average infectivity of infectious humans to mosquitoes , and since bites on those who have the most immunity account for a large fraction of bites , PR severely overestimates infectivity at high EIR . When we assumed that transmission-blocking immunity develops , as illustrated in Figure 3 , estimates of R0 ranged from below one to nearly 11 , 000 , with a median of 86 and an interquartile range of 15–1 , 000 . The extremely high estimates of R0 raise the question of this index's interpretation in finite human populations; when R0 exceeds the number of humans , what does R0 actually describe ? To interpret R0 , we simulated transmission in small well-mixed human populations of size H through one complete parasite generation with heterogeneous and homogeneous biting ( Figure 4; Methods ) . Let R0 ( H ) denote the expected number of humans who could trace an infection back to one human , and Z0 ( H ) the expected number of mosquitoes who could trace an infection back to one mosquito . ( To clarify our notation , R0 is synonymous with R0 ( ∞ ) , so when population sizes are effectively infinite , R0 = R0 ( ∞ ) = Z0 ( ∞ ) . ) R0 , R0 ( H ) , and Z0 ( H ) can differ , depending on the host population size . When these three indices don't differ , the assumption that populations are effectively infinite is reasonably good . When they differ by more than 10% , we call the populations “small . ” Small populations are defined by R0 and H , as well as the index of biting disparity , α , and the stability index , S . When the size of the human population was small and malaria transmission was very intense , R0 ( H ) was limited by the number of humans; obviously , R0 ( H ) ≤ H . If every human received exactly the same number of bites , some of them would have remained uninfected , by chance . With stochastic biting , there would be some variance in the number of bites received by each individual , even if the expected biting rates were uniformly distributed . Since humans received multiple bites , this tended to increase the proportion of bites that were absorbed by already infected humans , thereby reducing R0 ( H ) . When human population sizes were effectively infinite , each infectious bite landed on a different human . In finite populations , heterogeneous biting amplifies transmission , as measured by Z0 ( H ) , just as it does for infinite populations , because those who are bitten most infect a large number of mosquitoes [15 , 16] . Surprisingly , heterogeneous biting reduced R0 ( H ) below the expected number for homogeneous biting , especially when R0 was large and H was low ( Figure 5 ) . Heterogeneous biting reduced R0 ( H ) , i . e . , the 20% of individuals who were bitten most also absorbed 80% of the infectious bites . Thus , a larger fraction of infectious bites were “reinfection” events; the transmission amplification associated with heterogeneous biting was nullified by a “superabsorbing” effect when those same individuals received most of the infectious bites . The range of human population sizes that would be considered “small” differed for Z0 ( H ) and R0 ( H ) ( Figure 5 ) . Z0 ( H ) rises to R0 much more rapidly than does R0 ( H ) , when considered as a function of human population size , H . Some mosquitoes become infected and return to bite the same human again; reinfection of mosquitoes affects both R0 ( H ) and Z0 ( H ) . The fraction of mosquitoes that are reinfected depends mainly on the stability index , S , the index of biting disparity , α , and human population size . For reasonable estimates of S and α , Z0 ( H ) approaches R0 when the neighborhood includes less than 100 humans ( Methods ) . “Small” for R0 ( H ) depends on the ratio of R0 to H and the index of biting disparity . Some reinfection of mosquitoes does reduce R0 ( H ) , but this is a relatively unimportant effect for H > 25 ( Methods ) . The much larger effect is reinfection of humans . Obviously , when R0 and H are of comparable size , repeat infection of humans substantially reduces R0 ( H ) , but when the human population is several times larger than R0 , R0 ( H ) ≈ R0 , because very few people receive multiple bites . As a rule of thumb , R0 ( H ) approaches R0 when H > 2R0 . When 20% of people get 80% of the bites , the two measures are not close to one another until human population sizes are much larger: R0 ( H ) ≈ R0 when H > 100R0 . The asymmetry between R0 ( H ) and Z0 ( H ) as a function of R0 and H arises because of the large difference in the number of humans infected by each mosquito and the number of mosquitoes infected by each human . Mosquitoes have short lives , typically 1–2 wk . The expected number of humans infected per mosquito—cS by our assumptions—is typically much less than three . The infectious period in humans , by contrast , stretches out over several months . The number of mosquitoes that bite a human during that time can range upwards to several thousand , limited mainly by the ratio of mosquitoes to humans . The number of mosquitoes infected by a single human can be so large that it exceeds the number of humans available to be bitten . When a large number of bites are distributed back on a limited number of humans , a substantial fraction result in reinfection . The large range of R0 estimates suggests that malaria control presents a variable challenge across Africa . At low transmission intensities , local elimination of malaria might be a practical goal . At the highest transmission intensities , classic theory suggests that transmission would need to be reduced by a factor of thousands , or that greater than 99% of hosts would need to be protected from infection . The amplification asymmetry that defines the relationship between R0 , H , R0 ( H ) , and Z0 ( H ) suggests that malaria control measures set different targets depending on the control method deployed . Here , we consider the implications of the extreme variation in R0 for control in finite populations with heterogeneous biting , where a few humans might account for a very large fraction of all infectious bites . In such populations , control measures that target those who are bitten most will tend to disproportionately reduce transmission . To explore these ideas , we simulated malaria control . Because of differences in the way that control measures scale with human population size and alter transmission , we considered three categories of malaria control: host-based , vector-based , and mixed . Host-based methods , including antimalarial drugs or vaccines , reduce or completely neutralize transmission from hosts . Vector-based methods target vector populations in a general way: they lower the intensity of malaria transmission by reducing total vector density or adult lifespan . Mixed methods include insecticide-treated nets ( ITNs ) and indoor residual spraying ( IRS ) . Like vector-based methods , they achieve their greatest effects by killing vectors , but like host-based methods , they are deployed around hosts to whom vectors are attracted . Host-based methods include chemotherapy , chemoprophylaxis , and vaccines . Chemotherapy to clear infections would shorten the infectious period and reduce transmission . Obviously , case management does reduce the number of infectious individuals , but much larger reductions could be achieved through active detection of asymptomatically infected individuals followed by chemotherapy to clear infection . Since a person can become reinfected immediately after clearing an infection , more durable reductions would be achieved through chemoprophylaxis that completely neutralizes a host's ability to transmit . Similar effects would be also achieved through a vaccine that prevented infection , but no commercial vaccine for malaria is currently available or registered for public health use . For perfect targeting , we simulated neutralizing that fraction of the individuals who were bitten most ( Methods ) . With perfect targeting , herd immunity was achieved by neutralizing a relatively small fraction of hosts ( Figure 6 ) ; neutralizing transmission from those who are bitten most makes the most of superabsorbing . The threshold population coverage required to confer herd immunity increased approximately linearly with the logarithm of R0 , rising from around 20% of the human population when R0 was 50 , to 50% when R0 was 2 , 000 , much lower than the 98% and 99 . 95% coverage predicted to be necessary assuming homogeneous biting in infinite populations . When hosts were randomly neutralized , much higher coverage was required to achieve herd immunity; classic theory provided a useful guide , although the coverage required to achieve herd immunity was slightly lower in small populations ( Figure 6 ) . Vector-based methods include mass spraying and other methods that target adult mosquitoes or larvae with pesticides or that reduce larval habitat . Our analysis suggests that R0 provides a fairly good measure of the factor by which transmission would have to be reduced by vector control to eliminate malaria . In very small populations , there is some repeat infection of mosquitoes regardless of R0 , but as vector control reduces R0 , repeat infection of humans becomes much less common . Mixed methods merit a separate consideration from host- or vector-based methods because their success is often measured in terms of the proportion of hosts covered , and the effects on vector populations are more complicated than for purely vector-based control . Mixed methods reduce transmission from some hosts , but some vectors successfully feed despite ITNs or IRS , some mosquitoes are killed , and some vectors are repelled and attempt to bite again . We simulated targeted and random control with mixed methods ( Methods; Figure 7 ) . Like host-based methods , ITNs and IRS were very effective at reducing transmission when they were targeted , but the benefits also saturated after those who accounted for most of the bites were protected . Despite the promise of enormous reductions in R0 through reductions in the lifespan of vectors [27] , the total reductions in transmission from ITNs or IRS were limited . For the parameters we considered here , ITNs or IRS did not confer herd immunity , even with 100% coverage , for values of R0 well below our median estimate . The maximum reductions in transmission depended on the fraction of mosquitoes that were killed or deflected by ITNs or IRS , and there is substantial uncertainty about these parameters under field conditions . The maximum reductions were also affected by the stability index , a measure of an individual vector's transmission efficiency . The same level of transmission can be generated by a very large number of inefficient vectors , or a lesser number of efficient ones . ITNs and IRS were most effective at reducing transmission from very efficient vectors , i . e . , vector populations with a high stability index . The lower the stability index , the lower the potential proportional reductions in transmission . Since it was possible to achieve most of the reductions in transmission by targeting those who are bitten most , it might be possible to reduce costs by targeting . A side effect of ITNs or IRS was that the deflected bites were redistributed , so biting increased on those members of the population who were not protected ( Figure 7 ) . ITNs and IRS lower the risk of infection to unprotected individuals in the surrounding population by depleting vector populations , deflecting bites onto nonhuman hosts , or shortening vector lifespan . Despite the lower risk of infection overall , increased biting on unprotected hosts could increase their risk . Our analysis was focused on changes in R0 ( H ) and Z0 ( H ) , so it did not explicitly consider the risk of infection , as measured by either EIR or PR . To evaluate these questions , a different sort of analysis would be required .
R0 is an important metric for strategic planning for malaria control because it helps to set priorities and define realistic expectations about the outcome of control . Despite the importance of R0 , it has not been commonly estimated; the new estimates presented here increase the total number of published R0 estimates for malaria by an order of magnitude . The extremely large range of these R0 estimates suggests that a globally defined “one-size-fits-all” malaria control strategy would be inefficient . Where R0 is low , local elimination of malaria may be practical , even optimal . Where R0 is in the thousands , malaria may resist elimination even after heavy investments in multiple control measures [33] . In such populations , focused research to identify important aspects of local transmission would help to target control and achieve larger reductions . Mathematical modeling and R0 provide a quantitative framework for strategic planning , one that can be modified to suit the local micro-epidemiology [34] . Important factors for control include the density and distribution of humans , the distribution of larval habitat , the vector species and their biting habits , and the seasonal patterns of transmission . Our analysis here suggests that the size of the local human population is also an important factor to consider , and that different methods may be effective ( or cost-effective ) , depending on the distribution of humans and vectors . Thus , an important factor in evaluating the success of malaria control is the spatial scale of malaria transmission , which is determined by several factors . Mosquito flight distances may be shorter when human blood meals are close to oviposition sites , so the spatial scale of transmission is codetermined by human population density , the distribution of humans and vector habitat , vector ecology , and vector behavior [17 , 18] . The spatial scale is also affected by the movement of humans . The formulas that link commonly measured entomological and parasitological indices to transmission intensity , and that correct these estimates for vector ecology and human population density , provide obvious opportunities for extensive mapping of malaria endemicity to help guide and rationalize control . These opportunities are explored in detail elsewhere [35] . The large reductions in transmission from targeting control are only possible if those who are bitten most can be identified , as has been done for some vector-borne diseases [36] . The feasibility of targeting depends strongly on the underlying causes of heterogeneous biting . Potential causes include mosquito aggregation around places where adult mosquitoes emerge [17] or vectors oviposit [18]; also , some components of breath and sweat [37] and dirty linen [38 , 39] make some humans inherently more attractive to mosquitoes [39 , 40] . Other causes of differential biting include the use of bed nets , protective clothing , and repellants [41] , housing quality and design [42] , pregnancy [43] , alcohol consumption [44] , body size [45] , and defensive behavior [46] . With research , some of these may be exploited to identify and target those who are bitten most , and thereby improve malaria control . One practical idea is to target those with clinical malaria and presumptively treat their families and nearest neighbors with efficacious antimalarial drugs with antigametocidal properties ( i . e . , that clear the infectious stages ) [47] to clear infection and reduce the local reservoir . In low transmission areas , where a large fraction of new malaria infections result in clinical malaria , such targeting has demonstrably reduced transmission [48 , 49] . In high transmission areas , where a lower fraction of new cases result in clinical malaria , clinical malaria in young children may provide some indication of where drug treatment would be most effectively targeted . In such areas , the required reductions in transmission intensity are unlikely to be achieved by any single control measure . Where R0 exceeds a thousand , the additional widespread use of ITNs and supplementary targeted IRS may be required to achieve desired reductions in morbidity and mortality [33] . In small human populations , transmission may be effectively controlled by identifying those individuals who are most important for transmission and neutralizing their potential to transmit malaria . For example , consider an island that has only a few people , but many vectors . If one additional person came ashore infected with malaria , an epidemic would tend to ensue , if R0 ( H ) > 1 . It may not be possible to control the epidemic with ITNs ( i . e . , because Z0 ( H ) ≫ 1 ) , but malaria could be rapidly eliminated by clearing the infection from these individuals and preventing new infections with chemoprophylaxis . In large human populations , malaria could be controlled by targeting the same fraction of humans , but this might represent a very large number of people , so the costs may differ dramatically relative to control measures in small populations . Our analysis suggests that R0 provides a reasonably good estimate of the reductions in transmission intensity that would be required to eliminate malaria through vector-based control . Obviously , the decision to invest in vector-based control depends on many considerations . Like heterogeneous biting , the heterogeneous distribution of adult emergence rates from larval habitats would affect the benefits of larval control . If most of the adult mosquitoes could be eliminated by removing a few larval breeding sites , targeting larval habitats might produce a large gain for little effort . In the extreme case , if all the mosquitoes emerged from a well , the easiest solution might be to cover the well . Since the benefits are related to the number of humans who would benefit , vector-based control will be more cost-effective when there are many humans . In large , urban populations , it might be more cost-effective to target vector populations for control , because of the simple fact that there is much less area to treat and many more people who benefit [50] . The effectiveness of various malaria control methods depends on the context of local transmission , but several general principles derived from the classic modeling efforts are germane . First , since the infectious period for malaria can be extremely long and a substantial fraction of the P . falciparum reservoir resides in asymptomatic cases , the infectious period can be shortened and the reservoir of parasites reduced by the use of antimalarial drugs . Thus , effective antimalarial drugs can be important tools for malaria control as well as for treating clinical malaria , although this does raise concerns about the spread of resistance . Second , although the intensity of malaria transmission is exquisitely sensitive to the mortality rate of adult mosquitoes [27] , potential reductions in transmission intensity via manipulations of this parameter are limited by the fact that ITNs and IRS are not completely efficient; the maximum benefits depend on the fraction of mosquitoes that are killed or repelled and on aspects of the vector populations , especially the stability index . Because most of the reductions in transmission come from protecting a few humans , it is far more important to improve the killing effects of ITNs or IRS around those who are bitten most than to improve coverage on those who are bitten least; however , complete coverage and improved killing effects may be necessary to reach control goals . Finally , when host population sizes are small or transmission is very localized , targeted neutralization may be an extremely effective way to protect other people in the community from getting malaria . In some places , vector control may be an effective and cost-effective way to reduce the burden of malaria [2] , and it has had some historical success [51] , but it may not be cost-effective everywhere . In some of the African populations described here , where malaria transmission is very intense , no single control measure will be sufficient . Nevertheless , if the suite of interventions appropriate for the transmission regime could be implemented at the appropriate targeted scale in many malaria-endemic nations , the malaria-related millennium development goals could be achieved well before an effective vaccine is available . Clarifying the optimal mix of interventions and how these can be mapped and optimally targeted at scale thus remains an important direction for our collective future research .
Ross developed and Macdonald modified a mathematical model for the transmission of a vector-borne disease that is a simplified quantitative description of the parasite life cycle [11 , 12] . The parameter names , following Macdonald's notation , are given in Table 2 . The life-cycle model tracks the fraction of infected humans , X , and the fraction of infectious mosquitoes , Y , over time: In this system of equations , the parasite persists if R0 > 1 , where If R0 > 1 , the equilibria are given by the expressions Since the average mosquito lifespan is short ( i . e . , 1/g ≈ 10–20 d ) , but the malaria infections in humans last months ( i . e . , b/r ≈ 170 d [14] ) , the proportion of infectious mosquitoes adjusts rapidly to the proportion of infectious humans , i . e . , the sporozoite rate tracks PR when mosquito populations are constant ( but see the discussions by Aron and May [52] and by Smith et al . [17] ) . Thus , EIR is given by the formula where V denotes vectorial capacity , following the original definition ( see Table 2 ) [24] . Solving for V , we get By our notation R0 = bcV/r , so we can compute R0 by solving for vectorial capacity: Dietz [15] and Dye and Hasibeder [16] have demonstrated that R0 is higher because of heterogeneous biting: where α is the squared coefficient of variation of the human biting rate . In these equations , mortality during sporogony is counted , but the delay for sporogony is not [17] . These equations give expressions for R0 and equilibria , X̄and Ȳ , that are consistent with the simple assumptions of the classic model . These equations differ slightly from those given by Anderson and May , who write [9] , but the equilibrium would not be consistent with the standard assumptions when mortality during sporogony is incorporated by setting c′ = ce−gn [27] . Closely related delay equations are given by Aron and May [52] . An alternative approach incorporating a realistic incubation period was modeled by Smith et al . [17] . All these models assume constant per capita mortality for mosquitoes , and so they ignore important factors such as temperature-dependent mortality and senescence . Macdonald et al . 's equilibrium method estimates R0 from the force of infection [23]; usually , these estimates of h are based on the change in PR with age in cross-sectional surveys: so The estimates of b/r and α come from a nonlinear regression analysis using a model with superinfection ( i . e . , multiple infections ) [14] . Here , the connection between that model and the life-cycle model is explained . A generalized form of the life-cycle model tracks the fraction of the human population with some number of parasite “broods” [53–55] , denoted i . New broods are introduced by new infections at the happenings rate , which might depend on the number of broods present , hi , and these broods are cleared naturally , also depending on the number of broods present , ρi . The change in the fraction of uninfected humans is described by an equation: The change in the fraction of humans that are infected with i broods is given by This is an extremely general formulation of a model for infection , although the idea of a “brood” remains poorly defined . For different assumptions about hi and ρi , and for explicit assumptions about transmission of different broods by mosquitoes , it is possible to generate a very large number of models for infection in humans; some of these have been worked out by Dietz [56] . With a single brood , the dynamics reduce to the classical formulation . If there are a very large ( effectively infinite ) number of broods , then the force of infection is constant , hi = bE . For an infinite number of broods that clear independently , i . e . , ρi = ir , the distribution of brood number at equilibrium is Poisson with mean bE/r [55] , and the fraction infected is given by These estimates of R0 are based on Smith et al . 's estimate of b/r , which is based on the infinite brood and independent clearance model [14] . In turn , the formulas for R0 consider the invasion of a population by a single brood . The probability that a mosquito becomes infected , per bite , in the life-cycle model is denoted cX . In reality , transmission-blocking immunity and heterogeneous biting skew the probability that a mosquito becomes infected , per bite . Let X~ denote the probability that a mosquito becomes infected after biting a human ( i . e . , in the life-cycle model X~ = cX ) ; then , infection in mosquitoes follows the equation Following similar arguments as before , we get that vectorial capacity is given by the formula Because of transmission-blocking immunity , infectivity of humans declines as a function of EIR , denoted cE . Similarly , immunity at the liver stage affects the average infectivity of mosquitoes , denoted bE . Since R0 is defined for naïve populations , the formulas are based on infectivity in naïve hosts , c0 and b0 . Following similar arguments as before: Since our estimate of b/r may actually be an estimate of bE/r , we need to correct the estimate by the ratio BE = b0/bE . The bias introduced by transmission-blocking immunity depends implicitly on heterogeneous biting . With heterogeneous biting , mosquitoes bite individuals with index s at the rate sE; s is called a biting weight . Let X ( s ) denote the fraction of individuals with biting weight s that are infected , and let Γ ( s , α ) be the fraction of the population that has index s [14] . Finally , let c ( sE ) denote the average infectivity of humans who have a personal expected biting rate , sE . It follows that the probability a mosquito becomes infected after biting a human is We let c ( sE ) = c0e−γsE , so that , because of the development of transmission-blocking immunity , infectivity declines in those who are bitten most . Using the Γ distribution and the equations for superinfection , as in [14] , equation 19 can be solved: Similarly , prevalence , Ȳ , is given by [14]: We assume that a well-designed study would estimate Ȳ , while a mosquito sees X~ . The sampling bias index is . Using this formula , we can estimate R0 as a function of EIR and PR: We note that when EIR is low , σ ≈ c0 ( 1 + α ) , so this formula simplifies to the following: Here , we explore the interpretation of R0 in finite populations , of size H . This approach is motivated by the extremely high estimates of R0 ( or R0 ( ∞ ) ) , which in some cases may even exceed the local human population size . Here , R0 ( H ) is defined as the expected number of different individual humans that can trace an infection back to a single human after one complete generation of the parasite , and Z0 ( H ) is the number of mosquitoes that can trace an infection back to a single mosquito . Mathematical approaches to R0 have evolved since Macdonald [12] , and so have the definitions . We maintain the connection to Macdonald's original definition , in part , for historical continuity . Nowadays , R0 is computed using next-generation approaches [10 , 57] . By those definitions , the quantity that we compute is called . Next-generation approaches are linearized approximations , and R0 is an eigenvalue associated with asymptotic growth rates . Our reevaluation of R0 is motivated by a different case—when R0 and H are of comparable size—so asymptotic growth rates are not our primary interest . Since R0 ( H ) and Z0 ( H ) differ , it is possible that R0 ( H ) < 1 , but Z0 ( H ) > 1 . In finite populations where R0 ( H ) and Z0 ( H ) are both near one , malaria would be likely to random walk to extinction , in any case . To compute R0 ( H ) or Z0 ( H ) in heterogeneous populations , let i = 1 . . . H index humans in a population , and let masi denote their individual biting rates , where the distribution of biting weights , {s} , is constrained to have a mean of 1; . The proportion of bites that land on the ith individual is therefore si/H . First , we compute the number of infected humans , per human . While infectious , the ith human receives masi/r bites . Each mosquito biting the individual becomes infected with probability c , but some fraction of these mosquitoes return to bite the ith human again , so we need to discount multiple infection of mosquitoes . The fraction of bites on the ith human is si/H , so a short time after the ith human has become infected , the fraction of mosquitoes that were already infected by that human is The proportion of those bites that infect a different mosquito is Note that more than 90% of bites are new infections when H > 9cS , so reinfection of mosquitoes is a relatively small effect when H > 25 . Thereafter , the mosquito survives to become infectious with probability ce−gn , and then is expected to give a/g infectious bites . Thus , the total number of infectious bites that arise from the ith human is The jth human in that population is expected to be bitten at the rate Zisj/H , and each bite causes an infection with probability b . Thus , the probability that the jth individual remains uninfected is If the ith person is the index case , the expected number of infected humans is There are two reasonable expectations to be computed . First is the unweighted expectation: The second is the weighted expectation: We prefer this second , weighted expectation because it reflects heterogeneous biting , because those who are bitten most are most likely to be the index case , and because in the infinite human population limit , it converges to the formula for R0 derived by Dietz [15] and Dye and Hasibeder [16] . Note that α is the squared CV of {s} and that . From a single infectious mosquito , the expected number of bites that produce an infection is bS . The probability that the ith person becomes infected is Thereafter , that person gets masi/r bites before clearing an infection . The number of infected humans is A fraction c of all bites infect uninfected mosquitoes . As before , some fraction of mosquitoes are already infected . We consider only those infected mosquitoes that can trace their infection back to the index mosquito , so following the previous argument , the fraction of mosquito infections that are not reinfections is Thereafter , e−gn infected mosquitoes survive to become infectious . Therefore , the total number of infectious mosquitoes per infectious mosquito is given by the formula The fraction of newly infected mosquitoes increases rapidly as a function of H . In a very large population , T is less than bS; more than 90% of bites are new infections when H > 9cST > 9bcS2 . When transmission from humans is neutralized by a perfect vaccine or by chemoprophylaxis , infected humans continue to absorb bites , but don't infect any mosquitoes . We construct a vector of length H where Vj = 0 if an individual is protected , and Vj = 1 otherwise . With targeted protection , Note that Vj appears in the exponent to account for bites on neutralized individuals . To compute R0 ( H ) with neutralization , we compute the weighted expectation: Here , Vi removes protected individuals from the computation—if a person is protected , then it is not possible for him to be the index case , by assumption . After controlling vector populations , estimates of R0 ( H ) and Z0 ( H ) would be computed as before , but with different estimates of m or g . It is also possible that vector control would change the distribution of biting weights , but this is not a question that we have addressed here . When humans are protected from infection by ITNs or by IRS , some fraction of the mosquitoes that attempt to bite a protected human are killed , and some fraction are diverted onto other hosts . To model both effects , we assume that the biting weights describe the probability of finding a host during each attempt , that a fraction of biting attempts on protected humans kill the mosquito each visit ( denoted δ ) , that a fraction of mosquitoes successfully feed ( ψ ) , and that those mosquitoes that neither die nor successfully feed fly off to begin a new search . Of these , a fraction Q finds a human , again . Let N denote the set of people who are protected , then the fraction of visits that find a protected human is . We ignore the delay required to find another host , and assume that the vectors instantaneously reassort themselves onto hosts until they have either died or successfully fed . The fraction of mosquitoes that die is δP at the first attempt , plus δP times all those who failed to feed the first time and again find a protected human , and so on: Thus , ϕ is the fraction of human feeding attempts by vectors that result in mosquito death . With ITN use , the mosquito death rate increases to g′ = g + ϕa . By a similar argument , the feeding rate on the ith protected host is And the proportion of bites on the jth unprotected hosts increases to In a finite population , we compute R0 ( H ) and Z0 ( H ) as before , with new parameters describing human feeding , mosquito mortality , and biting weights ( which may not sum to one ) . Obviously , the success of ITNs depends on the baseline parameters , Q , δ , and ψ . Here , we simulate control for Q = 0 . 9 , δ = 0 . 3 , and ψ = 0 . 2 .
|
Each year malaria results in more than a million deaths . Controlling this disease involves understanding its transmission . For all infectious disease , the basic reproductive number , R0 , describes the most important aspects of transmission . This is the expected number of hosts that can trace their infection directly back to a single host after one disease generation . For vector-borne diseases , such as malaria , R0 is given by a classic formula . We made 121 estimates of R0 for Plasmodium falciparum malaria in African populations . The estimates range from around one to over 3 , 000 , providing much higher estimates than previously thought . We also show that in small human populations , R0 approximates transmission when counting infections from mosquito to mosquito , but overestimates it from human to human . Previous studies showed that transmission is amplified if some humans are bitten more than others . We confirm that such heterogeneous biting amplifies transmission counting from mosquito to mosquito , but it can also dampen transmission counting from human to human . Humans who are bitten most both infect a large number of mosquitoes and absorb many infectious bites . What does this mean for control ? When R0 is in the thousands , eliminating malaria may seem impossible . If transmission from the humans who are bitten the most can be targeted , however , local elimination can still be within reach .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"public",
"health",
"and",
"epidemiology",
"ecology",
"plasmodium",
"anopheles",
"homo",
"(human)"
] |
2007
|
Revisiting the Basic Reproductive Number for Malaria and Its Implications for Malaria Control
|
HIV infection can be effectively controlled by anti-retroviral therapy ( ART ) in most patients . However therapy must be continued for life , because interruption of ART leads to rapid recrudescence of infection from long-lived latently infected cells . A number of approaches are currently being developed to ‘purge’ the reservoir of latently infected cells in order to either eliminate infection completely , or significantly delay the time to viral recrudescence after therapy interruption . A fundamental question in HIV research is how frequently the virus reactivates from latency , and thus how much the reservoir might need to be reduced to produce a prolonged antiretroviral-free HIV remission . Here we provide the first direct estimates of the frequency of viral recrudescence after ART interruption , combining data from four independent cohorts of patients undergoing treatment interruption , comprising 100 patients in total . We estimate that viral replication is initiated on average once every ≈6 days ( range 5 . 1- 7 . 6 days ) . This rate is around 24 times lower than previous thought , and is very similar across the cohorts . In addition , we analyse data on the ratios of different ‘reactivation founder’ viruses in a separate cohort of patients undergoing ART-interruption , and estimate the frequency of successful reactivation to be once every 3 . 6 days . This suggests that a reduction in the reservoir size of around 50-70-fold would be required to increase the average time-to-recrudescence to about one year , and thus achieve at least a short period of anti-retroviral free HIV remission . Our analyses suggests that time-to-recrudescence studies will need to be large in order to detect modest changes in the reservoir , and that macaque models of SIV latency may have much higher frequencies of viral recrudescence after ART interruption than seen in human HIV infection . Understanding the mean frequency of recrudescence from latency is an important first step in approaches to prolong antiretroviral-free viral remission in HIV .
The development of highly potent antiretroviral therapy ( ART ) for HIV means that the virus can be effectively controlled in most treated patients . However , ART must be taken continuously , as interruption of ART is followed by the rapid recrudescence of virus from a quiescent ‘latent reservoir’ of infected cells . A major thrust of HIV research is to reduce the latent reservoir so that prolonged antiretroviral-free HIV remission can be achieved . A number of ‘latency reversing agents’ ( LRA ) are currently being developed to reduce the latent reservoir by reactivating latently infected cells [1–4] . Clinical studies of LRA in HIV-infected patients on ART have shown the ability to significantly increase cell-associated unspliced HIV RNA and in some studies increase plasma HIV RNA . However , these studies have not resulted in decreases in HIV DNA—a crude surrogate marker of latently infected cells—or measurable reductions in various measurements of the latent reservoir or antiretroviral free HIV remission [5–10] . A fundamental question in achieving HIV remission is what level of reduction of latently infected cells is required ? It is currently estimated that the reservoir of latently infected cells may be between one and 60 million cells [11–13] . Complete elimination of this would thus require reducing the size of the reservoir by at least one million fold . However , reducing the reservoir by smaller amounts may still produce significant delays between ART-interruption and viral recrudescence , allowing potentially for prolonged interruptions of therapy before viral recrudescence . Understanding factors that predict the duration of viral remission will be critical for the future design of eradication studies [14] . The dynamics of HIV reactivation from latency have not been examined in detail experimentally . However , it is clear from a number of studies that after ART-interruption there is generally a delay of about a week before viral rebound can be detected , and about half of the patients often experience rebound within the first two weeks or so [8 , 15–17] . However , a proportion of patients usually remains virus-free even after a month , suggesting variable dynamics of reactivation . From this observation , we can attempt to predict the underlying dynamics from our understanding of infection . Firstly , we expect that the latently infected cells might ‘attempt’ reactivation even during successful ART , because ART itself is not expected to affect the rate of initial latent cell reactivation . These reactivation attempts by latent cells may occur at some average frequency , and some fraction of these events will produce replication competent virus and thus be capable of initiating successful viral rebound . These reactivation events do not result in successful viral growth while therapeutic levels of ART are present . Thus , after ART there will be a period of ‘drug-washout’ before the virus is able to grow ( which will vary depending on the pharmacokinetics of the ART regime ) . Once ART levels have declined sufficiently that viral growth is possible , there may be some delay until the first replication-competent viral reactivation event occurs ( assuming reactivation is occurring randomly ) . Following the first successful reactivation , we expect virus levels will start at some low level , and then take some time to grow to the level of viral detection . Together , the time for drug-washout and viral growth create a ‘fixed delay’ before reactivation can be detected , and likely explain the fact that little rebound is usually detected in the first week after ART-interruption . After this fixed delay , if latent cells are reactivating randomly at some average frequency , this will lead to an exponential distribution in the time-to-reactivation observed , and may explain why some patients remain virus-free for longer periods . By fitting of the time-to-reactivation curve , we can estimate the frequency of successful reactivation from latency . In this study , we directly estimate the frequency of HIV recrudescence from latency following ART-interruption by analysing the time to detection of viral rebound from 4 independent patient cohorts undergoing ART-interruption . We find that the average frequency of successful reactivation from latency is approximately once every 6 days , around 24 times lower than previously estimated [18 , 19] . This low rate of successful reactivation has important implications for designing future eradication studies .
After interruption of successful ART , HIV rebounds to detectable levels within a few weeks in the majority of patients . This requires reactivation of latent cells bearing replication competent virus . The frequency with which this reactivation occurs is likely a function of the size of the latent reservoir ( which may vary substantially between individuals [20–22] ) , and the per-latent-cell probability of successive reactivation . If the initiation of viral growth after ART-interruption is a random event then the distribution of time-to-initiation will be exponential , and we could estimate the average frequency of initiation directly from the ‘survival curve’ of time-to-initiation of viral growth . However , since we are usually unable to detect the initiation of viral growth after ART-interruption , we instead measure ‘time-to-detection of virus’ at some threshold viral level . The actual time when we first detect virus is delayed both because of drug washout preventing viral growth immediately after interruption , and the time taken for the virus to grow from its level at initial reactivation to our threshold for detection of plasma virus ( Fig 1A ) . The duration of the delays due to drug washout and viral growth to the level of detection only affect the “shoulder” of the curve of time-to-detection by delaying the time until we could first detect virus ( Fig 1B ) . The average delay to viral detection is thus the sum of the average time between initiation of successful reactivation events , and the delay from reactivation to detection . There may be a distribution in both time-to-initiation as well as from reactivation to detection , both of which may affect the shape of the subsequent time-to-detection curve . To test this approach we first analysed the kinetics of time-to-detection of HIV in a published cohort of nine patients treated with the LRA panobinostat , and undergoing therapy interruption and biweekly monitoring of viral loads [8] . The threshold of detection of HIV viremia was 20 copies ml-1 , and virus was first detected between day 10 and day 45 across the patient group ( Fig 2A ) . To see if the observed time-to-detection was consistent with an exponential process , we plotted the ‘survival curve’ of time-to-detection in the cohort ( Fig 2B ) . This plot demonstrates an initial shoulder ( as expected due to drug washout and the time taken for viral growth ) , followed by a survival curve that conformed well to an exponential process . The exponential rate can be estimated from the survival curve , and equates to a frequency of viral reactivation of once every 7 . 6 days ( 95% confidence intervals ( CI ) = 6 . 5 , 9 . 1 ) . To test whether the exponential model was suitable , we performed a Chi-squared goodness-of-fit analysis , which indicated a good fit to the data ( p = 0 . 988 ) . Although the analysis above is consistent with an exponential process , this does not prove that this is the only source of delay . It has also been proposed that early stochastic events , differences in the initial level of replicating virus , or differences in viral growth rate may contribute to the delay until viral detection [18 , 23] . However , a comparison between a survival curve based on an exponential distribution with one based on a gamma distribution showed that the gamma distribution ( which incorporates multiple delays ) did not provide a significantly better fit ( p = 0 . 72 F-test ) . We can also use a modelling approach to understand the effects of these different factors on time-to-detection . For example , Pearson et al . [24] have estimated the distribution of delays arising from early stochastic events following primary HIV infection under different assumptions . Under most scenarios , the expected distribution of delays from early stochastic events is of the order of 1–3 days . Moreover , the importance of stochastic delays only becomes relevant in the presence of a low frequency of reactivation . In the presence of frequent reactivation , the stochastic delay in any individual reactivation event is overcome by the rapid arrival of the next reactivation ( see S1 File ) . Another potential cause of differences in time-to-detection is differences in initial levels of virus . That is , if the first latent cell to reactivate were to ‘seed’ the infection with a lower initial level of virus , then it will take longer for the virus to grow to the level of detection . However , given the growth rates of virus observed in these patients , the initial level of virus would have to vary by many orders of magnitude to produce the delays observed ( see S2 File ) . It is also possible that the distribution in time-to-detection arose because of slower viral growth in some patients . To investigate whether the observed differences in detection times could be due to slower viral growth , we estimated the viral growth rate from the serial viral load measurements after virus became detectable , and investigated whether later detection was associated with slower viral growth . We found no correlation between viral growth rate and when virus was first detected ( Fig 2C ) , indicating that slow viral growth did not explain the distribution of time-to-detection . Taken together , these results are consistent with the observed time-to-detection in this cohort being determined by a low rate of viral recrudescence from latency , with an average frequency of initiating viral replication of once every 7 . 6 days . A potentially confounding factor with this analysis is that patients were part of a trial of the LRA panobinostat , a histone deacetylase inhibitor , to assess its effect on the HIV reservoir under ART . We note that although panobinostat increased HIV in plasma and cell-associated unspliced HIV RNA , there were no changes in HIV DNA . Thus , it seems unlikely that panobinostat treatment significantly reduced the HIV reservoir . Alternatively , it is possible that panobinostat-induced activation might increase the frequency of viral recrudescence , although this seems unlikely given that the last dose of panobinostat was administered >36 weeks before ART interruption . Given the small number of patients in cohort 1 and their prior treatment with an LRA , it is important to confirm the estimated frequency of reactivation in other patient cohorts that have not received LRA . Therefore we obtained data on time to recrudescence for another three cohorts of patients undergoing ART-interruption , comprising an additional 91 subjects ( summarised in Table 1 ) . In the second cohort , 59 patients treated in primary infection underwent treatment interruption and weekly monitoring [15] ( Fig 3A ) . Estimating the frequency of initiation from the time-to-detection of virus ( at a threshold of 50 copies ml-1 ) we found an average frequency of once every 6 . 3 days ( CI = 5 . 7 , 7 . 1 ) ( Fig 3B ) , similar to our estimate from the panobinostat cohort . Estimation of viral growth rate was less accurate in this cohort , as patients were only sampled weekly . We compared viral growth rate in this cohort with the time-to-detection to once again check whether a difference in viral growth rate could explain the different time-to-detection of virus . We estimated viral growth using a ‘two-point’ growth estimate to compare growth rates of virus in patients where virus was first detected in different weeks . Using this approach , there was no difference in growth rate estimates for patients with virus detected in weeks two and three , but a slightly higher growth rate in week one ( Fig 3C ) . However , this estimate of growth rate is biased by the fact that viral loads at detection were lower in week one , and therefore we were estimating viral growth rate earlier in the growth phase , before it slows towards peak ( Fig 3D ) . Overall , differences in viral growth did not appear to play a major role in time-to-detection of infection in this cohort . We also analysed two other cohorts using data extracted from earlier publications on ART-interruption . The third cohort included 18 patients undergoing ART-interruption , where time-to-detection at a threshold of detection of 50 copies ml-1 was measured and viral growth rates were estimated ( Table 2 of reference [16] ) . The fourth cohort included 14 patients monitored on days 4 , 8 and 14 following ART-interruption ( using data from Fig 1 of reference [17] ) . Because of the small number of patients and timepoints in the fourth cohort , we included data from five sequential interruption cycles . Using the same method for estimating the frequency of initiation from the time-to-detection curves , we found very similar frequencies of viral recrudescence in these two cohorts ( every 5 . 1 , days CI ( 4 . 2 , 6 . 5 ) and 7 . 2 days CI ( 6 . 0 , 8 . 7 ) respectively , see Fig 4A–4B ) . In the third cohort viral growth rate was also estimated independently in the original study ( reference [16] , Table 2 ) , and again , viral growth rate was not significantly correlated with time-to-detection of infection ( Fig 4C ) , confirming that differences in viral growth played little role in the time-to-recrudescence in this study . Comparing all cohorts together , we found a trend for slightly higher frequencies of reactivation in cohort 3 , who initiated ART in chronic infection , and slightly lower frequencies of reactivation in patients treated in primary infection ( cohort 2 ) or with the LRA panobinostat ( cohort 1 ) . However , the frequency of recrudescence was not significantly different between the cohorts ( p- value = 0 . 059 , F-test ) . In addition , we used a Chi-squared test to assess whether the exponential model of reactivation frequency was suitable across the four datasets , and found that the data conformed well to this model ( p = 0 . 996 ) and that a survival curve based on a gamma distribution did not provide a better fit ( p = 0 . 5 , F-test ) . Overall , despite the different sampling regimens and study designs , the estimated frequencies of reactivation were similar across the four cohorts studied ( once every 5 . 1 , 6 . 3 , 7 . 2 , and 7 . 6 days ) , with an average frequency of once every 6 . 0 days ( CI 5 . 5 , 6 . 6 ) . The analysis of time-to-detection of HIV following ART-interruption suggests a relatively low frequency of recrudescence from latency , and thus a significant delay between successive reactivation events . If each reactivation event is ‘founded’ by virus produced by a single latently infected cell , this predicts that early after ART-interruption , the viral population would often be the progeny of a single latent cell ( in much the same way as virus observed early after sexual transmission is thought to arise from a single founder virion ) . Joos et al [25] have compared the diversity of the HIV plasma viral population present soon after ART interruption with the diversity present prior to commencing ART . They found a major narrowing of diversity after ART-interruption , suggesting monoclonal or oligoclonal origins of the plasma virus . Although the viral population after ART-interruption was not entirely homogeneous , they observed one or more ‘families’ of closely relate viruses , differing by only a few nucleotides , similar to the founder viruses observed after sexual transmission . They concluded from this that the viral population after ART-interruption represented random reactivation of latently infected cells , rather than continual seeding of virus . We accessed the viral sequence data from the Joos study ( Genbank accession numbers listed in the original publication [25] ) and reanalyzed this data in order to investigate the ratios of different ‘reactivation founder’ viruses in these patients . We observed six patients in whom it was possible to identify and count the frequency of founder viruses early after ART-interruption , and investigated the ratio of the number of copies of the most frequently observed founder to the next most frequently observed founder ( see S3 File ) . This ratio of founder copies is determined by both the delay until the next founder starts growing , and the overall growth rate of the virus . We then considered the distribution of these founder ratios , and used this to estimate the distribution of reactivation events and thus the average frequency of reactivation . We used maximum likelihood estimation to fit the ratios of founder copies observed in the Joos study to the theoretical distribution of ratios we would expect if founders reactivated λ times per day ( described in detail in methods ) . We found the average frequency of reactivation events ( 1/λ ) to be once every 3 . 6 days ( CI 1 . 98–6 . 62 days ) . The real delay between reactivation events is likely more than this , because in some cases ( marked with an asterisk in Fig 5A ) we can only estimate the minimum ratio ( for example , if all 16 sequences in a patient are from the same founder we can only say the frequency of the next founder is likely <1/16 , whereas it could be much lower ) . On the other hand , it is also possible that two latent cells bearing founder viruses that were identical in the sequenced region reactivated sequentially , and thus were classified as a single founder . We aimed to minimise the likelihood of this occurring by only selecting patients for this analysis with sufficient diversity of virus pre-treatment ( sequenced in the same region of the virus ) . In addition , it is likely that the latent reservoir was indeed more diverse than the circulating virus immediately before treatment , as it contains an archive of different viral strains . Thus , it seems unlikely that we are aggregating multiple identical founder viruses . Future studies using larger regions of the virus , and / or more in-depth sequencing approaches should provide more accurate estimates of the ratio of reactivation founders and the frequency of reactivation . However , our analysis of the ratios of reactivation founder viruses leads to very similar estimates of reactivation frequency to those obtained studying time-to-detection . Recent studies in macaques have suggested that very early treatment after SIV infection may also lead to delayed time-to-recrudescence after ART-interruption [26] . In this study they found that time-to-recrudescence was very short and significantly correlated with area-under-the-curve of viral load since infection , although significantly longer delays were seen only animals treated within 3 days of infection [26] . Using the same approach to estimate frequency of reactivation from time-to-detection of virus in the animals treated with ART at days 7 , 10 and 14 ( ie: excluding animals treated at day 3 ) , we found that the average frequency of initiation of viral replication in macaques was once every 1 . 7 days , compared to every 6 days in HIV ( Fig 4D ) . One explanation for this might be higher levels of HIV DNA in the macaques . However , the total number of HIV copies per million PBMC measured in the macaques just prior to ART-interruption seems similar to that reported in patients during ART [8 , 27] . Another reason for the higher reactivation rate in macaques may be the generally shorter periods of treatment ( 24 weeks of ART in SIV versus >12 months in the HIV studies ( Table 1 ) ) , which may have allowed less time for activated cells to decay and a steady-state of latently infected cells to be attained [28] . Alternatively , differences in immune activation or cytokine levels may also play a role . Regardless of the mechanism , this work suggests that short-term treated macaques may experience much higher rates of reactivation from latency compared to HIV patients even if treated early after infection . The primary goal in tackling HIV latency is to allow prolonged HIV remission in the absence of ART . Thus , a major question is how much we would need to decrease the latent reservoir in order to produce a durable delay in time-to-recrudescence and subsequent recommencement of ART ? A previous study estimated that a reduction in reservoir size of >2000 fold would be required to provide a one year average delay until reactivation [18] . However , that study assumed that viral reactivation was over 24 times faster than our estimates ( reactivation every 0 . 25 days ) , based on indirect modelling approaches published previously [18 , 19] . Our analysis demonstrates a much lower rate of viral reactivation , and thus much smaller reductions in the size of the latent pool would be needed for a one-year delay to reactivation . The required reduction in latent reservoir can be calculated as: R=Td ( 1 ) Where R is the required reduction in size of the latent reservoir , T is the length of delay until viral recrudescence and d is the average time between viral reactivation events ( ie: the baseline frequency of viral reactivation ) . For a baseline frequency of reactivation of once every 6 days ( the average over the four cohorts ) , our analysis predicts that a 61-fold reduction in the reservoir would provide an average one-year delay until recrudescence . Thus , for example , 12 rounds of therapy using an LRA that reduced the reservoir ( and reactivation rate ) by 30% would achieve an ≈72-fold reduction in the reservoir and hence an average one year ART-free control of viremia . Several recent case reports have suggested that very prolonged remission is possible if the reservoir can be reduced by early treatment or other interventions such as bone marrow transplantation . In the case of the ‘Mississippi baby’ , viral recrudescence was not observed until 27 months after ART-interruption [29] . Similarly , in two cases of haematopoietic stem cell transplantation in adults , viral recrudescence was not observed until 84 and 225 days after ART-interruption [30] . Our analysis indicates that the average frequency of reactivation is once every 6 days across the four cohorts we analysed . Therefore these three cases are respectively 135 , 14 and 37 fold longer than expected on average in these cohorts . One might speculate from these delays that the reactivation rate and reservoir size were respectively 135 , 14 and 37 fold smaller than average ( using Eq 1 ) . However , since reactivation is a random process , recrudescence is not always observed at the average time expected . Using the data we can estimate bounds for the likely frequency of initiation of viral replication ( and the extent of reservoir reduction ) based on observed time-to-detection of virus ( see Fig 5B ) . For example , given an observed time-to-detection of 84 days or more , it is highly unlikely that the reservoir was of the average size determined by analysis of our four cohorts ( probability for this is 8 . 4x10-7 ) . For a time-to-detection of 84 days to lie within the range expected for 95% of subjects , then the average frequency of recrudescence would have to be bounded below by 23 days and above by 3318 days . This suggests that in this case the latent reservoir was most likely between 3 . 8 fold and 553 fold smaller than the average size estimated from our four cohorts . Using the same approach in the case of the Mississippi baby , the maximum predicted reduction in viral reservoir ( top border of 95% CI ) is 5 , 300 fold . A corollary to these observations is that the rate of reactivation from latency and level of viral reservoir in the transplant patients is not decreased as much as might be predicted from the degree of chimerism seen in peripheral blood ( <0 . 001% of PBMC were of donor origin [30] ) . However , as noted by the authors of that study , the degree of chimerism in the patients’ tissues are likely significantly higher than that seen in PBMC [30] particularly as the patients received a reduced intensity conditioning regime . In addition , a number of recent studies have suggested that lymphatic sites may be a significant source of virus under therapy [31–33] . Thus , we speculate that reactivation from chimeric tissue sites might contribute to the observed reactivation rate . Overall , the wide error bars on estimates of potential reservoir size based on time-to-detection of individual patients ( Fig 5B ) suggest significant limitations in the use of time-to-detection to estimate reservoir size . Therefore , we also investigated the usefulness of time-to-detection assays in detecting the effects of LRA . A major question in clinical trials of LRA is how to measure changes in the latent reservoir . Approaches using detection of plasma HIV RNA , cell-associated HIV RNA , cell associated HIV DNA , as well as ex vivo quantitative outgrowth assays have been studied [7–10 , 34 , 35] . However , it is not clear whether these measures will reflect time to viral recrudescence after ART-interruption in vivo [8 , 36] . Since HIV remission involves essentially a prolonged time-to-detection of virus , direct measurement of time-to-detection following ART-interruption will ultimately be the most clinically meaningful endpoint . Treatment interruption studies to measure time-to-detection pose a number of ethical questions . Firstly , frequent treatment interruptions may increase morbidity or mortality compared to continuous treatment [37] , although it is less clear that occasional interruptions would have the same effect . Secondly , interruption may act to ‘replenish’ the viral reservoir , although this does not appear to occur quickly [22] . Thirdly , such studies would ideally require a control group , in order to compare time-to-detection in treated versus untreated patients . However , in addition to these factors , there are also a number of issues with the statistical power to detect delays in time-to-detection . Firstly , as indicated by the extremely wide error bars in our estimates of relative reservoir size in the Boston patients and Mississippi baby ( Fig 5B ) , time-to reactivation is not a useful measure of reservoir size in an individual , because of the random nature of reactivation . Time-to-detection is only useful in cohorts of patients . Once we understand time-to-detection as an exponential process , we can apply a power analysis to estimate how many patients would be required to identify differences in time-to-detection . Such an analysis suggests that to detect a 30% decrease in the reservoir ( and a 30% increase in the frequency of initiation of viral replication ) assuming a 100 day follow-up one would require >120 patients in each arm to have an 80% chance of detecting a difference ( Fig 5C ) . Thus , such studies are only likely to be useful in detecting rather large changes in the reservoir and rate of reactivation .
Our work provides the first direct estimates of the frequency of viral recrudescence from latency based on analysis of time-to-detection of plasma viremia from ART-interruption cohort studies . Previous studies have modeled the recrudescence of virus following ART-interruption , using a variety of approaches . This work has often focused on the dynamics of virus within the individual , and either did not estimate the frequency of viral reactivation [16 , 38] , or estimated a constant rate of production of virus , rather than the frequency of events [23] . Rong et al used a similar modeling approach to understand viral ‘blips’ during ART , and estimated that these were infrequent [39] . Pennings et al estimated the frequency of successful latent cell reactivation under ART as five times per day , based on the rate of development of drug resistance under ART and viral mutation rate [19] , and more recently Hill et al used this frequency to model the affects of LRA [18] . Our estimate of the frequency of successful reactivation from latency ( once every 5–8 days ) is based upon analysis of the distribution in time-to-detection in the patient population , and is substantially slower than these previous estimates . The relatively slow frequency of recrudescence has important implications for understanding how to prolong anti-retroviral-free viral remission . In addition , careful consideration of the dynamics of viral recrudescence is critical to designing successful future eradication studies . Our work suggests that rather than using indirect approaches to estimate reductions in the reservoir and predict delays in time-to-recrudescence , we should measure this directly . Previous studies of the number of latently infected cells under therapy have estimated that there are between 1 million and 60 million latently infected cells in the body , and the half-life of the latent reservoir is around 44 months [11–13] . One question that arises from this is whether reactivation from latency plays a significant role in the observed rate of decay of the latent reservoir , and whether periodic reactivation may lead to the reservoir ‘running dry’ [18 , 39 , 40] . If the current estimates of the number of latently infected cells and their decay rate are correct , this means the reservoir ‘loses’ on average at least 500 cells per day . Since we observe a successful reactivation from latency and reseeding of the viral reservoir only every 6 days , this suggests that reactivation would play a minimal role in the decay of the latent reservoir ( unless the reservoir is much smaller than previously estimated ) . Previous modelling has assumed that there may be many ‘abortive’ reactivation events for every successful reactivation leading to recrudescence [18] . This might occur , for example , if very early events in viral reactivation are controlled by the host immune response . However even considering this possibility , it seems unlikely that reactivation from latency is a major factor contributing to the observed half-life of the latent reservoir . Finally , if we assume that previous estimates of the reservoir size are correct , we can also estimate the average probability of an individual latently infected cell successfully initiating viral recrudescence on a given day , and the average time until an individual cell is likely to achieve this . Assuming a conservative reservoir size of one million latently infected cells per patient and the fact that on average a patient has only one successful reactivation every 6 days , we can calculate that an individual latent cell has only a 1 . 7 x 10−7 probability of initiating viral recrudescence each day . Thus , the average time for an individual latent cell to initiate infection ( assuming they all have the same probability of this ) is around 16 , 500 years ( ie: most latent cells will not successfully initiate an infection within the lifetime of the host ) . Although this per-cell probability of reactivation appears very low , it is perhaps worth considering how the reservoir is generally measured experimentally . The estimate of one million cells comes from the frequency of cells able to initiate viral growth in an in vitro viral outgrowth assay [12] . This assay involves stimulation of cells with PHA , and thus aims to estimate the number of ‘reactivatable’ latent cells with this strong and generalized stimulus . Reactivation in vivo may rely on antigenic stimulus , or random weak reactivation events , which activate a much smaller proportion of latent cells at any one time . Thus , it is not surprising that a much smaller number of cells is estimated to reactivate in vivo , than can be stimulated in vitro . Nonetheless , the in vitro quantitative viral outgrowth assay likely gives us a valuable measure of the size of the reservoir , as long as we recognize that only a fraction of these will actually reactivate in a given time . Given this potential for very prolonged quiescence of latent cells , it is not surprising that reactivation can be observed after prolonged periods of remission , as has been observed after transplantation and in the case of the Mississippi baby [29 , 30] . In our analysis we estimated a frequency of initiation of viral rebound for the different cohorts as if all patients in a cohort had the same frequency . However , recent studies have shown that reservoir size may vary substantially between patients and appears correlated with time to recrudescence after ART-interruption [20–22] , and it is highly likely that our patients also differed in reservoir size and rebound rate . To investigate this , we looked at whether time-to-detection was correlated for individual patients undergoing successive ART-interruptions in the SSITT trial ( cohort 4 ) . We found that time-to-detection was indeed correlated over multiple interruptions in individual patients ( Kendal’s concordance W = 0 . 47 , p = 0 . 013 ) . Thus , it seems likely that the frequency of reactivation we estimate for the cohorts represents the average frequency in the cohort , and there will be a distribution amongst individuals . In addition , we model reactivation as if it were the only mechanism affecting time to detection , and disregard the effects of viral growth because it is not correlated with time-to-detection . It is clear that differences in growth rates will inevitably affect time-to-detection , as slower growing virus will be seen later . However , unless the distribution of delays due to growth is large compared to the delays due to time-to-initiation , we would not expect growth to correlate well with time-to-detection ( as we have recently illustrated in the context of malaria infection [41] ) . There are clear limitations of our analysis of time-to-detection after ART-interruption , including the use of diverse cohorts , capturing patients at different times of infection , or after different interventions ( including the use of an LRA ) ( summarised in Table 1 ) . Our analysis was limited to ART-interruption studies with regular sampling after interruption , as this is required to capture the time-to-detection . Despite these obvious differences in the cohorts , we find very similar estimates of the average frequency of reactivation . These estimates were confirmed by a completely different approach , analyzing the ratios of ‘reactivation founder’ virus after ART-interruption . Our analysis suggests that much larger and more frequently sampled cohorts may be required to demonstrate differences in time-to-recrudescence amongst patients treated at different stages of infection or for different times ( consistent with the predictions of the power analysis ) . One apparent paradox of any estimate of frequency of reactivation is the observation of persistent low viral loads in patients on ART [42 , 43] . If infectious virus were continuously present , then there is no real concept of delay-to-reactivation ( and this argument applies equally to estimates of five reactivation events per day , or one every 6 days ) . However , the presence of reactivation founder virus suggests that viral growth is initiated by discreet reactivation events , rather than a constant ‘dribbling’ of virus from the latent reservoir [25] . Therefore it seems likely that the low levels of circulating virus detected under ART do not provide an immediate source of virus for reactivation . The development of therapies to purge the latent reservoir of HIV and produce prolonged antiretroviral-free HIV remission is a major priority . Understanding the frequency of recrudescence from latency is a crucial parameter in predicting the impact of interventions . Establishing the ‘normal’ rate of recrudescence from latency in HIV also allows us to assess the appropriateness of animal models and interventions , which can be judged on their ability to alter this parameter . A variety of approaches have been proposed to assess the effectiveness of latency reversing drugs . However , ultimately the test of LRA efficacy is the length of remission after ART-interruption . Future studies should determine the best predictors of time-to-recrudescence , so that these measures may be used as proxies to assess the efficacy of HIV eradication interventions .
This manuscript involves the analysis of previously published data from original human and animal studies published elsewhere ( summarized in Table 1 ) . Details of the ethical approval for the original studies may be found in the original publications . To study the dynamics of viral recrudescence , we assumed that the initiation of viral replication after ART interruption is a random event , occurring at a given frequency . Thus , the time-to-initiation will be exponentially distributed , and the proportion of patients without reactivation ( P ) will follow the equation: P=P0e−k ( t−t0 ) ( 2 ) Where P0 is the initial number of patients , k is the frequency of reactivation ( ie: reactivation occurs once every 1/k days ) , and t0 is the minimal time to detection ( as a result of ART-washout and the time taken for virus to grow from the initial level of viral infection to the level of detection ( summarized in Fig 1 ) . The equation was fitted to the data using the least squares method . In order to compare rates of reactivation between studies , we allowed the initial delay to detection to be an independent parameter for each study ( since both the ART drugs and threshold of detection varied between groups ) , and estimated the optimal frequency of recrudescence ( 1/k ) across all groups . To investigate whether the frequency of recrudescence ( 1/k ) was significantly different between groups we used an F-test . In order to estimate whether differences in viral growth could account for the observed delays to detection , we estimated viral growth rates from the viral load data , assuming exponential growth of the virus . We then investigated whether growth rate was correlated with time-to-detection , as would be expected if delayed detection occurred due to slower viral growth . In the first cohort of 9 patients sampled frequently , we used linear regression to estimate the slope of log-transformed viral load with time , using 2–5 sequential viral load measurements . In the second cohort of 59 patients sampled weekly , we estimated viral growth rate using a two-point estimate of the growth between the first and second positive viral load samples . Note that this may tend to underestimate viral growth if it growth slows as viral load increases . Moreover , there will be a tendency for patients detected with a lower viral load to have a faster growth rate ( because growth is measured at an earlier ( and thus faster ) stage of infection ) . The observation of lower viral loads in patients detected in the first week ( Fig 3E ) is likely an artefact of the pharmacokinetic delays before drug was fully eliminated and viral growth was possible in the first days after interruption . An additional assumption in our analysis is that the viral growth rate in plasma at the time of detection is reflective of ( or at least proportional to ) viral growth early after viral reactivation . In order to detect statistically significant differences of hazard ratios ( HR ) by Log-Rank Test with the level of significance α and power 1-β we need to have sufficient number of patients in each arm of the experiment . For estimation of this number we first need to estimate the number of events ( recrudescence of virus ) ( m ) and for this purpose used a formula , which assumes the equal number of patients in each arm [44]; m=4 ( zα/2+zβ ) 2/θ ( 3 ) where θ = Ln ( HR ) . However , if the rate of detection is not high enough to observe all patients in given time window of follow up , then the number of events will be lower than the total number of patients . Thus we need to correct the value of m by the fraction of patients with detectable virus at the end of the study . Assuming the exponential time to detection with the rate of detection estimated in our study ( k ) we can write the formula that relates the reduction in reactivation rate and the number of patients in one arm of the study . n=4 ( zα/2+zβ ) 2ln ( 1−p/100 ) 2 ( 1−e−kt ) ( 1−e−p100kt ) ( 4 ) where p is the percent reduction in reactivation rate , t is the time window of analysis . Sequence data on viral quasispecies after ART-interruption from the Joos study were obtained from Genbank ( Genbank accession numbers listed in the original publication [25] ) . The data were analysed using a ‘highlighter plot’ ( www . hiv . lanl . gov ) to identify the relationships between different viral species within a given patient ( see S3 File ) . Six patients were identified in whom we could distinguish and count the frequency of founder viruses early after ART-interruption , and this data was used to find the ratio of the number of copies of each observed founder virus in a patient to the next largest founder . To estimate the frequency of reactivation from the ratio of founder viruses , we assumed an exponential time-to-initiation of viral growth , and exponential growth of virus during the initial phase of infection . We can then write down a formula for the expected ratios ( R ) between the sizes of subsequent founders: R=V0egt1V0egt2=egΔ ( 5 ) Where g is the growth rate of virus ( = 0 . 4 day-1 ) , Δ = t1 ‒t2 is the delay between successive initiation events at times t1 and t2 , and V0 is the initial concentration of virus . The distribution of delays between the initiation of growth of different founders ( and thus their ratios ) will be determined by the frequency of initiation of viral growth after ART-interruption . We assume that that Δ has an exponential distribution with parameter λ and can then derive a formula for the probability density function ( PDF ) of the expected ratios ( h ( y ) ) using the formula for distribution function of a random variable . Where fexp ( λ , x ) is the probability density function ( PDF ) of the exponential distribution . The cumulative distribution function ( CDF ) of the ratios , H ( y ) , can be defined by: H ( y ) = Fexp ( λ , ln ( y ) /g ) ( 7 ) Where Ftrexp ( λ , x ) is the CDF of the exponential distribution . By using maximum likelihood estimation to fit the observed ratios between the number of copies of founders to h ( y ) we are able to estimate the rate of successful reactivation λ . We note that this analysis implicitly assumes that different founders grow at the same rate . It is also possible that individual founder viruses grow at different rates . However , as long as the growth rate is independent of the reactivation time , this should not significantly affect the expected distribution of founder ratios .
|
During treatment of HIV infection the virus persists in infected cells in a quiescent or ‘latent’ state . If treatment is stopped , then virus rebounds to detectable levels usually within 2–3 weeks . This is thought to occur due to release of infectious virus from a reservoir of long-lived latently infected cells . Reducing the number of latently infected cells should allow a prolonged period of HIV remission without antiviral treatment . A fundamental question is ‘how frequently does infectious virus emerge from the pool of latently infected cells ? ’ , and thus how much would we need to reduce the number of latently infected cells to produce remission ? Here we directly estimate the frequency of successful viral reactivation in four independent cohorts of patients undergoing treatment interruption . We find that active infection is initiated on average once every 5–8 days , considerably more slowly than previously thought . This has important implications for how much we need to reduce the number of latent cells in order to produce remission . Whereas previous analyses suggested that we would need to reduce the latent cell number 2000 fold to produce an average one-year remission , we show that reducing the latent cell number by 50–70 fold could achieve this aim .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
HIV Reactivation from Latency after Treatment Interruption Occurs on Average Every 5-8 Days—Implications for HIV Remission
|
It is a long-established fact that neuronal plasticity occupies the central role in generating neural function and computation . Nevertheless , no unifying account exists of how neurons in a recurrent cortical network learn to compute on temporally and spatially extended stimuli . However , these stimuli constitute the norm , rather than the exception , of the brain's input . Here , we introduce a geometric theory of learning spatiotemporal computations through neuronal plasticity . To that end , we rigorously formulate the problem of neural representations as a relation in space between stimulus-induced neural activity and the asymptotic dynamics of excitable cortical networks . Backed up by computer simulations and numerical analysis , we show that two canonical and widely spread forms of neuronal plasticity , that is , spike-timing-dependent synaptic plasticity and intrinsic plasticity , are both necessary for creating neural representations , such that these computations become realizable . Interestingly , the effects of these forms of plasticity on the emerging neural code relate to properties necessary for both combating and utilizing noise . The neural dynamics also exhibits features of the most likely stimulus in the network's spontaneous activity . These properties of the spatiotemporal neural code resulting from plasticity , having their grounding in nature , further consolidate the biological relevance of our findings .
Neuronal plasticity , both homeostatic and synaptic , is the central ingredient for the generation and adaptation of neural function and computation [1] . However , it remains mostly unclear how neurons in recurrent neural networks utilize neuronal plasticity to self-organize and to learn computing on temporally and spatially extended stimuli [2]–[4] . A full grasp of the principles of self-organization by plasticity in recurrent neural networks is jointly hampered by the diversity of existing neuronal plasticity mechanisms [5]–[7] and the limited understanding of their functions and cooperations , by the emergent nature of computation in recurrent systems , in the sense that computation is a collective phenomenon of the system as a whole and cannot be fully understood from the contribution of individual neurons [8] , [9] , and by the fact that neural systems are subject to noise [10]–[13] . In this paper , we simultaneously address these issues by studying the basic principles of self-organization in recurrent networks that arise from the interaction of synaptic and homeostatic intrinsic plasticity , and given that the network is subject to noise . To this end , we use numerical methods to explore the dynamics of nonautonomous , i . e . stimulus-driven , and plastic recurrent networks , and we provide a mathematical formalization for attaining a rigorously sound perspective ( see Methods ) . Incorporating synaptic plasticity with homeostasis goes back to Bienenstock , Cooper , and Monro's groundbreaking work known as the BCM theory [14] . Through rigorous mathematical analysis , the BCM theory predicted the necessity of a certain form of a sliding threshold , i . e . a homeostatic adjustment of neuronal excitability , for stabilizing the plastic afferent weights of a single neuron . Empirical findings supported the hypothesis of adjustable excitability and showed that it manifests through changes of neuronal properties at the soma [6] , [7] . While the BCM theory suggests homeostasis as a stabilization mechanism of synaptic weights with no direct influence on the neuron's encoding properties , Triesch proposed a homeostatic intrinsic plasticity ( IP ) mechanism that increases the neuron's encoding capacity and cooperates with synaptic plasticity ( SP ) to discover nonlinear independent features of the neuron's inputs [15] . These investigations , among others [16] , [17] , are very insightful in pinpointing how synaptic and homeostatic plasticity interact in single neurons . In addition , feedforward neural networks greatly simplify the analysis and understanding of self-organization and computation based on neuronal plasticity . For such architectures , both single plasticity rules , as well as combinations of different plasticity mechanisms , had been linked to neural computation , such as the formation of receptive fields [14] , the related identification of statistically-independent components [15] , [17] , [18] , and predictive coding [19] . However , it is important to note that neurons are embedded within large and highly recurrent networks [20]–[23] , and that an efficient use of neuronal resources entails distributed encoding schemes [8] , [9] . In addition , besides the spatial features of the world , its temporal structure should also be captured by the neural code [4] , [24]–[26] . Our understanding of neural information processing would greatly improve by extending the principles of self-organization to recurrent neural circuits , since the latter constitute the basic computational units in the cortex [22] . Lazar et al . were the first to study the emergence of computation from the interaction of different forms of plasticity on recurrent neural networks [27] , [28] . This study builds on their findings . However , we do not restrict the definition of computation to linear classifiers of the reservoir computing ( RC ) paradigm [3] , [29] , [30] . In addition to training linear classifiers for measuring the computational performance , we identified the necessity of analyzing the response of the recurrent neural network itself as an input-driven dynamical system [31] , [32] , and of concurrently viewing the network as a communication channel by taking an information-theoretical perspective [33] . Combining these tools enables us to understand how information is encoded in recurrent systems , how such encoding is developing from self-organization , and how noise is effecting both . Analyzing the dynamics of a large and , most importantly , input-driven neural system shaped by biologically-relevant plasticity is a hard task due to several methodological constraints . First , most analysis tools from dynamical systems theory are confined to small dynamical systems with very few degrees of freedom [34] . Exceptions are studies that circumvent this limitation by focusing on the low-dimensional collective dynamics of neural networks , e . g . , [35] , or studies that probe the high-dimensional phase space of the neural network , such as the classic example of Hopfield Networks [36] . Other instances of high-dimensional dynamical systems include ring networks and their coexisting periodic attractors [37] , stable heteroclinic orbits [38] , [39] , unstable periodic attractors [40] , and others [41]–[43] . The second and most important methodological constraint is that the use of standard dynamical systems theory is inappropriate , since it deals with autonomous systems only , i . e . systems with no explicit dependence on time . In reality , however , neural networks are subject to a flux of ever changing stimulation that renders them nonautonomous . A theory of nonautonomous dynamical systems is only recently taking shape as a branch of applied mathematics [31] , [32] . The fields of neural computation and computational biology are constantly contributing to the theory with concepts such as meta-transients and attractor morphing [37] , [44] , γ-systems [45] , and the nonautonomous dynamics of echo state networks [46] . A simple intuition of the difference between nonautonomous and autonomous systems can be stated as follows . Attractors of an autonomous dynamical system are defined by the system alone , and are therefore fixed . In contrast , attractors of a nonautonomous system are jointly defined by the dynamical system and its input . As the input changes , so does the attractor landscape of the system . This highlights the fact that studying computations in a driven system using the methods of autonomous dynamical systems is insufficient , since the input-induced changes of the system , i . e . changes of its attractor landscape , are ignored in that case . The third constraint is that the complexity of the dynamics increases due to the neural system's adaptability . The presence of plasticity imposes restrictions on the dynamics a network can exhibit , thus keeping the network dynamics in a regime that can support complex computations . To the best of our knowledge , no attempt prior to this work has been taken to combine high-dimensionality and nonautonomy with the consequences of plasticity on dynamics . We demonstrate that plasticity sculptures the stimulus-specific dynamic landscapes , and by that , serves in improving representation of the provided input . Moreover , neuronal plasticity can adapt and learn stimulus-induced sequences of such stimulus-specific landscapes . We thereby show that neuronal plasticity improves spatiotemporal computations . Given the above , we highlight and explain that spatiotemporal computations require two basic ingredients: a homeostatic mechanism that regulates neuronal activity , and synaptic learning that adapts the network's recurrent connectivity to the stimulus . We show that combining both types leads to a system that: first , learns the temporal structure of the input and carries out nonlinear computations , second , is noise tolerant , and third , even benefits from the presence of noise that sets the system to an input-sensitive dynamic regime . The paper is structured as follows . We first characterize the effects of self-organized adaptation that is based on synaptic and homeostatic intrinsic plasticity and their combination . For that , we use tasks where both random and temporally-structured inputs are reconstructed and predicted , as well as a task where nonlinear computations are performed . We estimate the network's self-information capacity ( its entropy ) , and its input-information capacity ( the mutual information between the input and the network ) . We then interlude to qualitatively analyze the resulting dynamics of plastic changes based on the theory of nonautonomous dynamical systems . We explain the superior computation of conjoining synaptic and intrinsic plasticity based on both the informational and dynamical analyses . Building upon that , we study network noise , and demonstrate how noise is combated and exploited through the interaction of synaptic and intrinsic plasticity .
The interaction of different forms of plasticity produces a rather complex emergent behavior that cannot be explained trivially by the individual operation of each . We therefore start with exploring the effects induced by the combination of spike-timing-dependent synaptic plasticity ( STDP ) and intrinsic plasticity ( IP ) . We compare the computational performance of recurrent networks trained either with both synaptic and intrinsic plasticity ( SIP-RNs ) , with synaptic plasticity alone ( SP-RNs ) , or with intrinsic plasticity alone ( IP-RNs ) , in addition to nonplastic recurrent networks , where the synaptic efficacies and firing thresholds are random . Following the plasticity phase , a network is reset to random initial conditions and the training phase starts . Output weights from the recurrent network to linear readouts are computed with linear regression so that the readouts activity is the optimal linear classifier of a target signal . The target signal depends on the computational task . That is followed by the testing phase , at which performance is computed . Performance is measured by the percentage of correctly matched readout activity to the target signal . Naturally , during simulation , the recurrent network is excited by a task-dependent external drive . The battery of tasks we deployed was designed to abstract a certain aspect of the spatiotemporal computations faced by biological brains , i . e . recalling past stimuli , predicting future ones , and nonlinearly transforming them . The memory task RAND x 4 , the prediction task Markov-85 , and the nonlinear task Parity-3 , as well as the plasticity models and simulation conditions , are detailed in the Methods section . Figure 2 shows that SIP-RNs significantly outperform both IP-RNs and SP-RNs in all tasks . Inputs from 3 time steps in the past are successfully retained far beyond chance level in the memory task RAND x 4 ( Figure 2A ) . Understandably , performance drops to chance level for future stimuli ( positive time-lags ) , since input symbols are equiprobable and their temporal succession carries no structure . Such is the case for the nonlinear task ( Figure 2C ) . It is worth noting that solving the nonlinear task Parity-3 requires recalling three successive stimuli , which adds to the computational load . The recurrent network , through learning the temporally-structured input of the task Markov-85 , boosts the readouts' ability to reconstruct past symbols in comparison to the structureless memory task RAND x 4 . It also allows for the prediction of future stimuli far beyond chance ( Figure 2B ) . STDP alone fails to provide the recurrent network with means to encode necessary information . This leads to SP-RNs performing at almost chance level in all tasks . Intrinsic plasticity , on the other hand , endues recurrent networks with an intermediate ability to sustain past inputs ( Figure 2A ) . IP-RNs also seem to learn the temporal structure of the input , as optimal linear classifiers are capable of predicting future stimuli ( Figure 2B ) . Intrinsic plasticity is , however , insufficient for nonlinear computations , as IP-RNs barely perform above chance in the nonlinear parity task . We also compare the performance of nonplastic kWTA networks with similar weight and threshold distributions as SP-RNs ( shown in gray in Figure 2 ) . They perform better than IP-RNs on the memory and nonlinear tasks , and worse on the prediction task . In all tasks , these nonplastic networks perform worse than SIP-RNs . We also show in Text S1 that nonplastic networks with comparable weight and threshold distributions as SIP-RNs also perform significantly lower than plastic networks . These results supply the evidence that the presence of plasticity enhances the computational power of recurrent neural networks ( see Text S1 for a discussion on heuristics for finding comparable random networks ) . No further analysis is carried out on these nonplastic networks , since the aim of this paper is to discern the effects of synaptic and intrinsic plasticity on spatiotemporal computations . Explaining the superiority of networks modified by deploying both STDP and IP starts from isolating the individual role of each plasticity mechanism in defining the spatiotemporal neural code . In that regard , a well-informed intuition is that STDP learns the basic structure of the input as the connectivity resulting from STDP reflects the input sequence transitions . IP , on the other hand , increases the neural bandwidth by introducing redundancy to the code , as IP leads to the longest periodic cycles in the spontaneous activity of kWTA networks ( See Figure 8 and Figure 4A in [27] ) . The spatiotemporal neural code , or the neural code for short , can be characterized by both the absolute capacity of the network activity to store information and by how network activity encodes the spatially and temporally extended network input . Entropy of the network activity measures its absolute capacity , i . e . the repertoire of network states that the network can actually visit and potentially assign to some input sequence . The assignment of a network state to an input sequence means that this particular network state encodes or represents that input sequence . Mutual information between network input sequences and network states quantifies the extent of how successful this assignment is . Not every visited network state needs be assigned an input sequence . A redundant code is reflected by input sequences being represented by multiple network states . Also , a network state might fail to encode an input , thus reflecting uninformative noise states . We investigate the neural code characteristics of kWTA networks by estimating both the entropy of the network state and the mutual information between network input sequences and network states . We drive the network by RAND x 4 input , and for computational tractability , we limit the estimation of mutual information to three-step inputs . An optimal encoder of this input sequence will then be a network with 6 bits of mutual information . The information-theoretical quantities are computed at intervals of the plasticity phase under the three plasticity conditions . At these intervals , the plastic variables are fixed and the driven network is reinitialized and run for a sufficient number of steps , and passed along with the input to the entropy and mutual information estimators . More details on how these measurements are carried out are found in the Methods section . Figure 3 shows how these measures develop through the plasticity phase ( For a discussion on the effects of longer plasticity exposure , see Text S2 ) . SP-RNs' entropy remains constant at 2 bits . This means that SP-RNs visit only 4 network states ( green in Figure 3A ) . However , these network states encode no information of the input sequence , as mutual information remains practically zero ( green in Figure 3B ) . We call this 2 bits input-insensitive code the minimal code , as it captures no more than a single possible succession of the 4 inputs . This effect is the result of the interaction between the machination of STDP and the initial firing thresholds and weights configuration . Transitions , such as in the input space , are to be stored in some of the synapses that connect neurons in the receptive field of with those in the receptive field of . At each time step , one transition , such as , could be easier to reinforce with the causal ( potentiating ) side of STDP for neurons having little higher excitability ( internal drive plus their own firing threshold ) . Without IP to tune down this excitability and with further contribution from the recurrency of the network , a positive feedback loop is generated , and this transition becomes more and more potentiated at the expense of others . This transition then becomes independent of the actual drive the network is receiving: the network becomes input-insensitive . On the other side of the entropy spectrum , we find IP-RNs . Through IP's constant adjustment of the neuronal excitability , many neurons contribute to the neural code and IP-RNs visit a large number of states . Entropy and the network state bandwidth are the highest ( blue in Figure 3A ) . One may view IP's effect as an introduction of intrinsic deterministic noise to the network activity . The increase in bandwidth of the network activity raises the odds for the random weights of an IP-RN to store an input sequence . In fact , many network states encode the same input sequence , resulting in a redundant code . However , without a synaptic reinforcement of representations , many states are visited due to the internal dynamics of the network , and not due to the external drive . These states remain uninformative and input sequences not successfully encoded: the mutual information ( blue in Figure 3B ) , and hence the classification performance , are low . The development of the neural code for SIP-RNs follows , however , a more interesting path . At the beginning , STDP has the upper hand and a 2 bits minimal code is generated . Through providing intrinsic deterministic noise , IP enriches the neural code by increasing redundancy and entropy ( orange in Figure 3A ) . At the same time , STDP incrementally associates different network states to different input sequences by adjusting the synaptic weights as seen from the increase of mutual information ( orange in Figure 3B ) . Then together , synaptic and homeostatic plasticity cooperate to create a code that is both redundant and input-specific . These properties are crucial for noise-robustness , as will be shown later in this text . A dynamical system's behavior depends on its past activity . Therefore , testing a system requires assuming plausible initial conditions . The recurrent neural network at hand , even though it is small in comparison to a real neural circuit , has a number of possible initial conditions too large for all its initial conditions to be tested . So far , we have chosen random initial conditions for the network activity following the plasticity phase . From now on , we choose the initial conditions systematically by reinitializing the network activity depending on a perturbation . This perturbation is applied to the end state of the plasticity phase , such that the end state of the plasticity phase and the initial state of the training phase are at a distance from one another . For details of how the initial conditions are selected depending on the parameter , we refer the reader to the Methods section . To discern the effect of this perturbation , we compute the performance of the trained system with the three combinations of synaptic and intrinsic plasticity . We do this both for a system that is perturbed and for a system that starts from the last state that the dynamics reaches at the end of the preceding plasticity phase . We find no difference between the two cases of initial conditions for either IP-RNs or SP-RNs . However , when the neural network is trained by both synaptic and intrinsic plasticity ( SIP-RNs ) , we find that the perturbed networks have better performance , as is illustrated in Figure 4A–C . The high performance of SIP-RNs that results from random initial conditions , as is shown in Figure 2 , is easily explainable . It stems from the fact that random initialization is merely a large perturbation , since the probability of choosing a random state from such a large set of possibilities that is at a small distance from a particular region of the state space is insignificant , compared to a state that is at a large distance . Moreover , we find that regardless of the task , larger perturbations result in higher average performance . This is also reflected in the neural code , where network state entropy and the mutual information with input correlate with higher perturbation ( see Figure 4D–E ) . This suggests that within the phase space of SIP-RNs there exist at least two dynamic regimes . Post-plasticity perturbation also provides the first sign of how SIP-RNs can benefit from noise , as it might put the system in the regime more suitable for computation . Optimal linear classifiers show that kWTA networks equipped with both homeostatic and synaptic plasticity are capable of creating spatiotemporal codes and performing nonlinear computation . Measuring entropy and mutual information allows for a quantification of the emerging neural code . But what are the geometric features of the neural code that allow for such computations ? How do network states represent the spatiotemporal input in a useful way ? A major part of the Methods section is devoted to developing the mathematical formalization of discrete-time nonautonomous dynamical systems . References to definitions , a proposition , and a theorem from that section are featured in the following results , as we apply these concepts to our model neural network . We view this treatment not merely as an exercise in mathematics . It allows for a rigorous description of the computational properties emerging from plasticity that are beyond the scrutiny of quantitative measures , such as linear classification performance and carried information . A consequence of these properties is also the two noise-related features we examine later . For a formal treatment of spatiotemporal computations which result from plasticity , we need to extend the theory of nonautonomous dynamical systems to provide a notion for representations , to specify how these representations allow for computations , and to discern the effect of plasticity in enhancing these representations for the sake of computation . But first , we start by identifying the modes of operation , i . e . the dynamic regimes , the model plastic neural network has , since not all regimes might be suitable for computation . According to Proposition 3 and Definition 6 , when subject to stimulation , kWTA networks are input-driven discrete-time dynamical systems . For such systems , two extremes exist regarding the degree of sensitivity the system exhibits in response to its input . At one extreme , the system shows no change of response for different inputs , so that it follows its own dynamics , as if no input exists . In such a mode of operation , the system is input-insensitive . The other extreme is when the system's response is different for each input and initial condition . A single system can show , in principle , multiple modes of operation , depending on the initial conditions . The set of initial conditions that show a single mode of operation defines a dynamic regime and a basin of attraction . In a first step , we visualize the high-dimensional response of the system to its input . To that end , we down-project the network activity to the first three principal components , and we study the effects of STDP and IP on the network's dynamics and input representations in this reduced 3-dimensional space ( Figure 5 ) . This analysis is performed on networks with Markov-85 input which fully demonstrate the relevant properties . It is important to note that while our analysis concerns the dynamics following the plasticity phase , we are still able to infer how it unfolds during this phase from the development of the neural code ( Figure 3 ) , as we make clear later . As suggested by the performance of SP-RNs ( Figure 2 ) and their neural code entropy and mutual information ( Figure 3 ) , their state space is dominated by an input-insensitive basin of attraction and these networks behave like autonomous semi-dynamical systems ( prefixing with “semi” refers to the fact that the dynamics needs not be invertible ) . This is confirmed by the asymptotic dynamics of SP-RNs , which is independent of the input ( Figure 5B ) . The dynamics within this dynamic regime follows the minimal code . The minimal code manifests itself through a period-4 periodic attractor which corresponds , in the case of Markov-85 input , to the most probable transition in the input space . This observation confirms the fact that STDP allows the system to learn the basic structure of its input . SIP-RNs exhibits similar dynamics at the end of the plasticity phase ( Figure 5C ) . However , as is evident from varying the perturbation parameter for SIP-RNs ( Figure 4 ) , the set of initial conditions that constitutes this input-insensitive basin is confined by a distance relation to the neighborhood of the periodic attractor: the probability of being in this basin diminishes the further away the initial conditions are from the input-insensitive periodic attractor . The increase of performance and the neural bandwidth of SIP-RNs for higher ( Figure 4 ) shows that outside of the input-insensitive dynamic regime there exists a different basin of attraction . Within this basin , the network is sensitive to input , and computations are possible . The observation that has no effect on IP-RNs and that they show intermediate performance and mutual information suggests that they are dominated by a dynamic regime with intermediate input-sensitivity . It also confirms that intrinsic plasticity is responsible for the emergence of the input-sensitive dynamic regime in SIP-RNs . Now that the dynamic regimes of trained networks with the three combinations of synaptic and intrinsic plasticity are identified , we next move to formulating the notion of representations inside the input-sensitive dynamic regime . Developing such a notion allows linking the theory of nonautonomous dynamical systems to a theory of spatiotemporal computations . To this purpose , we coin the term volumes of representation , which is a concept that describes the response of a nonautonomous dynamical system in respect to its drive . The volume of representation of some input sequence within some dynamic regime is the set of network states that are accessible through exciting the network with the corresponding input sequence , starting from all network states in this dynamic regime as initial conditions ( Definition 10 ) . The order of a volume is defined by the length of the input sequence it represents . We also introduce the volumes' inclusion property which hierarchically links the system's response to spatiotemporal input sequences to their sub-sequences . To visualize a network's volumes of representation , we sample the network's response . We do this because the size of the state space and the input-sensitive dynamic regime is too large , making a complete coverage impossible . Also , since volumes of representation can have complicated shapes in both the full and reduced state space , we approximate these volumes with ellipsoids . Figure 5D provides such an approximation to the volumes of representation of order-1 . The sample is a single sequence of 10000 Markov-85 inputs to a SIP-RN . Each volume is replaced by an ellipsoid . The center of this ellipsoid is the coordinates' average of the visited network states in the principal components space . Each of its semi-axes has a length that is the standard deviation from the mean of the corresponding coordinate . Also , according to the volumes' inclusion property , stated formally in the Methods section , a volume of representation of order-1 of some input includes all volumes of order-2 for sequences whose most recent input is . As such , Figure 5E , that depicts a similar approximation to all volumes of order-2 , is also a better approximation to volumes of order-1 . In Figure 5E , each order-1 volume consists of four order-2 volumes that are color-coded to match the rougher approximation in Figure 5D . In a supporting figure , we further show that this way of presentation is sufficient , compared to using percentiles of bootstrapped network states ( see Figure S1 ) . The volumes of representation provide a geometric view of spatiotemporal computations as the ability of the recurrent neural network to represent in its activity , in other words to encode , useful functions of the network's input sequences , and for these representations to be distinguishable and reliable . In the case of the tasks RAND x 4 and Markov-85 , the functions that the network activity represents are the identity , delayed or forecast . As shown in Figure 5D–E , the volumes of representation of SIP-RNs under Markov-85 input exhibit higher separability , which explains both their high classification performance and high mutual information . One also notices that the volumes of representation of order-2 that belong to the most probable transitions in the Markov-85 input , e . g . , , are also the most distant from one another ( Figure 5E ) . This results in the most probable transitions to be more easily distinguishable by optimal linear classifiers . In order to isolate the roles of synaptic and intrinsic plasticity in generating useful representations , we show in Figure 5A the order-1 volumes of representation of an IP-RN in response to Markov-85 input . Compared to the SIP-RN , these volumes are highly overlapping , which explains the lower classification performance . Also , the low mutual information between the network state and the input ( Figure 3 ) can now be explained by various network states belonging to multiple volumes of representation , at once . Also , many network states represent the same single input which is a signature of redundancy resulting from IP . These observations point towards STDP being the source of separability of representations in SIP-RNs , in addition to learning the structure of the input through situating the representations of the input's most probable transitions at further distances from one another . In the case of the task Parity-3 , the function that the network activity needs to represent is the sequential exclusive or operation over three successive binary inputs . As such , within the input-sensitive dynamic regime , two volumes of representation exists , each encodes one outcome of the nonlinear task Parity-3 . According to Definition 10 , these volumes are formed from an appropriate union of order-3 volumes of representation of the binary input . We provide an illustration of these two volumes of representation in Figure S2 . Here also , STDP provides the separability that allows these representations to be distinguishable , while IP gives the possibility of an input-sensitive and redundant regime to emerge , and , aided by STDP , for the volumes of representation to expand . The presence of dynamic regimes entails the existence of attractors , i . e . limit sets of the dynamics , that apply a pulling force on the dynamical system's activity and dictate its course of flow . In an input-driven dynamical system , attractors are not easily defined as sets of states . Instead , nonautonomous attractors are input-dependent moving targets of the dynamics , which adds a temporal aspect to their definition ( see Definition 8 ) . As follows , for our nonautonomous dynamical systems theory of spatiotemporal computations to be complete , we link the geometry of the computational entities , i . e . the volumes of representation , to the geometry of the nonautonomous attractors . This allows us to connect the features of the volumes of representation emerging from plasticity , namely , separability and redundancy , to the effects of plasticity on the nonautonomous attractor . To that end , starting from the volumes of representations , we define the perturbation set ( Definition 10 ) as a moving source of the neural activity towards its moving target , the nonautonomous attractor . Since the perturbation set changes with time , it is called a nonautonomous set ( Definition 7 ) . This also applies to nonautonomous attractors . The set of states constituting a nonautonomous set at a fixed time is called the set's t-fiber . We later show how the t-fibers of these nonautonomous sets relate to each other . In the input-insensitive dynamic regime , the dynamical system behaves as an autonomous dynamical system , and so does its attractor , which is the period-4 attractor in Figure 5B–C . In addition , the existence of a nonautonomous basin of attraction ( Definition 9 ) , that constitutes the input-sensitive dynamic regime in SIP-RNs , necessitates the existence of a nonautonomous attractor . It is not possible to fully identify the nonautonomous attractor by looking into the nonautonomous dynamics . This is because the attractor is not fixed in space and because the dynamics almost never converges to it . However , we prove in Theorem 11 . 1 that in an input-driven discrete-time dynamical system , and within a basin of attraction , the nonautonomous attractor is a subset of the basin's perturbation set , and that the t-fibers of a nonautonomous attractor are subsets of the t-fibers of the perturbation set . Given this result , the location of the nonautonomous attractor within the state space of the network can be approximated by the perturbation set . The perturbation set summarizes how the network activity passes from one volume of representation to another , at every time step , according to the input's transition statistics . We replace the time dimension in Figure 5D by arrows that correspond to the transitions in Markov-85 input . The volume of representation visited at time is the volume corresponding to the input at that time , and it forms the t-fiber of the perturbation set . Instead of defining the asymptotic dynamics of the model neural network within the input-sensitive basin of attraction by a single nonautonomous attractor with different t-fibers , we can define it by multiple autonomous attractors , each belonging to a particular input . According to Theorem 11 . 2 , within the input-sensitive basin of attraction , there exists for each input , an autonomous attractor ( Definition 4 ) of the autonomous semi-dynamical system defined by . The theorem also shows that this attractor is a subset of the volume of representation of . Theorem 11 . 3 further shows that the basin of attraction of the autonomous attractor is also the input-sensitive basin . Accordingly , the network dynamics undergoes a bifurcation at each time step the input changes its identity . A bifurcation is a change in the topological properties of invariant sets , such as attractors . We observe bifurcations in the input-sensitive regime of kWTA networks . The topological property undergoing the change is the loss of stability of the periodic attractor associated with an input , and the appearance of an attractor with a different period and location that is associated with the input . Figure 5F shows the autonomous periodic attractors associated with each Markov-85 input within the input-sensitive basin of attraction of a SIP-RN . Each of these attractors is also a t-fiber of the nonautonomous input-sensitive attractor . While these autonomous attractors are depicted in one state space , overlaying them in a single plot serves only in illustrating the geometric relations between them . In reality , these attractors do not coexist . Each autonomous attractor appears in the phase space of the network when its associated input drives the network , and the attractor from the previous time step disappears . The geometry of the nonautonomous attractor within an input-sensitive dynamic regime is very important regarding spatiotemporal computations . In fact , computations are completely defined according to the relative positions of the nonautonomous attractor's t-fibers to one another , and to the volumes of representation . An attractor consists of limit points of a basin of attraction . Thus , it exerts a pulling force on the network states that define the volumes of representation . So , if the t-fibers of a nonautonomous attractor are close to one another in the state space of the network , different volumes will be overlapping and computations will be difficult to carry through . Such is the case in IP-RNs . On the other hand , distant t-fibers of the nonautonomous attractor result in separate volumes of representation and better spatiotemporal computations , which is the case in SIP-RNs ( Figure 5D–F ) . Also , the number of states comprising the t-fibers of the nonautonomous attractor effects the redundancy of representations . As intrinsic plasticity increases the number of states of these t-fibers , the perturbation set becomes more redundant . Given the above , while the perturbation set contains the nonautonomous attractor , it is the attractor that defines how the perturbation set , and as a consequence the volumes of representation , extends in space . For a correct characterization of spatiotemporal computations according to the geometry of the nonautonomous attractor and function representations , we borrow the concept of meta-transients [44] . A transient activity of an autonomous ( semi- ) dynamical system is the trajectory its dynamics follows as it approaches a fixed attractor . Alternatively , an attractor of an input-driven dynamical system changes constantly . This leads the trajectory pursued by the dynamics to switch its course , so as to keep track of its moving target . Such an input-dependent trajectory is termed a meta-transient . When the input changes , the meta-transient passes from one volume of representation to another , i . e . the dynamics bifurcates and the meta-transient approaches the vertexes of the current attractor , while being repelled from the others that are now unstable . It is in this geometric relation to the different attractors ( or t-fibers ) that computation resides . In fact , as a proof of principle , the autonomous attractors of SIP-RNs were allocated . This was done by clamping each input for a sufficient time until the dynamics converges to that input's periodic attractor . Then , optimal linear classifiers were fitted to perform the three computational tasks . As training data , the Hamming distances between the meta-transient and the vertexes of these periodic attractors were used . Figure S3 shows the performance resulting from this computational procedure , which outperforms both SP-RNs and IP-RNs . While the performance is far from what is achieved directly from the activity of SIP-RNs , especially in the nonlinear task Parity-3 , it is important to note that distance is a very rough compression of the geometric relations between the meta-transient and the autonomous attractors . For instance , distance does not allow the distinction between network states that are symmetrical in relation to the autonomous attractors . We now outline how the interaction of homeostatic and synaptic plasticity gives rise to spatiotemporal computations through developing useful representations . To this end , we combine the analysis of dynamic regimes , volumes of representation , and autonomous and nonautonomous attractors ( Figure 5 ) with the informational-theoretic intuitions regarding the evolution of the neural code ( Figure 3 ) . At the beginning of the plasticity phase , STDP has the upper hand and it generates a minimal code of the input . This is evident from the 2 bits network state entropy ( Figure 3A ) and the close to zero mutual information with input ( Figure 3B ) at the beginning of the plasticity phase of SIP-RNs . The minimal code captures , through an input-insensitive periodic attractor , the most probable transitions in the input ( Figure 5B ) . Another feature of the input-insensitive periodic attractor is the high separability of its vertexes in the state space of the SIP-RN . At the same , IP time succeeds in reducing the excitability thresholds of some neurons , such that more network states become accessible at the vicinity of the vertexes of the input-insensitive attractor: entropy increases alongside the potential for redundancy . STDP concurrently assigns these network states to the inputs that induce them: mutual information and redundancy increase . This incremental process manifests dynamically in the appearance of the input-sensitive basin of attraction , and the associated appearance and expansion of volumes of representation ( Figure 5D–E ) . Due to the highly separate vertexes of the input-insensitive attractor and the neighborhood relations of the volumes with these vertexes , the volumes of representation are highly separate . This shows that the input-insensitive dynamics is a necessary prerequisite for the emergence of spatiotemporal computations , as it sets the stage for the appearance of separate representations that also carry the structure of the input . The emerging dynamics can also be viewed through formulating the SIP-RN during the plasticity phase , as an input-driven dynamical system parametrized by the weights and the excitability thresholds . Through varying the parameters of the system with STDP and IP , the dynamics at some point in the parameters space bifurcates from one stable dynamics , the input-insensitive dynamics , to two stable dynamics with the appearance of the input-sensitive attractor in whose basin computations are realizable . This also applies to each member of the family of semi-dynamical systems with the appearance of new dynamics and the associated new periodic attractor ( Figure 5F ) . Equipped with different vantage points to describe the information processing properties of plastic recurrent neural networks , we now turn to ask a central question: what does an information processing system like the brain require in order to be noise-robust ? We state the following hypothesis . Noise-robustness is an effect of the interplay between 1 ) a redundant code that provides multiple possible encodings of an input , and 2 ) separability of representations which allows for a margin of noise without obscuring the identity of the input . The analysis of the neural code ( Figure 3 ) shows how IP increases the potential for redundancy by increasing the neuronal bandwidth . STDP could exploit this potential redundancy by assigning multiple neurons to the same input . Viewing the network dynamics in the principal components space , on the other hand , made clear that STDP ensures separability in the volumes of representation ( Figure 5D–E ) . This also suggests that the recurrent network should be more robust to noise , the more recent the decoded input is , as the margin of noise becomes smaller for older inputs . The expansion of volumes of representation in IP-RNs also points towards a higher potential redundancy . We test the hypothesis and the role of STDP and IP interaction in noise-robustness by injecting nondeterministic noise into the recurrent network . Following the plasticity phase , we deploy a certain rate of random bit flips on the network state that reserves the kWTA dynamics , i . e . if some neuron is silenced due to noise , another neuron is selected at random and it fires instead . Different networks with different input statistics will amplify the same amount of noise to a varying extent . The shaded area in Figure 6 marks the ratio-of-noisy-spikes range within the network states of 100 recurrent networks . For all tasks and networks , we measured performance of optimal linear classifiers on both the noise-free and noisy network states , and computed the relative change in performance . We compare the change in performance for each time-lag with the ratio of noisy spikes . To understand how this comparison aids in characterizing noise-robustness , we rely on an example . If 10% of a network's spiking activity has been replaced by noise , spikes being the carriers of information , 10% of the information in the network would be lost . However , if the activity of other neurons within the network is a replica of half the lost spikes , only 5% of the information would be lost , and the performance of the linear classifiers would decrease just as much . Having the change of performance below noise level is evidence of noise-robustness due to redundancy and intrinsic plasticity . Information carried by the network cannot deteriorate beyond the amount of noise; the ability to perform computations , on the other hand , is another story , since distinguishing between representations is a necessary condition for computation . Noise can lead to an overlap in the volumes of representation , which hinders the information processing capability of the recurrent neural network , since overlapping representations are indistinguishable and prone to over-fitting by decoders , linear or otherwise . However , when volumes of representation are well separated due to STDP , and redundancy is at play , performance will not exceed the amount of noise in the network: noise-robustness is still achieved . Figure 6 shows that redundancy and separability are assuring noise-robustness in the three tasks . The effects are the strongest for the task RAND x 4 . The change of performance never exceeds the range of noise for all time-lags . The change of performance on the task Markov-85 remains below the range of noise for few time-lags in the past and it remains within the bounds of the noise range for older stimuli . The networks then are still capable of tolerating noise , while the volumes of representation are becoming more overlapping . The decrease of noise-robustness for larger time-lags in the past confirms our suggestion that volumes of representation become less separate for older inputs . The analysis of order-2 volumes of representation ( Figure 5E ) also suggests that less probable transitions of the input are more prone to noise . This , however , was not tested . The task Parity-3 is noise-robust for 0-time-lag only and with the change in performance being within the noise range . This is understandable , since for each time-lag , order-3 volumes of representation and the associated volumes of the Parity-3 function should be separate and redundant . These observations confirm our hypothesis that redundancy and separability are the appropriate ingredients for a noise-robust information processing system , such as our model neural network . These properties being the outcome of STDP's and IP's collaboration , suggest the pivotal role of the interaction between homeostatic and synaptic plasticity for combating noise . Now that we have demonstrated the contributions of STDP and IP in combating noise , we turn to investigating noise's beneficial role . We have seen that perturbation at the end of the plasticity phase provides a solution to the network being trapped in an input-insensitive regime . Besides viewing perturbation as a form of one-shot strong noise , which is , biologically speaking , an unnatural phenomenon , what effect would a perpetual small amount of noise have on the dynamics of the recurrent neural network ? We again deploy a certain rate of random bit flips on the network state that reserves the kWTA dynamics . Unlike the previous section , we do not restrict noise to the training and testing phase , but apply it also during the plasticity phase . We also do not reset the network activity after the plasticity phase , i . e . the perturbation parameter is set to 0 . Figure 7A–C compares the performance of optimal linear classifiers on the three tasks for different levels of noise . For all tasks , some levels of noise resulted in a significantly higher average performance than the noiseless case . The task Markov-85 had the highest average performance at the largest level of noise , while the tasks RAND x 4 and Parity-3 , where the input was uniformly random , had the highest performance at the third and fourth levels of noise , and the average performance dropped substantially at the fifth level of noise . In all tasks , performance was far off the levels it reached in the noiseless case ( Figure 2 ) . Information-theoretical quantities are again measured on networks with RAND x 4 input . As expected , the network state entropy increases monotonically with noise ( Figure 7D ) . Mutual information , on the other hand , starts dropping for noise larger than the third level ( Figure 7E ) . This is also expected from the change of performance ( Figure 7A ) . Noise then appears to provide , in some of the SIP-RNs , the necessary means to escape the input-insensitive dynamics . At some levels , however , the network activity becomes dominated by noise beyond the compensatory effects of redundancy and separability achieved through plasticity . In addition , more unstructured noise during the plasticity phase delays the creation and expansion of useful volumes of representation , thereby hindering computations further .
We demonstrated how the interaction of synaptic learning and homeostatic regulation boosts memory capacity of recurrent neural networks , allows them to discover regularities in the input stream , and enhances nonlinear computations . We provided a geometric interpretation of the emergence of these spatiotemporal computations through analyzing the driven dynamic response of the recurrent neural network . We view computations as a geometric relationship between representations of functions over stimuli , representations that consist of network states , and the asymptotic dynamics of the network , i . e . attractors . Accordingly , Figure 8A shows a possible driven-dynamics viewpoint on computation , which is the following . As the stimulus changes , a bifurcation occurs where the current attractor of the network becomes unstable , while another stabilizes according to the current stimulus . That leads the network dynamics to change its course towards the new stable region , or attractor , of the state space , and away from the previous attractors that are all unstable . As such , this path of the network activity , i . e . the meta-transient [44] , is defined by both the stimulus sequence and the locations of the network's attractors . Together , they lead the meta-transient to pass through particular representations which encode computations . An equivalent alternative to the chain of bifurcations between autonomous attractors is that of a single nonautonomous attractor that behaves as a stimulus-dependent moving target of the dynamics . We showed that a successful implementation of these spatiotemporal computations requires the interaction of synaptic and homeostatic intrinsic plasticity which generates useful representations in the dynamics of excitable cortical networks . Figure 8 schematically illustrates the stimulus-driven dynamical viewpoint of spatiotemporal computations and the effects of plasticity . Synaptic plasticity produces stimulus-insensitive dynamics that captures the temporal structure of the input . Intrinsic plasticity increases the neuronal bandwidth by increasing sensitivity to stimuli , which reduces the dominance of the stimulus-insensitive dynamics . This , in combination with synaptic plasticity , generates stimulus-sensitive attractors and redundant representations around them . These stimulus-sensitive components are pulled apart by the stimulus-insensitive dynamics , so that the structure of the input is preserved , and the separability of representations is higher and computations are realizable . We pointed out throughout the text that computation is an emergent property of the recurrent network , and that it cannot be fully understood from the individual contribution of the parts , be it neurons or plasticity mechanisms . It might appear contradictory to that statement that the analysis was often concerned with the isolated role of each single plasticity mechanism . However , the quantitative assessments of computations point back to the emergent and collective aspect of computation . Namely , measured on SIP-RNs , neither performance of linear classifiers nor mutual information with input can be accounted for by a linear relationship between the respective quantities measured on SP-RNs and IP-RNs . In fact , the performance of networks where the recurrent weights and firing thresholds are adapted separately , and then combined following the plasticity phase , is far less than the performance of SIP-RNs , where intrinsic and synaptic plasticity are mutually active ( see Figure S4 ) . This further consolidates the claim that computations in SIP-RNs emerge from the interaction of STDP and IP , and not from their isolated contributions . It also points back to the formation of separate and redundant representations from the continuous interplay of these two mechanisms . We also illustrated the combined role of synaptic and homeostatic intrinsic plasticity in creating noise-robust encoding through the generation of a redundant neural code . Many studies have investigated the redundant nature of neural information transmission in many cortical regions , and have justified this expensive allocation of neural recourses by redundancy serving as an error-correction strategy that provides neural assemblies with the capacity to average out noise [10] , [47]–[50] . Tkaik and colleagues have shown that in the presence of noise , a maximum entropy model of the retina increases redundancy for higher noise levels . A side effect of their model is that stimulus representations become highly separate , which increases the tolerance margin of noise and enhances information transmission [51] . Our model was able , through local plasticity mechanisms , to capture both of these properties , achieved in [51] through optimality principles , and to lead to a noise-robust population code ( Figure 8B ) . Namely , synaptic plasticity enhances the separability of representations through the pulling force of the input-insensitive attractor , while intrinsic plasticity perturbs the network states and increases redundancy when interacting with synaptic plasticity , which allows for alternative representations of similar input sequences . Another point of similarity with the model of Tkaik and colleagues [51] and with empirical findings [52] , [53] is the remnant fingerprint of the most common stimulus in the network's spontaneous activity , which manifests in our model neural network in the stimulus-insensitive dynamics ( Figure 5B–C ) . In addition to combating noise , our model explores a potential benefit from its presence . We pointed out the necessity of the stimulus-insensitive dynamics for the emergence of computation in the model neural network . The stimulus-insensitive attractor provides the baseline dynamics for the appearance of highly separate representations , and thus , the excitable dynamics necessary for computations . Getting from the input-insensitive regime to the excitable one depended , however , on the ad hoc reinitialization of the network activity at the end of the plasticity phase . Noise provides an alternative . During the plasticity phase , noise shallows the boundaries between the two basins of attraction , which reduces the dominance of the stimulus-insensitive attractor . After the plasticity phase , noise supplies the small perturbations needed to get the network activity to the sensitive dynamics where computations are possible . This solution , in comparison to reinitializing the network activity , is more inferior , specifically because noise also delays the learning of representations . We postulate that another homeostatic plasticity mechanism , synaptic scaling , might contribute to the shallowing of the attractor boundary by constraining the strength of synapse bundles between neural subpopulations ( e . g . , between and ) . For instance , synaptic scaling was necessary for implementing spatiotemporal computations in self-organizing recurrent networks ( SORN ) [28] , but no analysis of the dynamics of these networks was done . Testing this hypothesis is , however , beyond the scope of this work . It is also tempting to connect the topology of the attractor landscape of SIP-RNs to neuropathology and to a model by Pfister and Tass [54] . They suggest that two stable regimes of recurrent network activity , a synchronous pathological regime and an asynchronous healthy regime , coexist , and that their coexistence is a necessary condition for the functioning of a model of deep brain stimulation . In their model , the stimulation of the recurrent network destabilizes the synchronous dynamics through inducing STDP . The destabilization drives the network activity towards the healthy asynchronous basin of attraction . By eliminating the stimulation , the energy hill between the two dynamic regimes rises again and the network remains in the healthy dynamics . Our study has shown how these two coexisting dynamic regimes and their associated forms of activity might come into being through neuronal plasticity . We also suggested noise as a possible mechanism for avoiding the unhealthy dynamics . Further analysis is necessary to investigate how the interaction between noise and different plasticity mechanisms might contribute to our understanding of neurological disorders . Our analysis of spatiotemporal computations was restricted to Markovian dependencies in the temporal structure of the stimulus or to no dependencies at all . This is often not the case in natural stimuli faced by animals and humans , where the Markov property does not always hold . Lazar et al . have shown that SIP-RNs are capable , to a certain degree , of performing predictions on second-order Markov chains [27] . However , optimal encoding of non-Markovian stimuli and performing computations over them require forms of spike-timing-dependent plasticity that are less myopic to the temporal dependencies than what we considered in this work ( Figure 1B ) . For instance , Brea and colleagues have shown that storing and reproducing a non-Markovian sequence in a recurrent neural network require a nonlocal form of STDP with more complex temporal dependencies between pre- and post-synaptic spikes [55] . While their model was not concerned with carrying through spatiotemporal computations of the kind we presented here , it successfully reproduced the stored non-Markovian input in the spontaneous activity of the neural network . This refers to a point of similarity to the simpler case we presented here , where Markovian input was stored and recalled in the spontaneous activity of the input-insensitive dynamics . In any case , while spatiotemporal computations over non-Markovian stimuli and the necessarily more complex plasticity mechanisms that lead to their emergence , are not considered here , we view the concepts and methodology developed above as a general framework for future studies . In this article , we provided a first analysis of the combined role of synaptic and intrinsic plasticity on the emergent dynamics of recurrent neural networks subject to input . We redefined computations in relation to these emergent dynamics and related that to properties of the neural code . We also considered how the neural dynamics interact with noise , both as a nuisance to combat , and as a driving force towards healthy neural activity . The model we used is simplified , however , both in network architecture and plasticity mechanisms . While this simplification is necessary for mathematical convenience , biology never cares for formal abstractions , for the brain is a complex information processing system that is rich with a variety of neuronal morphologies and functions . The plastic changes the brain undergoes are neither confined to the two mechanisms we dealt with here , nor are they uniform across different regions . On the other hand , mathematical formalization of computation and adaptability allows the identification of unifying principles in computational biology , in general , and neural computations , in particular . We intended the current article as a step in that direction .
In this paper , the model recurrent network is of the k-Winner-Take-All ( kWTA ) type [27] that consists of memoryless binary neurons from which only neurons are active . The discrete-time dynamics of the recurrent network at each time step is given by ( 1 ) where is the network state . The nonlinear function sets the units with the highest activities to 1 ( spiking ) , and the rest to 0 ( silent ) . As such , the population firing rate is held constant at , and there is no need to introduce inhibitory neurons to balance excitation and inhibition . Recurrent synaptic efficacy is defined by the weight matrix with being the efficacy of the synapse connecting neuron to neuron . Self-coupling is avoided by setting diagonal elements to 0 . defines neuronal firing thresholds that modulate the neurons' resistance to firing , and hence , their excitability . is the external drive whose dynamics depends on the task performed . More formally , the set of possible network states is a metric space: Definition 1 . Given the set of all binary vectors of size , we define the Hamming metric by the function: According to this metric , the distance between two vectors of is the number of bits at which these two vectors differ . The Hamming metric is a proper metric on strings of fixed length which is the case for . The pair then forms a metric space . It is also equivalent to the norm on the set Y , which allows us to define the Hamming length of a vector as the Hamming distance between and the 0-vector , i . e . . Given the kWTA dynamics ( see Equation 1 ) , the network activity is restricted to the set: ( 2 ) Since , the pair forms a metric space as well . Distances between subsets of can be measured using the Hausdorff metric , which we also denote . We are concerned with the interplay of two forms of plasticity in enhancing the computational capability of the model recurrent network . Spike-timing-dependent synaptic plasticity ( STDP ) is a set of Hebbian/anti-Hebbian learning rules , where synaptic efficacy is modified according to the relative firing time between pre- and post-synaptic neurons [56] . We adapted a simple causal STDP learning rule by which a synapse is potentiated whenever the pre-synaptic neuron fires one time step before the post-synaptic neuron , and is depressed when a post-synaptic spike precedes a pre-synaptic spike by one time step: ( 3 ) where is the synaptic plasticity learning rate set to 0 . 001 . To prevent weights from switching signs or growing uncontrollably , we enforce hard bounds such that the weights remain within the interval [0 , 1] . Competition between synapses due to STDP leads to neurons with synapses that won the competition to fire consistently and those who lost the competition to be constantly silent [57] . To counteract this pathological state , the time-averaged firing rate for a neuron is modulated through homeostatic modification of its excitability threshold using intrinsic plasticity ( IP ) [6] , [7]: ( 4 ) where is the intrinsic plasticity learning rate set to 0 . 001 . This rule uses subtractive normalization to pull the time-averaged firing rate of each neuron closer to the population firing rate . Neural circuits in different brain regions adapt to best serve the region's functional purpose . To that end , we constructed three tasks , each of which resembles in spirit the demands of one such canonical function . We then , under the stimulation conditions of each task , compared the performance , information content , and dynamical response of networks optimized by combining both STDP and IP with networks that are optimized by STDP alone or IP alone . In all tasks , the network is subject to perturbation by a set of inputs . The receptive fields of non-overlapping subsets of neurons are tuned exclusively to each input . As such , each input has its corresponding receptive field in the recurrent neural network . When an input drives the network , all neurons receive a positive drive , while the rest receive none . Readouts are trained on the current network state to compute a function over input sequences , and being time-lags at which target inputs are applied where positive lags corresponds to future inputs and negative lags to past ones . We restrict time-lags to the range . In a first task , RAND x 4 , we assessed the capacity of the network to retain memory of past stimuli within its activity . The recurrent network is driven by four randomly drawn inputs . The receptive field of each input consists of 15 neurons , and one optimal linear classifier is trained for each input/time-lag pair , i . e . fires when and is silent otherwise . The second task , Markov-85 , explores the ability of the recurrent network to discover temporal regularities in its input . The recurrent network receives one of four possible inputs generated from a Markov chain with 85% probability for to be followed by followed by followed by and followed by All other transitions occur with a 5% probability . Again , the receptive field of each input consists of 15 neurons , and one optimal linear classifier is trained for each input/time-lag pair . With the third task , Parity-3 , we exploit the nonlinear expansion provided by the recurrent neural network . Here , the network is subject to binary input , where each symbol has a receptive field of 40 neurons . The task is to identify the parity of a sequence of three successive inputs . This means that given an input sequence , an optimal linear classifier fires when , and is silent otherwise . is the nonlinear exclusive or binary operation . Even though every task used here is very much simplified compared to stimuli usually processed by neural systems , we would still like to link the basic properties of every task presented here to a realistic case processed by a human or an animal . The property of the memory task RAND x 4 that we want to emphasize is that a neural system must be able to process rapidly changing stimuli that are only shortly presented . That property is partly reminiscent of retinal input , which is rather stationary during moments of fixation , and rapidly changing due to saccadic eye movements . However , it needs to be noted that saccadic eye movements might be difficult to predict and may appear rather random , but are very likely structured and stimulus-dependent . This motivated the prediction task Markov-85 that models temporally structured and rapidly changing sensory input that is shortly presented . Such input could either be generated by retinal input and saccadic eye movements , or by the whisking behavior and the produced neural activity in the barrel cortex of a mouse . In addition , nonlinearities are prevailing in natural stimuli , and to highlight the necessity of processing these stimuli , we used the nonlinear task Parity-3 . Such computational demands can be easily motivated by occlusion in vision , where pixel intensities do not sum up linearly at points where one object occludes another in the visual field . Again , we stress that none of these tasks is a good model of real processing in neural systems in nature . However , each is sharing individual aspects that are motivated by real life examples . In order to isolate the role of STDP and IP in shaping the computational and information processing properties of the recurrent network , we compared networks trained by both STDP and IP , with networks that are trained by STDP alone or IP alone . Throughout all experiments , we trained networks of neurons on either the STDP+IP condition , the STDP condition , or the IP condition for a plasticity phase of time steps . For convenience , we call a recurrent network trained with both synaptic and intrinsic plasticity SIP-RN . In contrast , we name a recurrent network that learned with a single plasticity mechanism either SP-RN or IP-RN . is set to 12 , the initial weights are chosen uniformly on the interval [0 , 0 . 1] with 10% connectivity probability , and thresholds are drown from a Gaussian distribution with 0 mean and 0 . 1 standard deviation . Under the IP condition , to assure that weights' distribution is not different from conditions where STDP modifies the synaptic efficacies a pre-plasticity phase of similar length to the plasticity phase precedes the latter , where both STDP and IP are active . Afterwards , the weights structure is destroyed by random shuffling and the plasticity phase starts where IP is turned on . In all experiments where the performance of optimal linear classifiers is estimated , the plasticity phase was time steps long . Afterwards , weights and thresholds are held fixed , the network state is reset to a random initial state , and the training phase starts where linear classifiers are trained using linear regression on time steps , followed by a testing phase of performance for another time steps . At the beginning of the training phase , the network state is reset to a random initial state . If the network dynamics is multistable , this resetting could bring it to a different regime than where the network was at the end of the plasticity phase . To test this possibility systematically , we perform the following post-plasticity perturbation . Given some perturbation parameter . We assume the network state at the end of the plasticity phase is . Instead of randomly choosing the initial network state for the training phase , we choose a network state such that the condition holds . To satisfy this condition , is chosen as follows . In the network state , firing neurons and silent neurons are randomly selected . The firing neurons are then silenced and the silent neurons are set to firing . According to the RC paradigm , an input signal undergoes a nonlinear feature expansion by projecting into a recurrent neural network of nonlinear units . The network recurrency also provides a sustained but damped trace of past inputs ( echo state [29] or fading memory [30] ) to propagate through the network . The network state is then read out by simple linear units through linear regression . Following the plasticity phase , the network activity during the training phase ( 5 ) provides the training data for all optimal linear classifiers , where denotes matrix transpose . The target signal of output neurons for a particular time-lag is clamped in a supervised fashion to ( 6 ) where depends on the task and is the cardinality of the set of possible values which the target signal can take . equals for the tasks RAND x 4 and Markov-85 . Output weights for each time-lag are then computed using linear regression through ordinary least squares ( 7 ) where is the Moore-Penrose pseudoinverse of a matrix , and is the regular inverse of square matrices . These optimal linear classifiers are then validated on the network activity ( 8 ) during the testing phase . First , a pre-estimate of the target signal is computed for each time-lag: ( 9 ) Only one output neuron fires each time step for each time-lag , and this is specified through winner-take-all on the rows of . This leads to the final estimate . The classification performance for each time-lag is finally computed as the percentage of correct classifications: ( 10 ) On multiple occasions , both the self-information capacity of the network state and its dependence on input was measured . Entropy measures self-information capacity which is the expected value of information carried by the network activity and is given by ( 11 ) where is the base-2 logarithm , so that entropy ( and mutual information ) are measured in bits . Mutual information measures the dependence of the network activity on a corresponding input sequence and is given by ( 12 ) In computing entropy and mutual information , we used the algorithm and code developed in [58] that computes entropy from an adaptive k-nearest-neighbor estimate of probability density functions . This allows for reliable estimates of these quantities with far fewer samples in comparison to other algorithms . Nevertheless , due to the high number of channels we have ( 100 neurons ) , and to truncate unnecessary computation time , samples from the network activity are first transfered to the principal components space , and only components that carry 95% of the information are passed to the mutual information estimator . We always considered inputs from the task RAND x 4 and we computed the mutual information between samples of the network state and the three most recent inputs . We encoded each of the four input symbols by a 3-bits code to ensure equal pairwise Hamming distances between symbols . For all cases but one , as few as 5000 samples of the network state and input sequence were enough to reliably estimate entropy and mutual information . The exception was computing mutual information between input and IP-RN activity , which demanded a higher number of samples ( 500000 time steps ) and very long computation time , as covering 95% of the information required no less than 60 principal components . For a full understanding of the emerging information processing properties of the interaction of synaptic and intrinsic plasticity , it was necessary to rely on and develop concepts from the newly emerging mathematical theory of nonautonomous dynamical systems [31] , [32] . Throughout what follows , the correspondence of the introduced concepts to our model is clarified . First , autonomous dynamics are defined , since they form a special instance of the nonautonomous case . Definition 2 . Let be a metric space with a metric A discrete-time semi-dynamical system is a function that satisfies Equation 1 defines the driven or nonautonomous kWTA dynamics . The autonomous alternative is given by the discrete-time difference equation ( 13 ) where is the kWTA nonlinearity defined as above . To relate Equation 13 to Definition 2 , the function ( the solution mapping ) is chosen such that ( 14 ) where is function composition . For to be an autonomous semi-dynamical system , it has to satisfy the three conditions of Definition 2 . The first two conditions are trivial , as they result from the definition of function composition . We turn to prove the third condition , namely , the continuity of . We first observe that is merely the t-fold composition of the function and since the composition of continuous functions is continuous , it is sufficient to prove the continuity of Proposition 3 . The kWTA function from Equation 13 defined on the metric space is continuous , i . e . Proof . For all and all , we choose . For all , if the Hamming distance , and have to be equal , since the kWTA dynamics restricts the distances between any two states to the set . As such , since is a metric , , which is always smaller than . Ergo , is continuous . We note that the proof to Proposition 3 becomes trivial if we consider a result from topology which states that any function from a discrete topological space to another is continuous . However , the proof is interesting in that it shows that has a stronger form of continuity , that is , is uniformly continuous , since the proof shows that there exists a packing radius such that either . With the proof of Proposition 3 , we conclude that the kWTA autonomous dynamics in Equation 13 generates a discrete-time semi-dynamical system . A dynamical system is a semi-dynamical system with invertible dynamics , which is not the case for kWTA networks . However , for all intents and purposes , being a semi-dynamical system is sufficient for formalizing the nonautonomous dynamics of the model network . Characterizing the computational properties of the model neural network requires defining invariant sets and attractors . Definition 4 . Let be a discrete-time semi-dynamical system generated by an autonomous difference equation on a metric space . A subset is invariant under , and is positively invariant if . is an attractor of if the following conditions hold: For the kWTA dynamics , the second condition is assured , since is discrete and finite , which makes all subsets compact . The third condition assures that no subset of satisfies the invariance and compactness conditions . Another important concept is that of a basin of attraction which associates each attractor with the region of the state space that converges to that attractor: Definition 5 . Let be a discrete-time semi-dynamical system generated by an autonomous difference equation on a metric space . The basin of attraction of an attractor of is defined by Unlike autonomous ( semi- ) dynamical systems , the elapsed time is not sufficient to find the solution for nonautonomous dynamics: both the start and end times must be specified . Accordingly , we now define a discrete-time nonautonomous dynamical system as a process . In what follows , we will make use of the set . Definition 6 . Let be a metric space with a metric A discrete-time process is a function that satisfies We now turn to formulating the driven kWTA difference equation ( see Equation 1 ) as a discrete-time process . We first note that for a particular task , a set of possible inputs is defined . For completeness , this set covers the autonomous case by including the 0-vector . For each member of this set , we define a separate map . The set of maps with cardinality defines a family of discrete-time autonomous semi-dynamical systems . These maps are chosen either randomly for the tasks and Parity-3 , or in a more structured fashion for the task Markov-85 . In either case , the kWTA discrete-time nonautonomous dynamics in Equation 1 can be rewritten in the form ( 15 ) which generates a solution mapping ( 16 ) The solution mapping satisfies the three properties of a process . The first two properties are a product of the definition of function composition , and the continuity condition is proven exactly as in Proposition 3 . Given the above , the family of discrete-time autonomous difference equations on the metric space generates a process , and thus , it defines a particular kind of nonautonomous dynamical systems termed an input-driven dynamical system . It is important to point out that an input-driven dynamical system is not defined for a particular input sequence , but for all input sequences drawn from its input set . This becomes more explicit if one considers the alternative skew product definition of a nonautonomous dynamical system , where the input is treated as a driving autonomous dynamical system [31] , [32] . We compare the two definitions of nonautonomous dynamical systems in Text S3 . We now cover a few important concepts that will aid in defining the dynamic behavior of the model neural network . Attractors in nonautonomous dynamical systems are defined on nonautonomous sets , relating strongly to the concepts of invariance and entire solutions . Definition 7 . Let be a discrete-time input-driven dynamical system generated by the family of autonomous difference equations on a metric space . A subset is called a nonautonomous set , and for all , the setis called the t-fiber of . is said to be invariant under if for all . An entire solution of is an invariant set under whose t-fibers are the singleton sets that are the images of the function such that An important property of invariant nonautonomous sets is that they consist exclusively of entire solutions ( for a proof , see Lemma 2 . 15 in [31] ) . Nonautonomous attractors are nonautonomous sets . As such , they consist of entire solutions as well . There are several types of attractors of nonautonomous dynamical systems . Only of interest to our model neural network are forward attractors , so we drop the qualifier ‘forward’ and substitute it with ‘nonautonomous’ . Definition 8 . Let be a discrete-time input-driven dynamical system generated by the family of autonomous difference equations on a metric space . A nonautonomous set is a nonautonomous attractor of if the following conditions hold: As in the autonomous dynamics of kWTA networks , all subsets of are compact . The third condition assures that no subset of satisfies the invariance and compactness conditions . One may generalize the concept of a basin of attraction in an autonomous dynamical system to the nonautonomous case . This concept associates each nonautonomous attractor with the region of the state space that converges to that attractor: Definition 9 . Let be a discrete-time input-driven dynamical system generated by the family of autonomous difference equations on a metric space . The nonautonomous basin of attraction of a nonautonomous attractor of is defined by Spatiotemporal computations requires encoding different input sequences in the states of the neural network . The set of network states accessible from some initial conditions within a basin of attraction through perturbing the network with a particular input sequence defines this sequence's volume of representation . Definition 10 . Let be a discrete-time input-driven dynamical system generated by the family of autonomous difference equations on a metric space . Given an input sequence and a basin of attraction , a subsetis called the volume of representation of the input sequence within the basin . The sequence length defines the order of this volume . The nonautonomous set whose t-fibers are order-1 volumes of representation is called the perturbation set within . Also , given a function on input sequences such that , the setis the volume of representation of given g . It is straightforward to show that , within a basin of attraction , the volume of representation of some sequence is a superset of the volume of a sequence for all , and that the volume of is equivalent to the union of the volumes of for all . We term this property the volumes' inclusion property . The concept of ‘volumes of representation’ allows us to state the following theorem on the nature of attractors in discrete-time input-driven dynamical systems: Theorem 11 . Let be a discrete-time input-driven dynamical system generated by the family of autonomous difference equations on a metric space , and let be a compact nonautonomous basin of attraction . The following holds: Proof . This theorem allows us to characterize the properties and relations between autonomous and nonautonomous attractors of kWTA networks , where all subsets of are compact due to 's finiteness and discreteness . Namely , it allows us , within some compact basin , to allocate the nonautonomous attractor's t-fibers as subsets of the t-fibers of the perturbation set , and it shows that the autonomous attractor of the input at time t is the t-fiber of the nonautonomous attractor . It is possible for a process to behave locally or globally as an autonomous ( semi- ) dynamical system . That is equivalent , in the case of input-driven dynamical systems , to being input-insensitive . Definition 12 . Let be a discrete-time input-driven dynamical system generated by the family of autonomous difference equations on a metric space . A state is said to be input-insensitive if for all . An input-insensitive basin is a basin of attraction that consists entirely of input-insensitive states . This definition implies that the volumes of representation of a particular order and the t-fibers of each nonautonomous set within this basin are equivalent , including the perturbation set and the nonautonomous attractor: they reduce to autonomous sets . The input-insensitive attractor becomes the autonomous attractor of each discrete-time semi-dynamical system generated by a difference equation .
|
The world is not perceived as a chain of segmented sensory still lifes . Instead , it appears that the brain is capable of integrating the temporal dependencies of the incoming sensory stream with the spatial aspects of that input . It then transfers the resulting whole in a useful manner , in order to reach a coherent and causally sound image of our physical surroundings , and to act within it . These spatiotemporal computations are made possible through a cluster of local and coexisting adaptation mechanisms known collectively as neuronal plasticity . While this role is widely known and supported by experimental evidence , no unifying theory of how the brain , through the interaction of plasticity mechanisms , gets to represent spatiotemporal computations in its spatiotemporal activity . In this paper , we aim at such a theory . We develop a rigorous mathematical formalism of spatiotemporal representations within the input-driven dynamics of cortical networks . We demonstrate that the interaction of two of the most common plasticity mechanisms , intrinsic and synaptic plasticity , leads to representations that allow for spatiotemporal computations . We also show that these representations are structured to tolerate noise and to even benefit from it .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"circuit",
"models",
"mathematics",
"neural",
"homeostasis",
"neural",
"networks",
"computational",
"neuroscience",
"biology",
"nonlinear",
"dynamics",
"neuroscience",
"learning",
"and",
"memory"
] |
2014
|
Spatiotemporal Computations of an Excitable and Plastic Brain: Neuronal Plasticity Leads to Noise-Robust and Noise-Constructive Computations
|
Two lineages of Salmonella enterica serovar Typhimurium ( S . Typhimurium ) of multi-locus sequence type ST313 have been linked with the emergence of invasive Salmonella disease across sub-Saharan Africa . The expansion of these lineages has a temporal association with the HIV pandemic and antibiotic usage . We analysed the whole genome sequence of 129 ST313 isolates representative of the two lineages and found evidence of lineage-specific genome degradation , with some similarities to that observed in S . Typhi . Individual ST313 S . Typhimurium isolates exhibit a distinct metabolic signature and modified enteropathogenesis in both a murine and cattle model of colitis , compared to S . Typhimurium outside of the ST313 lineages . These data define phenotypes that distinguish ST313 isolates from other S . Typhimurium and may represent adaptation to a distinct pathogenesis and lifestyle linked to an-immuno-compromised human population .
Salmonella enterica isolates can infect a range of animals and humans , causing a spectrum of disease syndromes ranging from gastroenteritis through to typhoid and an asymptomatic carrier state [1] . From a clinical perspective S . enterica serovars have been classically assigned to two broad groups , typhoidal or non-typhoidal Salmonella ( NTS ) . Typhoidal Salmonella include the human restricted S . enterica serovar Typhi ( S . Typhi ) , the cause of the systemic disease typhoid fever , which is strictly transmitted within the human population independently of a zoonotic reservoir . NTS , on the other hand , are predominantly associated with self-limiting gastroenteritis , largely originating from zoonotic reservoirs with human-to-human transmission regarded as being relatively rare [2] . Invasive NTS ( iNTS ) disease in sub-Saharan Africa does not fit well into the classical view of salmonellosis . NTS has emerged as a significant cause of invasive human disease , exceeding S . Typhi in many parts of the region as the leading cause of invasive salmonellosis . Humans can be predisposed to this disease by immune suppression or co-infections , which include severe malaria in children and HIV in adults [3 , 4] . Invasive NTS clinical syndrome is somewhat dissimilar to both typhoid fever and gastroenteritis , and includes non-specific fever and only sporadic or limited diarrhea [5] . High case fatalities have been reported in children and adults in the absence of adequate treatment [6–9] . We recently reported that the emergence of iNTS disease within the sub-Saharan region has been associated with the emergence of two closely-related , multi-antibiotic resistant lineages of S . Typhimurium that belong to multilocus sequence type ( MLST ) ST313 [10] . Phylogenetic analysis indicated that these ST313 lineages emerged independently in recent decades , in close temporal association with the HIV pandemic [10] . As no obvious zoonotic source of ST313 S . Typhimurium has been identified , it has been postulated that these lineages may be undergoing host adaptation to humans and may be transmitted , at least in part , directly from human-to-human [11] . Additionally , emergence of the lineages was concomitant with acquisition of multidrug resistance ( MDR ) including chloramphenicol in one lineage . Genome sequencing of a representative ST313 isolate , D23580 from Malawi , identified distinct genetic signatures not present in other sequenced non-ST313 S . Typhimurium [5] . For example , the genome of D23580 exhibited considerable genome degradation with some similarity to that observed in S . Typhi [5] . Genome degradation , in the form of the accumulation of so-called pseudogenes , is a signature of some host restricted pathogens including Bordetella pertussis [12 , 13] , S . Typhi [14 , 15] , S . Paratyphi [15 , 16] and S . Gallinarum [16] . Here , a population-based approach was used to assess how genome degradation emerged within the ST313 lineages . In addition , we used a range of approaches to phenotype representatives of the ST313 in an effort to link the genotypic differences to metabolic and virulence-associated phenotypic differences .
ST313 isolates fall into two closely related phylogenetic lineages that are distinct from other S . Typhimurium ( Fig . 1A , S1 Fig , S1 Table ) [10] . Previous genome sequence analysis of one ST313 isolate , D23580 from lineage II , revealed both gene acquisition ( e . g . novel phage elements ) and genome degradation ( e . g . deletions and pseudogenes ) in comparison to S . Typhimurium in other lineages [5] . To ascertain genome variation within the overall ST313 population , we analysed whole genome sequences of 129 ST313 isolates using the ST19 S . Typhimurium SL1344 genome as a reference . We first identified the synonymous and non-synonymous single nucleotide polymorphisms ( SNPs ) in the S . Typhimurium ST313 lineages compared to the reference ( Fig . 1A , Table 1 ) . The dN/dS ratios for the parental branch of ST313 , and that of each lineage since divergence from the last common ancestor , were similar ( 0 . 41 ± 0 . 13 s . d and 0 . 35 ± 0 . 007 s . d , respectively , Table 1 ) . Thus , the dN/dS values were smaller than one but still elevated and similar to those expected for recently evolved lineages where time has been too short for purifying selection to act to a significant level [17 , 18] . The relatively high proportion of non-synonymous SNPs in the two lineages may also represent segregating polymorphisms rather than fixed mutations . Many of these SNPs , both synonymous and non-synonymous , were in metabolic genes and genes involved in degradation of small molecules in both lineages , when compared with SL1344 ( Table 1 , Fig . 2A ) . A proportion of the SNPs were also found in genes with no assigned function . To further characterise the acquisition or loss of genetic material by the ST313 lineages we analysed the whole genome sequence of the additional sequenced isolates . There was little variation in the arrangement of genes within the major virulence-associated Salmonella pathogenicity islands ( SPIs ) , including SPIs -1 to -6 , -9 , -11 to -14 , and -16 ( Fig . 1B ) . The previously described prophage elements BTP1 , BTP3 , BTP4 and BTP6 [5] were present in all isolates and putative deletion events have led to the loss of the phage remnants SLP281 and Fels2 compared to S . Typhimurium SL1344 and other non-ST313 isolates included in the analyses . Sequences with similarity to the phage SLP289 were found in a subset of lineage I isolates from Uganda and Kenya but were absent from the rest of the ST313 population . Whole or partial sequences of the S . Typhimurium DT104-associated prophage 5 ( Fig . 1B ) were present in both ST313 lineages , although there was no obvious pattern to the distribution of particular rearrangements of this phage within the ST313 tree . Thus , these data catalogue the major insertions and potential deletions that have occurred since the divergence of ST313 lineages from the last common ancestor . The ST313 sequences were next analysed for evidence of pseudogene formation arising from nonsense SNPs and frame-shift mutation caused by insertions or deletions ( <20bp ) impacting on all ST313 S . Typhimurium used in our analyses . Ten pseudogenes were present in all ST313 isolates but intact in SL1344 , including the genes ttdA , ratB and SL1567 ( Fig . 1A , S3 Table ) . ttdA encodes L ( + ) -tartrate dehydratase , involved in glyoxylate and dicarboxylate metabolism [19] . The gene ttdA is also a pseudogene in S . Typhi and S . Paratyphi A . The ratB gene which encodes an outer membrane protein implicated in intestinal persistence in a murine model , is also a pseudogene in S . Typhi , S . Paratyphi A , S . Paratyphi B and the fowl-restricted S . Gallinarum [5 , 20 , 21] . Other pseudogenes found in all sequenced ST313 include SL1567 , a membrane associated protein with different independently acquired nonsense SNPs in lineage I and II , SL2747A , a putative exported protein which may be involved in phospholipid biosynthesis and a transposase , SL1873 ( S3 Table ) . Five pseudogenes were a result of frame-shift mutations ( Fig . 1A , S3 Table ) . These affected the genes SL2990 , SL3733 , SL1475 , SL13051 , SL2653 , which are predicted to be involved in transcriptional regulation , metabolism and transport and annotated as possessing conserved hypothetical functions , respectively . Most of these genomic signatures represent degradation that occurred before divergence from the last common ancestor of both ST313 lineages . In addition to shared genome degradation , lineage-specific nonsense SNPs , frame-shift mutations and small deletions are also present in genes of isolates from ST313 lineages I or II . These represent degradation that occurred after divergence of the two ST313 lineages . Three additional candidate pseudogenes were found in all lineage I isolates . These were in prfH , a peptide chain release factor , phnT , a probable ATP-binding component of 2-aminoethylphosphonate transporter and ybjZ a putative ABC transporter ( Fig . 1A , S3 Table ) ; Lineage I isolates also harbour a 700 bp partial deletion within a putative phage gene SL1477 . In lineage II , three candidate pseudogenes were a consequence of nonsense SNPs while there was an insertion within SL2214 , a putative phage protein in an O-antigen modification locus . The lineage II candidate pseudogenes arising from nonsense SNPs include the gene encoding a conserved hypothetical protein , SL2659 and the membrane proteins yhjU and SL4223 . The sseI gene , which encodes a type III effector , is inactivated by an IS200 element in all lineage II but not lineage I isolates in our collection . Genes associated with allantoin metabolism or transport e . g . allB , gcl , glp and ybbW are likely pseudogenes in lineage II . Interestingly genes associated with allantoin metabolism are also inactivated in S . Typhi , S . Paratyphi A and S . Gallinarum . All lineage II isolates possess a partial deletion of the pipD gene , encoding a SPI-5 associated protein implicated in persistence in murine macrophages and fluid secretion in bovine models [22 , 23] ( S2 Table ) . The phoP/phoQ regulated genes , pagO and pagM harbour deletions in all lineage II isolates ( Fig . 1A , S2 Table ) . It is important to note that a number of phoP/phoQ-regulated genes are associated with virulence [24] and pagO has been previously linked to virulence in porcine models [25] . Additionally , a 4 . 2kb region encoding plasmid stability proteins was also deleted in all lineage II isolates ( S2 Table ) . There is a statistically significant over-representation of surface/membrane associated and exported proteins inactivated in both lineages ( p value <0 . 05 ) ( Fig . 2B ) . A systematic analysis of 576 metabolic activities was performed using Biolog phenotype microarrays ( PM ) [26] on three representatives each of the two ST313 lineages and four S . Typhimurium ST19 isolates including SL1344 that acted as experimental controls and comparators ( S1 Table ) . A principal component analysis ( PCA ) ( Fig . 3A ) and a hierarchical clustering of Biolog signal values ( Fig . 3B ) were employed to assess the data sets ( S1 Text , S2 Fig ) . The results from both analyses support the conclusion that ST313 isolates share similar metabolic capacity distinct from ST19 S . Typhimurium isolates . For example , analyses of cellular respiration over 48 hours of incubation showed that ST313 isolates exploit particular carbon sources such as meso-tartaric acid ( meso-tartrate ) and tricarballylic acid more readily than S . Typhimurium ST19 isolates included in the experiment ( Table 2 ) . Conversely , the ST19 isolates , which includes SL1344 , utilised carbon sources such as L-tartaric acid and dihydroxyacetone ( Table 2 ) . The differing metabolism of L-tartaric acid and meso-tartaric acid by ST313 and ST19 corroborates the observation that ttdA , encoding the stereo-specific enzyme tartrate dehydratase , is a pseudogene in ST313 isolates [27 , 28] [29] . To determine if ST313 isolates are virulent in a mouse systemic infection model we orally inoculated genetically susceptible mice ( NRAMP1- , C57bl/6 ) with representative ST313 isolates from lineage I and II . The resulting data showed that the tested ST313 isolates are indeed able to colonise systemic sites in this model ( S3 Fig ) . We therefore investigated the ability of representative isolates of ST313 ( A130 from lineage I and D23580 from lineage II ) to induce an inflammatory response in the caecum of orally inoculated streptomycin pre-treated C57bl/6 mice , compared with SL1344 ( Fig . 4 ) . No significant difference in Salmonella colonisation of the caecum was evident at 48 hours post-inoculation ( S4 Fig ) . SL1344 induced pronounced inflammation characterised by marked oedema in the submucosa with moderate to marked cellular inflammatory infiltrates in the submucosa and mucosa , with numerous crypt abscesses and erosive changes in the surface epithelium ( Fig . 4C & 4D ) . However , these pathological signatures were less common in mice infected with A130 ( Fig . 4E ) or D23580 ( Fig . 4F ) , although there was some evidence of mild to moderate submucosal oedema and mild inflammatory cell infiltration into the submucosa and mucosa . The epithelial surface changes and crypt abscesses were also much less prominent . SL1344ΔorgA , a SPI-1 defective derivative induced similar levels of inflammatory cell infiltration into the mucosa and submucosa of the caeca to A130 and D23580 ( Fig . 4B ) . Uninfected caecum exhibited no noticeable oedema or neutrophil infiltration ( Fig . 4A ) . The histopathological scores of the replicate experiments summarised in Fig . 4G & 4H illustrate these observations . In further experiments , groups of streptomycin pretreated 129P2/olaHsd mice were independently inoculated with the same S . Typhimurium isolates as in the previous experiment in C57bl/6 mice . Similar differences in intestinal pathology in the caecum were observed 48 hours post-inoculation ( Fig . 4H ) . The overall virulence profiles observed in the streptomycin-treated mouse model of colitis were also evident in a bovine ligated ileal loop model . In these studies , ligated segments of the mid-ileum of two calves were infected in triplicate with representative invasive ST313 isolates from the two lineages ( lineage I—A130 & 5597; lineage II- D23580 & 5579 ) and compared to bovine virulent ST19 S . Typhimurium strains ST4/74 , DT104 and IR715 ( S1 Table ) and internal negative controls . Secretory and inflammatory responses in this model are strongly influenced by SPI-1 ( prgH mutation; Fig . 5 ) , as previously described [30] . In pair-wise t-tests , a significant difference in fluid accumulation was detected 12 hours post-inoculation between ST19 and ST313 isolates in almost all cases ( Fig . 5 ) . Mean values for the secretory response to the three ST19 isolates also differed significantly from the mean value for the four ST313 isolates ( p = 0 . 02 ) . Recruitment of 111Indium oxinate-labelled polymorphonuclear leukocytes ( PMN ) relative to the negative control ( PMN influx ) also differed significantly for a number of ST19 and ST313 isolates in pair-wise combinations ( Fig . 5 ) . Though the difference in mean values for PMN influx for all ST19 vs . ST313 was marginally not significant ( p = 0 . 065 ) , the difference was significant for PMN recruitment to the luminal contents by ST19 vs . ST313 ( p = 0 . 04 ) .
Here , we have identified lineage-specific signatures and phenotypic changes that differentiate ST313 from other S . Typhimurium , including isolates associated with gastroenteritis . These findings extend on previous analyses of D23580 , a lineage II ST313 isolate to the broader ST313 population [5] . We identified and highlighted lineage specific gene acquisition and loss events , some common to both ST313 lineages and others restricted to either lineage I or II . Among the collective changes that have accompanied the emergence of ST313 are a relatively high proportion of genomic changes found in metabolic genes ( Fig . 2A ) . This is worthy of note since altered metabolic capacity has previously been associated with adaptation of Salmonella serotypes to extra-intestinal niches [31] . The high numbers of SNPs within this class of gene could also be indicative of evolutionary pressure acting on the ST313 isolates . Examples of genes within this group include ttdA , that are also found either deleted or are pseudogenes in host-restricted or host-adapted serovars such as S . Typhi , S . Paratyphi A , S . Paratyphi B and S . Gallinarum . This ability to utilise less common carbon sources such as meso-tartrate and tricarballylic acid by the ST313 isolates may positively influence their fitness in a new ecological niche . Degradation of aerobic metabolic genes in the isolates of the two lineages may suggest a preferential loss or reduction of aerobic metabolic capacity in ST313 . This observation could be indicative of a heightened ability for anaerobic metabolism following internalisation within macrophages , as anaerobic respiration and metabolism takes precedence over aerobic metabolism within this niche . These metabolic activities impact the interaction of the pathogen with the host in the intracellular niche and have implications for intracellular compartmentalisation within tissues such as the bone marrow , as previously reported [32 , 33] . Further evidence for the clonality of the two epidemic lineages in sub-Saharan Africa is emphasized by the predominantly conserved pattern of the known S . Typhimurium phages and genomic islands . Although we present details of common SNPs in ncRNAs found in intergenic regions of the genomes in ST313 ( S1 Text ) , the impact of these SNPs on regulation and subsequently on metabolism or virulence-associated phenotypes are difficult to predict and will be the subject of future investigation . Surface proteins are often antigenic in nature and are part of the first line of contact with the host immune system . Changes in surface proteins can thus impact on the host response to colonisation and invasion by these pathovariants . The so-called ‘stealth’ methods employed by host-adapted serovars to evade host gut inflammatory responses leading to increased invasive capability have been well documented [34] . Thus , potential inactivation of genes such as pagO [23 , 25] , pipD [22] , ratB [20] and sseI [5] in ST313 isolates is interesting in this regard ( Table 3 ) . These proteins are all exported or membrane surface-associated proteins implicated in the establishment of gastrointestinal infection or long-term systemic infections in animal models . Such defects have been observed in the differential virulence profiles observed in pigs infected with S . Typhimurium , which cause non-fatal but acute enteritis , and S . Choleraesuis , which is host-adapted and frequently causes a severe systemic disease in pigs . In a porcine ligated ileal loop model , S . Typhmiurium elicits a profound inflammatory response , which subsequently controls and confines the pathogen to the intestinal mucosa . Conversely , the host-adapted S . Choleraesuis replicated slowly and elicited weaker pro-inflammatory responses both of which may facilitate avoidance of the host immune response by stealth [35] . Although we do not show a direct causal link , it is also possible that the inactivation of these virulence-associated proteins has led to the reduction in the enteropathogenic potential of isolates in the lineages I and II ( Fig . 4 & 5 , respectively ) . S . Typhimurium ST313 is frequently associated with iNTS disease in sub-Saharan Africa . However , the extent to which this genotype is also associated with gastroenteritis in this region is poorly understood . iNTS disease syndrome is distinct from typhoid fever and gastroenteritis and thus lack an established animal model of infection . Mice with a defective Nramp1 gene are also susceptible to invasive NTS disease , so we used this infection model to determine if ST313 isolates differed in their ability to colonise systemic sites organs of the reticuloendothelial system and the gall bladder compared with the non-ST313 SL1344 . All ST313 exhibited some virulence in the mice and colonised to a similar level to that observed for SL1344 ( S3 Fig ) . This is consistent with reports that ST313 isolates can establish systemic infections in different models of infection [36] . We also evaluated the ability of ST313 isolates ( A130 , D23580 ) and ST19 ( SL1344 ) to invade eukaryotic cells growing in vitro was using Hep2 cells . Although , all three isolates showed evidence of invasion , internalisation and replication in epithelial cells over a time course of 24 hours , A130 and D23580 consistently showed lower invasion compared to SL1344 in this in-vitro model ( p < 0 . 001 at 24 hours ) ( S5 Fig , SI Text ) . We have detailed the shared genomic and phenotypic variation that may contribute to the adaptation of these new pathovariants to the novel niche provided by immunocompromised humans , identifying several changes that are consistent with those found in host-adapted lineages of S . enterica . The high proportion of metabolic genes implicated in the degraded gene component in lineages I and II of ST313 is a signature that is an emerging narrative among invasive pathogens in enterobacteriaceae including Salmonella [16] , Shigella [37] , Yersinia [38] and E . coli [39] . Our results thus suggest adaptation within a particular human population in ST313 . However , the possibility of asymptomatic carriers or environmental reservoirs being integral components of iNTS transmission networks also exists . Elucidating these networks and defining the relationship between zoonotic , environmental and human isolates remains the subject of much needed on-going research .
Bacterial isolates used in this study have been described in Okoro , et al . , 2012[40] . See S1 Table . All bacteria were grown on Luria-Bertani ( LB ) medium; single colonies were incubated in LB Broth overnight at 37°C . Descriptions of specific growth conditions for experiments are given in the corresponding segments below . dN/dS was calculated using the formula adapted from Holt et al . , 2008 ( N/n ) / ( S/s ) , where N = sum of nonsynonymous SNPs , n = nonsynonymous sites in non-repetitive protein-coding sequences , S = sum of synonymous SNPs , s = synonymous sites in non-repetitive protein-coding sequences[14 , 41] To investigate the origin of SNPs reported on the tree , SNPs were reconstructed back to the phylogenetic tree using parsimony and optimised by both ACCTRAN ( accelerated transformation ) and DELTRAN ( delayed transformation ) [42] . Both methods gave comparable results and so the results from the DELTRAN optimisation are presented here . The DELTRAN method allocates or maps SNP origins along the phylogenetic branches as close to the tips as possible [43] . This enabled frameshift mutations and premature stop codons that reduced the length of CDSs relative to their annotation in the reference genome to be detected . SNP positions , type and quality were manually confirmed by checking reads against the reference sequence and visualised using BamView[44] . Paired-end sequence reads of each isolates were mapped to the multi-fasta sequence features of either insertion sequences , phages or pathogenicity islands using the Burrows-Wheeler Aligner software BWA[45] , with minimum base call quality of 50 , minimum mapping quality of 30 , and minimum read depth of 4 . Isolates from each of identified lineages were analyzed separately , by lineage . A cut-off value of < 30% of reads mapped to the length of the feature was selected as an indication of absence and > 70% as presence of the region of interest in an isolate . A heat map of the analysis based on the selected cut-off values was generated . Culture and inoculum preparation were preformed according to a modified manufactures’ protocol ( see S1 Text ) . A total of 576 assays were performed for each isolate , with each isolate represented by three biological replicates . Bacteria were incubated for 15–48 hours at 37°C and bacterial respiration on each assayed metabolite was measured by colorimetric redox assay . The metabolic activity and kinetics data files of each strain over time were exported from the OmniLog phenotype MicroArray ( PM ) program suite . Further analysis proceeded as described previously in R[46] . Signal values were calculated as in Homann et al . , 2005 [47] . Log signal values displayed a clear bimodal distribution corresponding to non-respiring ( background dye reduction ) and respiring modes . Normal distributions were fitted to each mode , and strains were defined as respiring on a particular substrate if all 3 replicates were at least 4 times more likely to originate from the respiring distribution . Significant differences in respiration rates between isolates were assessed using a moderated t-test with the LIMMA R package [48] . P-values were corrected using the Benjamini and Hochberg method [49] to control for the false discovery rate . Results presented here are for respiration for up to 15 hours and 48 hours . Results with adjusted p-values of <0 . 05 and signal value differences ( positive or negative ) greater than or equal to 100 at 48 hours were selected as significant . Results similar to the 48 hour profiles with adjusted p-values of <0 . 05 were also reported for the 15 hour profiles . The functions of metabolites significantly utilized to a greater or lesser degree by the invasive isolates commonly relative to SL1344 were identified for each cluster , and the list of the associated metabolites generated and analysed with Pathway Tools [50] to put them in a wider context and predict the metabolic pathways that were involved . All mouse experiments were conducted in compliance with the Animals ( Scientific Procedures ) Act 1986 under Home Office project licence 80/2596 with the consent and approval of the Ethical Review Committee of the Wellcome Trust Sanger Institute , UK . Mice were sacrificed by cervical dislocation at the end of the experiment . Calf ligated ileal loop experiments were conducted in compliance with the Animals ( Scientific Procedures ) Act 1986 under Home Office project licence 30/2485 with the consent and approval of the Ethical Review Committee of the Institute for Animal Health , UK . General anaesthesia was induced by intravenous administration of propofol and maintained by inhalation of isoflourane in oxygen for the duration of the study . Calves were given an overdose of intravenous sodium pentobarbitone at the end of the study . S . Typhimurium isolates were grown in LB agar supplemented with appropriate antibiotic selection and incubated overnight at 37°C . Single colonies were used to inoculate LB broth and incubated overnight at 37°C . Approximately 1x107 were inoculated into each mouse . Two experiments were conducted and a total of 5 and 3 mice per isolate were used for the infections in first and second experiment , respectively . Specific pathogen-free SPF female mice C57BL/6 ( groups of five ) or female 129P2/olaHsd mice , ( groups of three ) , 6–8 weeks old , were treated by oral gavage with 0 . 2 mL of 100 mg/mL streptomycin by oral gavage . At 20 hours after streptomycin treatment , mice were infected with 1x106 ( C57BL/6 mice ) or 1x107 ( 129P2/olaHsd mice ) of S . Typhimurium in 0 . 2 ml of PBS pH 7 . 4 or treated with sterile PBS ( control ) by oral gavage . At 48 hours ( C57BL/6 ) and 72 hours ( 129P2/olaHsd ) , mice were culled and two caecal tissue samples taken for enumeration of viable S . Typhimurium or were fixed with formalin for subsequent wax embedding , sectioning and tissue staining with haematoxylin/eosin ( H/E ) staining . Enumeration of bacteria was conducted by plating serial dilutions of caecal tissue homogenates on LB agar containing the appropriate antibiotics . Colonies were counted after overnight incubation at 37°C . The H/E stained caeca were histopathologically assessed and scored using a 4-point scale of 0 , 1 , 2 , or 3 , for five markers of vascular and cellular inflammation by using a modification of methods described in Kim , J . J . et al . , 2012 [51] as follows; mucosal inflammatory cellular infiltration predominantly by neutrophils ( PMNs ) , the presence of crypt abscesses ( neutrophils within the lumen of the crypts in the mucosa ) , erosive and reactive changes to the epithelial surface of the mucosa , the amount of submucosal oedema assessed by the increase in thickness of the submucosa and the level of submucosal inflammatory cellular infiltration predominantly by neutrophils . Salmonella-induced secretory and inflammatory responses in calves were quantified essentially as described previously [30] . Briefly , two 4-week-old Friesian bull calves were placed under terminal general anaesthesia , a laparotomy performed and the mid-ileum flushed with sterile PBS . In each calf , twenty-seven 6 cm loops with 1 cm spacers were constructed by ligation of the gut with surgical silk . Representative invasive ST313 isolates from the two lineages ( lineage I—A130 & 5597; lineage II- D23580 &5579 ) and bovine virulent ST19 strains SL1344 , DT104 and IR715 . IR715 is a nalidixic acid-resistant derivative of strain ATCC 14028 [52] . Triplicate loops in each calf were inoculated in a semi-randomized order with c . 1x109 CFU of the indicated S . Typhimurium strains grown to mid-logarithmic phase in LB broth at 37°C . Three loops in each calf were inoculated with an equivalent volume of sterile LB broth as a negative control . After inoculation the mid-ileum was returned to the abdominal cavity for 12 h then the animals given an overdose of pentobarbitone sodium . At post mortem examination , loops were excised and the volume of fluid accumulated recorded and normalized to loop length [volume ( mL ) /length ( cm ) ] . To quantify inflammation , c . 80 mL of jugular blood was collected at the start of the experiment and PMN isolated and labelled with 111Indium oxinate as described[30] then injected into the donor calf within 1 h of loop inoculation . Gamma-radioactivity associated with the mucosa and contents of each loop was normalized to loop length ( counts per minute/cm ) , then the mean PMN influx for each set of triplicate loops determined by dividing the mean value for test strains by the value for the negative control . Values shown are the mean ± standard error of the mean ( SEM ) from two independent animals .
|
Salmonella enterica is a diverse species , isolates of which can colonise or infect many different animals , including humans and can cause different disease syndromes . S . enterica can be sub-typed using serology into serovars . Isolates from some serovars , known as generalists , can infect multiple hosts ( e . g . S . Typhimurium ) and usually cause gastroenteritis . However , other serovars exhibit host adaptation or even restriction . Host-adapted serovars such as S . Dublin show preference for a particular host but can also infect other hosts , while host-restricted serovars are capable of infecting only a single host ( e . g . S . Typhi in humans ) and frequently cause febrile systemic disease ( typhoid ) . In this study , we use genotypic and phenotypic methods to investigate clinical isolates representative of populations of two recently emerged S . Typhimurium lineages of type ST313 associated with invasive disease in sub-Saharan Africa . Our results identify potential characteristics in these isolates that may be associated with adaptation to invasive disease in humans with a compromised immunity .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[] |
2015
|
Signatures of Adaptation in Human Invasive Salmonella Typhimurium ST313 Populations from Sub-Saharan Africa
|
Cellular and systemic responses to low oxygen levels are principally mediated by Hypoxia Inducible Factors ( HIFs ) , a family of evolutionary conserved heterodimeric transcription factors , whose alpha- and beta-subunits belong to the bHLH-PAS family . In normoxia , HIFα is hydroxylated by specific prolyl-4-hydroxylases , targeting it for proteasomal degradation , while in hypoxia the activity of these hydroxylases decreases due to low oxygen availability , leading to HIFα accumulation and expression of HIF target genes . To identify microRNAs required for maximal HIF activity , we conducted an overexpression screen in Drosophila melanogaster , evaluating the induction of a HIF transcriptional reporter . miR-190 overexpression enhanced HIF-dependent biological responses , including terminal sprouting of the tracheal system , while in miR-190 loss of function embryos the hypoxic response was impaired . In hypoxic conditions , miR-190 expression was upregulated and required for induction of HIF target genes by directly inhibiting the HIF prolyl-4-hydroxylase Fatiga . Thus , miR-190 is a novel regulator of the hypoxia response that represses the oxygen sensor Fatiga , leading to HIFα stabilization and enhancement of hypoxic responses .
Cells and organisms exposed to environmental stress mount complex adaptive responses in order to maintain homeostasis . In mammals , hypoxic stress triggers cellular and systemic modifications , such as metabolic switches [1 , 2] , erythropoiesis [3 , 4] , angiogenesis and vasodilation [5 , 6] , resulting in reduced oxygen consumption and increased oxygen transport to hypoxic tissues . Responses to hypoxia are principally mediated by a family of transcription factors named Hypoxia Inducible Factors ( HIFs ) [7–12] , that are heterodimers composed of an oxygen regulated α-subunit ( HIFα ) and a constitutive β-subunit ( HIFβ ) [13 , 14] . HIFα activity is controlled by different mechanisms [15] , the most prevalent being oxygen-dependent regulation of protein stability . In normoxia , HIFα is hydroxylated on two specific prolyl residues within the oxygen-dependent degradation ( ODD ) domain , enabling binding to the von Hippel-Lindau ( VHL ) tumor suppressor protein , a component of the elongin BC/cullin-2/VHL ubiquitin-protein ligase complex , which targets HIFα for degradation at the 26S proteasome [16–18] . HIFα hydroxylation is catalyzed by specific prolyl-4-hydroxylases ( PHD1-PHD3 ) that are 2-oxoglutarate and Fe ( II ) -dependent dioxygenases [19 , 20] . Since PHDs use molecular oxygen as a co-substrate of the reaction , in hypoxia their activity is inhibited . Consequently , in hypoxia HIFα is not hydroxylated , accumulates , translocates to the nucleus , dimerizes with HIFβ and binds to HIF-responsive elements ( HREs ) , thus promoting transcription of target genes [21–23] . We and others have demonstrated that Drosophila melanogaster has a hypoxia-inducible transcriptional response that is homologous to that of mammals [24] , with Similar ( Sima ) [25] and Tango ( Tgo ) [26] being the homologs of HIFα and HIFβ , respectively [27 , 28] , and Fatiga ( Fga ) the only Drosophila PHD enzyme [29 , 30] . We have previously shown that the microRNA ( miRNA ) machinery is required for full activation of the Sima-dependent transcriptional response to hypoxia , both in cell culture and in vivo [31] . Yet , the individual miRNAs involved in Sima regulation remained unrevealed . Here , we performed an overexpression screen in Drosophila embryos aimed at defining miRNAs that regulate the hypoxic response , and identified specific miRNAs whose overexpression enhances Sima-dependent transcription . One of these miRNAs , miR-190 , is induced in hypoxia , is necessary for Sima-dependent gene expression and promotes terminal tracheal cell sprouting . Finally , we found that miR-190 directly targets the HIF prolyl hydroxylase fatiga transcript on its 3’UTR , thereby inhibiting its expression . We propose that miR-190 positively regulates Sima-dependent transcription by inhibiting the oxygen sensor Fatiga , which is the main negative regulator of the hypoxic response .
To identify miRNAs involved in the response to hypoxia in Drosophila , we performed an overexpression screen in stage 14–17 embryos . The rationale was that since suppression of the miRNA machinery inhibits the hypoxic response [31] , overexpression of certain specific miRNAs could potentially enhance this response . For the screen , we utilized a collection of 93 fly lines ( S1 Table ) to overexpress individual miRNAs under control of a breathless-Gal4 ( btl-Gal4 ) driver , and a HIF/Sima-dependent LacZ reporter ( HRE-LacZ reporter ) as a read out ( Fig 1A; [27] ) . This transgenic reporter was not expressed in normoxic embryos , but induced at 5% O2 in a Sima-dependent manner ( Fig 1A; [27] ) . In fatiga homozygous mutant embryos ( fga9 ) , Sima protein accumulates [29] , and hence , expression of the reporter was strongly upregulated even in normoxia [29] , being this induction suppressed by expression of sima RNAi ( Fig 1A and S1 Fig ) . Given that the biological effect of Drosophila miRNAs is often mild , we sought to conduct the screen under sensitized conditions . To define an appropriate sensitized condition of the hypoxia response system , we used a UAS-fatiga RNAi line ( fatigaRNAi; [32] ) whose effect is modest . In normoxia , expression of fatigaRNAi had no effect on HRE-LacZ reporter induction , while at mild hypoxia ( 11% O2 ) , β-galactosidase expression was readily detectable in these embryos ( Fig 1A ) . In embryos bearing only the btl-Gal4 driver , no induction of the reporter was observed under these same conditions ( Fig 1A ) . In strong hypoxia ( 5% O2 ) , reporter expression was enhanced in the fatigaRNAi line in comparison to wild type controls ( Fig 1A ) . Since mild hypoxia ( 11% O2 ) represented a sensitized condition for the hypoxia response machinery , in which potential effects of miRNAs regulating the system might become evident , we performed the screen by exposing the embryos that overexpressed miRNAs at 11% O2 for 4 h; isogenic embryos that did not overexpress any miRNA were used as negative controls ( Fig 1B ) . The screen was carried out in triplicate; overexpression of most miRNAs had no effect on HRE-LacZ reporter expression ( Fig 2A and 2B ) , but 4 out of the 93 tested miRNAs , namely miR-190 ( Fig 2C and 2G ) , miR-274 ( Fig 2D and 2G ) , miR-280 ( Fig 2E and 2G ) and miR-985 ( Fig 2F and 2G ) , scored as positives in the screen , inducing expression of the reporter . miR-970 , one of the many miRNAs that had no effect on reporter expression , was randomly chosen as a negative miRNA control , and used in the rest of the experiments carried out in this work . We focused our studies on miR-190 , whose occurrence in vivo has been experimentally validated by high-throughput sequencing of small RNA libraries generated from different tissues and developmental stages [33 , 34] . In order to confirm miR-190 participation in the Fatiga/Sima pathway , we began by studying biological responses characteristic of Sima accumulation . We previously reported that fatiga loss-of-function mutations provoke accumulation of high levels of Sima in normoxia , resulting in lethality at the pupal stage [29] and an increased number of terminal ramifications in 3rd instar larval tracheae [35] . Since our results suggested that miR-190 is a positive regulator of Sima ( Fig 2 ) , we tested whether overexpression of miR-190 can also induce similar developmental phenotypes , and to what extent they depend on Sima activity . When overexpressed with an engrailed-Gal4 ( en-Gal4 ) driver , miR-190 , but not the control miRNA ( miR-970 ) , was associated with lethality at pupal or pharate adult stages ( Fig 3A ) . Knock-down of sima by RNAi completely rescued the lethality caused by miR-190 overexpression , suggesting that lethality was indeed due to Sima accumulation ( Fig 3A ) . In addition , coexpression of Fatiga B , one of the isoforms of the Drosophila HIF prolyl hydroxylase , also rescued the lethal phenotype ( Fig 3A ) , further suggesting that over-accumulation of Sima was the causal factor . When expressed alone , neither sima RNAi nor Fatiga B overexpression had effects on viability ( Fig 3A ) . Tracheal terminal cells of Drosophila 3rd instar larvae are plastic and ramify in response to hypoxia ( Fig 3D and 3E; [36] ) in a Sima- and Fatiga-dependent manner [35 , 37] . As we previously reported , the number of terminal branches with more than 1 μm diameter ( “thick terminal branches” , TTBs ) of the dorsal branch of the 3rd segment of 3rd instar larvae is a sensitive parameter to quantify terminal tracheal branching after physiological or genetic interventions [35] . To investigate whether miR-190 can also modulate this process , we overexpressed miR-190 under control of the tracheal terminal cell-specific driver dSRF-Gal4 . In normoxic larvae overexpressing this miRNA , we observed a significant increase in the number of TTBs ( Fig 3C and 3F ) in comparison to controls expressing the Gal4 driver only ( Fig 3B and 3F ) , or larvae overexpressing an unrelated miRNA ( miR-970 ) ( Fig 3F ) . To investigate if this increase of ramification depends on Sima , we coexpressed miR-190 along with a UAS-simaRNAi , and observed complete reversion of the phenotype , attaining these larvae a normal number of TTBs ( Fig 3F ) . Expression of the sima RNAi on itself did not induce changes in tracheal terminal sprouting . These results indicate that overexpression of miR-190 can induce Sima-dependent tracheal terminal sprouting , a typical physiological response to hypoxia . To get additional evidence that miR-190 participates in the HIF pathway , we analyzed genetic interactions between miR-190 , fatiga and sima , by assessing induction of the HRE-LacZ reporter as a read out . Overexpression of miR-190 with a btl-Gal4 driver in mild hypoxia enhanced expression of the HRE-LacZ reporter ( Fig 2 ) in comparison with control individuals expressing an unrelated RNAi ( Fig 4A ) ; co-expression of this miRNA along with sima RNAi suppressed this enhancement ( Fig 4A ) . Overexpression of miR-190 along with Fatiga B , a highly active isoform of the oxygen sensor Fatiga [30] , sharply decreased induction of the reporter ( Fig 4A ) . These results indicate that miR-190 enhances the HIF pathway , antagonizing the activity of the prolyl-4-hydroxylase Fatiga . To analyze further these genetic interactions , we utilized miR-190 null mutant embryos ( miR-190KO , [38] ) . Unlike the previous experiments in which the HRE-LacZ reporter was utilized in heterozygosis ( Figs 1 , 2 and 4A ) , the reporter was used in homozygosis to favor reporter induction in wild type embryos exposed to mild hypoxia ( Fig 4B ) . Noteworthy , this induction was suppressed in miR-190KO mutants ( Fig 4B ) , confirming that miR-190 contributes to Sima-dependent transcription . In fatiga homozygous mutant embryos ( fga9 ) , induction of the reporter occurs ( Figs 1A and 4B; [29] ) , and interestingly , this expression was not altered in miR-190KO homozygotes ( Fig 4B ) , indicating that miR-190 operates upstream of the fatiga gene . Taken together , our genetic interactions data are consistent with a model in which miR-190 inhibits Fatiga , resulting in an enhancement of the hypoxic response . Having analyzed HRE-LacZ reporter induction upon miR-190 loss- and gain-of-function , we studied if miR-190 affects the expression of endogenous Sima target genes . We measured mRNA levels of two well-established Sima targets by real time RT-PCR , namely fatiga B ( fgaB ) and heat shock factor ( hsf ) [30 , 39] in embryos with gain- or loss-of-function of miR-190 . Ubiquitous overexpression of miR-190 with an actin-Gal4 ( act-Gal4 ) driver in embryos maintained in normoxia or exposed to mild hypoxia ( 11% O2 ) for 4 h induced upregulation of fgaB and hsf transcripts in comparison to control embryos carrying only the act-Gal4 driver or overexpressing a control miRNA ( Fig 5A and 5B ) . We confirmed these results in Drosophila S2R+ cells , where overexpression of miR-190 also resulted in upregulation of both fgaB and hsf mRNAs , in comparison with cells transfected with the empty vector ( S2 Fig ) . Next , we examined whether hypoxic induction of the HIF target genes fgaB and hsf is affected in miR-190 knock-out ( miR-190KO ) homozygous embryos or in embryos heterozygous for miR-190KO and the rhea79a microdeletion that covers the rhea locus [40]; miR-190 is encoded in an intron of the rhea gene [33 , 34] ( S3 Fig ) . Hypoxic induction of both HIF target genes was severely impaired in miR-190 loss-of-function embryos ( Fig 5C and 5D ) , indicating that miR-190 is necessary for HIF activation . The results described so far demonstrate that miR-190 positively regulates Sima . Therefore , to investigate the mechanisms of Sima regulation by miR-190 , we measured sima mRNA abundance following miR-190 overexpression . Using a ubiquitous act-Gal4 driver , we overexpressed miR-190 in embryos exposed to either normoxia or mild hypoxia ( 11% O2 ) for 4 h , and measured sima mRNA levels by quantitative real time RT-PCR . No differences were detectable in sima transcript levels , either in normoxia or in mild hypoxia ( S4 Fig ) , indicating that the miR-190 regulatory mechanism is independent of sima transcription or mRNA stability . To identify direct targets of miR-190 , we searched for target genes related to HIF-dependent response to hypoxia using publicly available database . The miRNA target prediction database miRanda ( www . microrna . org ) [41–43] predicted two potential miR-190 binding sites within the 3’ UTR of the prolyl-4-hydroxylase fatiga , the main negative regulator of Sima . To determine whether miR-190 can regulate fatiga expression , we used a transgenic reporter construct that directly responds to Fatiga activity . This ubiquitously expressed reporter construct consists of a Green Fluorescent Protein ( GFP ) fused to the Sima oxygen-dependent degradation ( ODD ) domain , which is rapidly degraded when Fatiga is active . Conversely , the fusion protein accumulates when Fatiga activity diminishes ( Tvisha Misra and Stefan Luschnig , personal communication ) . We overexpressed miR-190 or a control miRNA with an engrailed-Gal4 driver in the posterior compartment of wing imaginal discs , and analyzed the behavior of the GFP-ODD reporter by confocal microscopy . While expression of the control miRNA ( miR-970 ) did not induce changes in GFP-ODD reporter levels , expression of miR-190 resulted in increased GFP signal in the posterior compartment of the discs ( Fig 6A–6H ) , indicating a stabilization of the GFP-ODD reporter , and suggesting downregulation of Fatiga . A Red Fluorescent Protein ( RFP ) expressed under the same ubiquitous promoter was used as an expression reference construct . RFP labeling was homogenous throughout the disc and therefore unaffected by expression of the miRNAs ( Fig 6C and 6D ) . To investigate whether fatiga is a direct target of miR-190 , we analyzed the expression of a luciferase reporter in which the firefly luciferase coding sequence is fused to the 3’UTR of fatiga ( Fig 6I ) . The experiment was carried out in S2R+ cells transfected with a plasmid driving the expression of miR-190 , in comparison to cells transfected with an empty vector; miR-12 and its specific luciferase reporter [31 , 44] were utilized as a positive control of the system ( S5 Fig ) . Importantly , transfection of the plasmid expressing miR-190 strongly reduced luciferase activity of the reporter containing the fatiga 3’UTR , as compared to control cells transfected with the empty vector ( Fig 6J ) . To assess binding specificity of miR-190 , we mutagenized the strongest miR-190 recognition site within the fatiga 3’UTR ( Fig 6I ) . The reporter bearing the mutant binding site became insensitive to the expression of miR-190 ( Fig 6J ) , confirming specificity of the miRNA . Collectively , these data demonstrate that miR-190 directly targets and downregulates fatiga . We next investigated if miR-190 expression is regulated by oxygen . RT-qPCR analysis revealed a significant increase of miR-190 expression in wild type embryos exposed to hypoxia ( 5% O2 for 4 h ) , in comparison to controls maintained in normoxia ( Fig 7A ) . To determine if hypoxic induction of miR-190 depends on Sima , we analyzed miR-190 levels in embryos exposed to hypoxia and expressing sima RNAi . sima knock-down did not affect miR-190 hypoxic induction , suggesting that upregulation of miR-190 in hypoxia is independent of Sima ( Fig 7A ) . To investigate if miR-190 upregulation in hypoxia is regulated at a transcriptional level , we evaluated the expression of pre-miR-190 . As depicted in Fig 7B , pre-miR-190 expression increased in hypoxia as compared to normoxia , and this induction was again unaffected after sima knock-down . These results suggest that hypoxic upregulation of miR-190 occurs at a transcriptional level , in a Sima-independent manner . Given that miR-190 is encoded in an intron of the rhea gene ( S3 Fig ) , we investigated if rhea transcript levels are also upregulated in hypoxia . Similarly to miR-190 , rhea was upregulated in hypoxia in a Sima-independent manner ( Fig 7C ) . As a control of the effect of sima silencing , we assessed in the same embryos the expression of fatiga B , which is a well-known Sima target [30] . As shown in Fig 7D , fatiga B transcript levels were strongly increased in hypoxic wild type embryos , and this induction was reduced upon sima knock-down . Taken together , this set of experiments suggests that miR-190 is transcriptionally induced in hypoxia , as part of the rhea transcript , in a Sima-independent manner ( S3 Fig ) .
Drosophila melanogaster has proved to be a useful model for studying the function of miRNAs as regulators of developmental programs , as well as in the maintenance of cellular homeostasis [38] . In the current work , we have carried out an in vivo screen , aimed at the identification of miRNAs involved in HIF-dependent hypoxic responses in Drosophila . Among 93 miRNAs tested , we identified miR-190 , miR-274 , miR-280 and miR-985 as positive regulators of Sima-dependent transcription . In mammalian cells , several miRNAs have been reported to participate in the response to hypoxia . Certain miRNAs , such as miR-20b , miR-199a , miR-155 , miR-122 , miR195 , miR-335 , miR-33a and miR-18a inhibit HIFα expression directly by binding its 3’UTR [45–52] . Other miRNAs , such as miR-424 , miR-184 , miR-210 , miR-130 , miR-494 , miR-21 and miR-17 regulate HIFα expression positively through indirect mechanisms [53–60] , which involve inhibition of negative regulators of this transcription factor . For example , miR-424 directly targets and reduces the expression of cullin2 ( CUL2 ) , a scaffold component of the ubiquitin ligase complex that targets HIFα for degradation in the 26S proteasome [53] . Likewise , miR-184 inhibits another cardinal regulator of HIFα: the factor inhibiting HIF-1 ( FIH-1 ) , an asparagine hydroxylase that hydroxylates HIFα , thereby inhibiting its association with the p300 transcriptional coactivator [54 , 61] . Another interesting example is the direct silencing of the succinate dehydrogenase complex subunit D ( SDHD ) by miR-210: inhibition of SDHD leads to accumulation of its substrate , succinate , which is in turn a product of HIFα prolyl hydroxylase ( PHD ) activity with inhibitory effects on the enzyme [62] , which finally results in HIFαstabilization [55] . In this study , we have shown that miR-190 directly targets and downregulates the oxygen sensor fatiga , thereby exerting positive regulation on the hypoxia master transcription factor Sima ( Fig 8 ) . miR-190 is induced in hypoxia , a condition in which Fatiga activity is also inhibited due to low oxygen availability ( Fig 8 ) , providing a mechanism by which miR-190 enhances the strength of the hypoxic response . To our knowledge , this is the first report of a miRNA that directly downregulates an oxygen sensing prolyl-4-hydroxylase . As documented in the miRNA database miRBase ( www . mirbase . org ) , miR-190 is broadly conserved in evolution , not only within the Drosophilid lineage [34] , but also in distant taxa , including mammals . In most mammalian species , two miR-190 family members occur , miR-190a and miR-190b . The miR-190a locus lies in an intron of talin2 ( TLN2 ) , which encodes a high molecular weight cytoskeletal protein . Remarkably , Drosophila melanogaster miR-190 is encoded in an intron of the gene rhea ( S3 Fig ) , the homolog of talin2 ( TLN2 ) . Intron 53 of human TLN2-001 ( which is 12 , 893 nucleotides long ) and intron 14 of rhea-RB ( which is 356 nucleotides long ) only share sequence similarity within the miR-190 locus [33 , 34 , 63–70] , reflecting the physiological relevance of this miRNA and perhaps some biological link with Rhea/Talin2 . Interestingly , human PHD3 ( also known as EGLN3 ) , which is one of the three mammalian homologs of Drosophila Fatiga [19] , has a predicted binding site for miR-190a , according to the miRNA target prediction databases TargetScan ( www . targetscan . org ) [71] and miRDB ( mirdb . org ) [72] , even though with a relatively low score in both cases . Thus , it is possible that miR-190-dependent regulation of HIF-prolyl hydroxylases is conserved in evolution . We found that Drosophila miR-190 is induced in hypoxia . Interestingly , mammalian miR-190 is upregulated in different types of cancer , including hepatocellular carcinoma [73 , 74] , primary myelofibrosis [75] , pancreatic [76] , breast [77–79] , rectal [80] and papillary thyroid cancer [81] . Hypoxic microenvironment is a common feature of many solid tumors [82 , 83] , and most primary human cancers and their metastases exhibit increased levels of HIFα [84] . In addition to intratumoral hypoxia , genetic and epigenetic alterations can also stimulate HIF activity within tumors [82 , 84 , 85] . HIF promotes angiogenesis [82 , 86] , metabolic switches [87] , metastasis [88] and chemo/radio-resistance of cancer cells [89 , 90] , and high levels of HIF are associated with poor patient prognosis and increased mortality [84 , 91] . On the other hand , many different miRNAs have been shown to play pivotal roles in cancer development , functioning as oncogenes or tumor suppressors [92–94] . Given that miR-190 is upregulated in diverse cancer types , our findings open the possibility that miR-190 contributes to HIFα stabilization in cancer cells , thereby enhancing tumor progression . In line with this possibility , miR-190 directly inhibits the PH domain leucine-rich repeat protein phosphatase ( PHLPP ) , a tumor suppressor protein that inactivates the kinase AKT through Ser437 dephosphorylation [95–97] . In human bronchial epithelial cells , trivalent arsenic ( A3+ ) induces the expression of miR-190 , which binds the 3’UTR of PHLPP transcript , decreasing PHLPP protein levels [95 , 96] . As a consequence , AKT phosphorylation and activation increase , finally resulting in vascular endothelial growth factor ( VEGF ) expression [95] , which is induced following AKT activation [98] . Another bona fide miR-190 target is IGF-1 , which is significantly reduced in serum of patients with hepatocellular carcinoma . Accordingly , miR-190b is upregulated in tumor tissues , contributing to insulin resistance through downregulation of IGF-1 , which is associated with poor prognosis [73] . Thus , miR-190 favors carcinogenesis through distinct pathways . Importantly , strengthening the notion of a possible involvement of miR-190 in mammalian responses to low oxygen , miR-190 is induced by hypoxia in a rat model of hypoxic pulmonary artery hypertension ( PAH ) [99–102] . miR-190 directly targets and represses the expression of Kcnq5 , a member of the voltage-gated K+ channel family , resulting in augmented vasoconstriction of the pulmonary artery , a hallmark of hypoxic PAH [101] . In summary , the results reported here increase our understanding of the network controlling HIF-dependent responses to hypoxia , and open the possibility of analyzing the regulation exerted by additional miRNAs which may be part of this complex network .
The UAS-miRNA fly collection utilized in this study was previously described [103] . The following fly stocks were from the Bloomington Drosophila Stock Center ( Indiana University , Bloomington , IN , USA ) : w1118 , breathless-Gal4 , engrailed-Gal4 , dSRF-Gal4 , actin-Gal4 , UAS-GFP , UAS-white RNAi and miR-190KO . The following stocks were from the Vienna Drosophila RNAi Center: UAS-fatiga RNAi ( VDRC 103382 ) , UAS-sima RNAi ( VDRC 106504 ) . The HRE-LacZ reporter [27] , UAS-Fatiga B [30] and fga9/TM3 [29] lines were generated in our laboratory and previously described . The rhea79a [40] mutant was kindly provided by Nicholas Brown . Hypoxia was applied in a Forma Scientific 3131 incubator , by regulating the proportions of oxygen and nitrogen . To obtain synchronized individuals , embryos were collected on egg-laying agar plates for 4 h , and then incubated at 18°C or 25°C in normoxia until the desired stage . When necessary , embryos or first-instar larvae were sorted to obtain the desired genotypes using a fluorescent Olympus stereomicroscope MVX10 . For X-gal stainings , embryos were dechorionated in bleach for 1 min , incubated with heptane for 5 min , fixed with glutaraldehyde 0 . 5% in PBS for 20 min and then washed three times for 5 min in PT 0 . 3% ( PBS containing 0 . 3% Triton-X 100 ) . Samples were incubated 1 h with the staining solution ( 5 mM K4Fe ( CN ) 6 , 5 mM K3Fe ( CN ) 6 , 0 . 2% X-gal ) at 37°C . After three washes with PT 0 . 3% , samples were analyzed using an Olympus stereomicroscope MVX10; and photographed after mounting in glycerol 80% with an Olympus BX60 microscope equipped with an Olympus DP71 digital camera . The screen was performed using the Fly Condo ( Flystuff , San Diego , CA , USA ) , which contains 24 independent chambers , allowing for high-throughput collection of Drosophila embryos . In each chamber , we placed adult males bearing the btl-Gal4 driver and the HRE-LacZ reporter , together with females of one miRNA line or a wild type line ( w1118 ) as a negative control . Embryos from the offspring were collected in the 24-well stainless steels mesh plate provided with the condo and subjected to hypoxia ( 11% O2 ) , for 4 h . Next , we evaluated the expression of the HRE-LacZ reporter performing X-gal stainings of the embryos within the mesh plate . First-instar larvae were placed in fresh vials , at a density of 20 individuals per vial . When they reached the third-instar wandering stage , larvae were anesthetized with ether and ramifications of the terminal cell of the trachea in the dorsal branch of the third segment were counted and photographed using bright-field microscopy . Embryos were incubated under hypoxia ( 11% , 8% or 5% O2 ) or normoxia for 4 h , at 25°C . Next , total RNA was isolated using Trizol reagent ( Invitrogen , Carlsbad , CA , USA ) from embryos of stages 14–17 . Genomic DNA was removed from RNA samples using Ambion’s DNase ( Ambion , Austin , TX , USA ) . Samples ( 1 μg ) were reverse-transcribed with the M-MLV Reverse Transcriptase ( Invitrogen , Carlsbad , CA , USA ) , following the manufacturer´s instructions , using oligo-dT as a primer . The concentration and integrity of RNA and cDNA were determined using Nanodrop ND-1000 spectrophotometry and gel electrophoresis . The resulting cDNA was used for quantitative real time PCR , using a MX3005P instrument ( Stratagene , La Jolla , CA , USA ) . The real time PCR reaction contained: 1 μL Sybr Green 1/1000 , 0 . 3 μL ROX reference dye 1/10 ( Invitrogen , Carlsbad , CA , USA ) , 0 . 2 μL of Taq DNA Polymerase Recombinant ( Invitrogen , Carlsbad , CA , USA ) , 2 . 5 μL Buffer 10X , 1 μL MgCl2 50 mM , 0 . 5 μL dNTP mixture 10 mM ( Invitrogen , Carlsbad , CA , USA ) , 1 μL of sense primer 10 μM , 1 μL of anti-sense primer 10 μM , 5 μL of template cDNA 1/30 , 4 . 2 μL glycerol 30% and 8 . 3 μL of H2O . The thermal cycling conditions were the following: 95°C for 10 min , followed by 40 cycles at 95°C for 30 s , 60°C for 1 min and 72°C for 1 min , finishing with a cycle for the melting curve of 95°C for 1 min , 60°C for 30 s and 95°C for 30 s . Relative mRNA expression was normalized using rpl29 , rpl32 or GAPDH as internal controls . For quantification of miR-190 levels , the NCode VILO miRNA cDNA Synthesis Kit ( Invitrogen , Carlsbad , CA , USA ) was used , following manufacturer´s instructions . The 2S rRNA was used for normalization in quantitative real time PCR determinations of miR-190 . Larvae were dissected in PBS , fixed in 4% formaldehyde ( Sigma , St . Louis , MO , USA ) for 40 min at room temperature and then washed three times for 10 minutes in PT 0 . 3% ( PBS containing 0 . 3% Triton-X 100 ) . Samples were blocked with bovine serum albumin 5% in PT 0 . 3% ( PBT ) for 2 h and then incubated with the primary antibody in PBT overnight at 4°C . After washing three times for 15 min with PT 0 . 3% , tissues were incubated for 2 h at room temperature with the secondary antibody in normal goat serum 5% diluted in PT 0 . 3% . Next , samples were washed , imaginal discs were separated and mounted in glycerol 80% . Images were analyzed and captured using a Carl Zeiss LSM510 Meta Confocal Microscope . We used mouse anti-Engrailed ( 1:100; Developmental Studies Hybridoma Bank , Iowa , IA , USA ) primary antibody and donkey anti-mouse Cy5 ( Jackson , ImmunoResearch Laboratories Inc . , West Grove , PA , USA ) secondary antibody . Fluorescence of GFP and RFP was analyzed without antibody staining . The copper-inducible pMT/V5-His plasmid ( Invitrogen , Carlsbad , CA , USA ) was utilized as a backbone vector for generating reporter constructs . For generation of pMT-Luciferase renilla ( pMT-Renilla ) , the coding sequence of renilla luciferase was subcloned from a pRL-SV40 vector ( Promega , Madison , WI , USA ) into HindIII/XbaI sites of pMT/V5-His . For the pMT-Luciferase firefly reporter construct ( pMT-Firefly ) , the firefly luciferase coding sequence was subcloned from pGL3 vector ( Promega , Madison , WI , USA ) into EcoRI/XbaI sites of pMT/V5-His . fatiga 3’UTR sequence was generated by PCR from cDNA prepared from Drosophila yellow white embryos and cloned into XbaI/ApaI restriction sites of the pMT-Firefly plasmid . The primers utilized were: Forward ( Fw ) : 5’-GCTCTAGACCCAAGCCGACAGCGCAGCT-3’; Reverse ( Rv ) : 5’-GCCATTGGGCCCCATCAGCTCAGGCTTTTGTTTA-3’ . Point mutations in miR-190 binding site at the fatiga 3’UTR were introduced by nested PCR with the following primers: Fw: 5’-CTGTAAATCATGAAGTATGTATATTTATGCCCTCGCTACATATTGTATG-3’; Rv: 5’-CATACAATATGTAGCGAGGGCATAAATATACATACTTCATGATTTACAG-3’ . The Luc-CG10011 3’UTR reporter and pAc-miR-12 were a gift from E . Izaurralde [44] . The pAc-miR-190 overexpression plasmid was kindly provided by M . Milán [104] . The pAc-5 . 1/V5-His ( Invitrogen , Carlsbad , CA , USA ) was used as a negative control . Semi-adherent Schneider ( S2R+ ) Drosophila cells were maintained in Schneider Drosophila medium ( Sigma , St . Louis , MO , USA ) supplemented with Penicillin ( 50 U/ml , Invitrogen ) , Streptomycin ( 50 μg/ml , Invitrogen ) and 10% fetal bovine serum ( Invitrogen , Carlsbad , CA , USA ) at 25°C in 25 cm2 T-flasks . Cells were seeded in 24-well plates at a 35000 cells per well density and 0 . 3 μg of total DNA was transfected employing the Effectene transfection reagent ( Qiagen , Valencia , CA , USA ) . All pMT-Firefly-3’UTR constructs were co-transfected at a 1:1 proportion with pMT-Renilla to normalize transfection efficiency . Expression of luciferase from pMT vectors was induced 24 h after transfection by addition of 0 . 7 mM CuSO4 for 7 h . Firefly and renilla luciferase activities were measured by the Dual-Glo Luciferase Assay System ( Promega , Madison , WI , USA ) , following the instructions of the manufacturer , in a Veritas Microplate Luminometer ( Turner BioSystems ) . Data are expressed as mean ± standard deviation ( SD ) . Infostat Statistical Software was used for statistical analysis . Comparisons were performed using one- or two-way analysis of variance ( ANOVA ) followed by Fisher's protected least significant difference ( LSD ) as post hoc test , or unpaired two-tailed Student's t-test . Data were tested for normality ( Shapiro–Wilks test ) and variance homogeneity ( Levene test ) to use parametric statistical analysis . If data did not fulfill these statistical criteria , Welch's correction or the Kruskal-Wallis one-way ANOVA non-parametric test were used . A p<0 . 05 was considered statistically significant .
|
Sufficient oxygen supply is essential for animal survival . When cells or organisms are exposed to low oxygen levels ( hypoxia ) , a complex molecular response is triggered , enabling adaptation to this stressful condition . A key mediator of this response is HIF , a transcription factor that induces the expression of a set of genes that mediate the adaptive response to hypoxia . The most important regulation of HIF is exerted by a family of prolyl-4-hydroxylases ( PHDs ) , which prevent HIF accumulation under normal oxygen levels and lift this inhibition of HIF only in hypoxia . This pathway is highly conserved among metazoans , including humans and the fruit fly Drosophila melanogaster . microRNAs ( miRNAs ) , which are small ( ~22 nucleotides long ) , non-coding RNAs that control gene expression post-transcriptionally , play central roles in stress responses . In the present study , we have performed a screen in Drosophila and identified miRNAs that regulate HIF-dependent adaptations to hypoxia . We found one miRNA , miR-190 , that is induced in hypoxia and in turn enhances HIF-dependent biological responses , as well as the expression of HIF-inducible genes . The mechanism of action of miR-190 involves the inhibition of the Drosophila PHD , thereby positively regulating HIF-dependent responses to hypoxia at the molecular and organismal level .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"invertebrates",
"rna",
"interference",
"gene",
"regulation",
"oxygen",
"animals",
"animal",
"models",
"developmental",
"biology",
"micrornas",
"model",
"organisms",
"drosophila",
"melanogaster",
"hypoxia",
"molecular",
"biology",
"techniques",
"epigenetics",
"embryos",
"drosophila",
"research",
"and",
"analysis",
"methods",
"embryology",
"genetic",
"interference",
"hyperexpression",
"techniques",
"gene",
"expression",
"chemistry",
"molecular",
"biology",
"insects",
"molecular",
"biology",
"assays",
"and",
"analysis",
"techniques",
"arthropoda",
"gene",
"expression",
"and",
"vector",
"techniques",
"biochemistry",
"rna",
"chemical",
"elements",
"cell",
"biology",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"metamorphosis",
"physical",
"sciences",
"non-coding",
"rna",
"larvae",
"organisms"
] |
2016
|
miR-190 Enhances HIF-Dependent Responses to Hypoxia in Drosophila by Inhibiting the Prolyl-4-hydroxylase Fatiga
|
Elevated plasma triglyceride ( TG ) levels are an established risk factor for type-2 diabetes ( T2D ) . However , recent studies have hinted at the possibility that genetic risk for TG may paradoxically protect against T2D . In this study , we examined the association of genetic risk for TG with incident T2D , and the interaction of baseline TG with TG genetic risk on incident T2D in 13 , 247 European-Americans ( EA ) and 3 , 238 African-Americans ( AA ) from three prospective cohort studies . A TG genetic risk score ( GRS ) was calculated based on 31 validated single nucleotide polymorphisms ( SNPs ) . We considered several baseline covariates , including body- mass index ( BMI ) and lipid traits . Among EA and AA , we find , as expected , that baseline levels of TG are strongly positively associated with incident T2D ( p<2 x 10-10 ) . However , the TG GRS is negatively associated with T2D ( p=0 . 013 ) , upon adjusting for only race , in the full dataset . Upon additionally adjusting for age , sex , BMI , high-density lipoprotein cholesterol and TG , the TG GRS is significantly and negatively associated with T2D incidence ( p=7 . 0 x 10-8 ) , with similar trends among both EA and AA . No single SNP appears to be driving this association . We also find a significant statistical interaction of the TG GRS with TG ( pinteraction=3 . 3 x 10-4 ) , whereby the association of TG with incident T2D is strongest among those with low genetic risk for TG . Further research is needed to understand the likely pleiotropic mechanisms underlying these findings , and to clarify the causal relationship between T2D and TG .
Along with age , body weight , family history , and other characteristics , triglyceride ( TG ) level is a major risk factor for cardiovascular disease and type-2 diabetes ( T2D ) . In many prospective studies , TG at baseline is positively associated with T2D incidence , independently of body mass index ( BMI ) and other risk factors [1] , although the direction of causality between TG and T2D is unclear [2] , and somewhat ambiguous . Over thirty single nucleotide polymorphisms ( SNPs ) have been found to be associated with plasma TG levels through meta-analysis of genome-wide association studies ( GWAS ) [3 , 4] . Using a case-control analytical design in over 30 , 000 individuals , we previously examined the association of 17 genetic risk scores ( GRS ) for various traits , including lipid traits , with current T2D status ( importantly , adjusting only for sex and age ) , and observed an unexpected , albeit not statistically significant , pattern suggesting that being at high genetic risk for low-density lipoprotein cholesterol , total cholesterol , or TG could be protective against T2D [5] . At the same time , another group found in a cross-sectional study that genetic risk for dyslipidemia is negatively associated with fasting glucose and HbA1c [6] after adjustment for lipid phenotypes . As illustrated and discussed by Li et al . [6] , by not adjusting for TG , the negative direct path between TG genes and T2D is weakened by the positive path linking TG genes , TG , and glycemic traits . Previous studies have shown that GCKR variation is associated with both elevated TG and reduced fasting glucose and T2D risk , [7 , 8] potentially implicating its role in hepatic de-novo lipogenesis [9] , through which it would promote glucose uptake , glycolysis , and subsequently increase plasma TG levels [10] . In a Mendelian randomization study , no evidence to suggest a causal role for TG in T2D or related traits was identified , although the authors did identify a “suggestive protective association” of genetically-raised TG levels with T2D risk [11] . Finally , Qi et al . identified a significant positive association between genetic risk for TG and T2D risk . However , in that analysis there was no adjustment for plasma TG levels which would have biased the results towards the null , according to Li et al . [6] . Likely more important , is their exclusion of SNPs with known pleiotropic effects , such as GCKR [12] . Here , we seek to further examine these seemingly paradoxical findings by drawing from three prospective cohort studies and across two racial groups , and by considering baseline measures of BMI , high-density lipoprotein cholesterol ( HDL-C ) , low-density lipoprotein cholesterol ( LDL-C ) , TG , fasting glucose ( FG ) and fasting insulin ( FI ) as covariates . Specifically , we hypothesize that baseline TG is associated with incident T2D , that genetic risk for TG is negatively associated with T2D incidence , and that the association between TG and incident T2D is stronger among those individuals with low genetic risk for TG .
Among EA , TG was strongly positively associated with T2D incidence ( hazard ratio ( HR ) = 1 . 14 , 95% CI [1 . 09–1 . 18] , p = 8 . 8 x 10–11 ) , conditioning upon age , sex , BMI , and HDL . The association of the TG GRS with incident T2D , including no other covariates , was not statistically significant , but was in the negative direction ( HR = 0 . 997 , 95% CI [0 . 994–1 . 000] , p = 0 . 06 ) . Upon conditioning for age , sex , and BMI , the TG GRS was significantly negatively associated with T2D incidence ( HR = 0 . 996 , 95% CI [0 . 993–0 . 999] , p = 0 . 008 ) . Finally , upon conditioning for age , sex , BMI , TG , and HDL , the TG GRS was significantly negatively associated with T2D incidence ( HR = 0 . 991 , 95% CI [0 . 988–0 . 995] , p = 1 . 2 x 10–7 ) . We did not include LDL-C as a covariate as it was not significantly associated with T2D incidence in a model which included BMI , TG and HDL-C . Including LDL-C in the full model resulted in a similar association of TG GRS with T2D ( HR = 0 . 991 , 95% CI [0 . 988–0 . 994] , p = 2 . 6 x 10–8 ) . Including FG and FI as additional covariates resulted in a similar but weaker negative association of the TG GRS with T2D ( HR = 0 . 994 , 95% CI [0 . 990–0 . 998] , p = 3 . 1 x 10–3 ) . Among AA , TG was strongly positively associated with T2D incidence ( HR = 1 . 18 , 95% CI [1 . 12–1 . 23] , p = 3 . 6 x10-11 ) , conditioning on age , sex , BMI , and HDL . The TG GRS was negatively but not significantly associated with T2D incidence ( HR = 0 . 996 , 95% CI [0 . 991–1 . 001] , p = 0 . 087 ) , conditioning only on age and sex . As in EA , LDL-C was not significantly associated with T2D incidence in any model . Conditioning upon age , sex , BMI , HDL , and TG , the TG GRS association with T2D incidence was negative , but not statistically significant ( HR = 0 . 995 , 95% CI [0 . 991–1 . 000] , p = 0 . 055 ) . Fig 1 shows the TG GRS association on a study-by-study basis in both EA and AA . We examined the association of the TG GRS with T2D incidence , conditioning only on race , in a combined dataset of EA and AA . We find that the TG GRS is negatively and significantly associated with T2D incidence ( HR = 0 . 997 , 95% CI [0 . 994–0 . 999] , p = 0 . 013 ) . Upon conditioning on race as well as age , sex , TG and HDL , we find a very strong negative association of the TG GRS with T2D ( HR = 0 . 993 , 95% CI [0 . 990–0 . 995] , p = 7 . 0 x 10–8 ) . Of 31 TG-associated SNPs , 24 ( 77 . 4% ) , and 15 ( 48 . 4% ) are negatively associated with T2D incidence in EA and AA , respectively ( see Table 2 ) . However , none are significantly associated with T2D incidence . SNPs that are negatively associated with T2D among EA , at a nominal level of significance ( p<0 . 05 ) are those in/near APOA1 , PLA2G6 , GCKR , CETP , and CILP2 . Among AA , no single SNP was at least nominally significant . However , we did observe the same negative trend of association for all SNPs that are nominally significant in EA , except for PLA2G6 . Among EA , we find a significant interaction of TG and TG GRS ( HRinteraction = 0 . 997 , 95% CI [0 . 995–0 . 999] , p = 1 . 1 x 10–3 ) , with the inclusion of age , sex , BMI , TG and HDL as covariates . Fig 1 shows the TG x TG GRS interaction HRs on a study-by-study basis in EA and AA . Fig 2 shows the association of TG with incident T2D in each of three tertiles of TG GRS . Among EA , we observe a stronger association of TG with incident T2D among individuals with a low TG GRS ( HR = 1 . 24 , 95% CI [1 . 16–1 . 32] , p = 8 . 0 x 10–10 ) compared to those with a high TG GRS ( HR = 1 . 12 , 95% CI [1 . 05–1 . 19] , p = 4 . 6 x 10–4 ) . Among AA , we do not find a statistically significant interaction of TG and TG GRS ( HRinteraction = 0 . 998 , 95% CI [0 . 993–1 . 0003] , p = 0 . 39 ) , although the trend is similar as that in EA ( see Fig 2 ) . Including all possible TG x covariate and TG GRS x covariate interaction terms in the model results in a TG x TG GRS interaction that has the same direction as the model without all possible interaction terms , but that is not statistically significant ( EA: p = 0 . 15 , AA: p = 0 . 38 , combined: p = 0 . 06 ) . In a combined analysis , conditioning upon race as well as the other covariates , we find a statistically significant interaction of the TG GRS and TG on T2D ( HRinteraction = 0 . 997 , 95% CI [0 . 995–0 . 999] , p = 3 . 3 x 10–4 ) . Among the 31 individual SNPs , only rs2929282 in FRMD5 shows a significant interaction ( p = 2 . 2 x 10–4 ) with TG on T2D incidence among EA ( see Table 2 ) .
Using data from three prospective cohorts , we show that genetic risk for elevated TG is negatively associated with T2D incidence . Furthermore , our findings suggest that the risk conferred by TG is greatest among individuals with low genetic predisposition for TG . We find a similar pattern in both EA and AA , despite 1 ) the smaller sample size of AA , 2 ) the fact that TG SNPs were mainly identified in EA , and 3 ) that the association of the TG GRS with TG among AA in our sample is relatively weak , compared to that in EA . Our findings are in accord with those of Klimentidis et al . [5] and Li et al . [6] . By adjusting analyses for phenotypic levels of TG and HDL , we were more easily able to identify the negative association of the TG GRS with T2D , and thus confirm the likely pleiotropic nature of TG genes . However , our results differ with respect to the specific SNPs identified . Li et al . identified SNPs in/near CETP , MLXIPL , PLTP , GCKR , APOB , APOE-C1-C2 , CYP7A1 and TIMD4 as being most strongly negatively associated with glycemic traits . Among these , we find that only CETP and GCKR SNPs are negatively associated with T2D , albeit at a nominal significance level . All others , except APOE , were directionally consistent . Differences in identified SNPs between our study and the Li et al . study may be attributed to , among other things , a difference in the outcome measured ( T2D vs glycemic traits ) . Indeed , the genetic determinants of glycemic traits ( i . e . normal glycemic variation ) has been shown to differ somewhat from the genetic determinants of T2D [13] . We find that rs964184 near APOA1 was most strongly negatively associated with T2D , according to the p-value . To our knowledge there is no previous report of an APOA1 variant being negatively associated with T2D or related trait . The variation near the APOA1 gene is thought to increase TG levels through impairment of the capacity of APOA1 to bind with lipoprotein lipase ( LPL , the enzyme responsible for lipolysis ) , resulting in decreased LPL-mediated lipolysis of TG [14 , 15] . In the DIAGRAM meta-analysis , the G allele ( TG increasing allele ) at this SNP is non-significantly ( and positively ) , associated with a greater risk of T2D ( p = 0 . 15 ) [16] . We suspect that this non-significant result ( and positive direction ) is due to the fact that TG is typically not included as a covariate in GWAS of T2D . Previous studies have shown that variants in GCKR may have pleiotropic effects of increasing TG and decreasing insulin resistance [7 , 8 , 17–19] . We confirm this in our study . However , our results suggest that GCKR is just one of several drivers of the pleiotropic effect of TG genes on T2D risk . A recent functional analysis of GCKR variant rs1260326 found that it is highly expressed in hepatic cells in response to glucokinase ( GCK ) activity , as compared to pancreatic islet cells , indicating its functional role inside liver cells [10 , 20] . Furthermore , the researchers found the GCK-inhibiting activity , mediated through fructose-6 phosphate , to be significantly attenuated among those with the rs1260326 variant , thus potentially enhancing glucose uptake in the liver and further increasing levels of TG precursor substrate , malonyl-CoA [10] . This influx of glucose inside the liver and resulting increase in the synthesis of malonyl-CoA may account for low glucose and high TG levels . A previous study has also implicated the CILP2 locus as having opposite putative effects on TG and T2D [21] . The SNP discovery efforts of large GWAS meta-analyses could have down-prioritized SNPs involved in lipid variation among individuals with T2D , if T2D cases were excluded from the analysis . This could potentially have the effect of only discovering SNPs involved in lipid variation among individuals who are non-diabetic . Although T2D cases were excluded from the ARIC study in the Teslovich et al . meta-analysis , for most other studies , T2D cases were included in the analysis , and analyzed separately from controls [3] . These findings have implications for how we understand the causal relationship between TG and T2D . They suggest that additional studies are needed to closely examine the biological and causal links connecting lipid and glycemic phenotypes . For example , several SNPs associated with FI have been found to be associated with higher TG [13 , 22] providing support for a causal pathway in which a T2D-related phenotype causes elevated TG . In conclusion , further research is needed to understand the molecular and physiological mechanisms underlying the putative pleiotropic nature of TG-associated genes .
We used phenotypic and genotypic data of European- and African-Americans ( EA , AA ) from three prospective cohort studies conducted in the United States for a combined sample size of 13 , 285 . We used data from 7 , 868 EA and 2 , 089 AA participants from the Atherosclerosis Risk in Communities ( ARIC ) study , which is a multi-center prospective study of men and women between the ages of 45 and 64 to investigate risk factors associated with atherosclerosis [23] . We used data from 3 , 430 EA participants from the Framingham Heart Study Offspring Study ( FHS ) , which is a prospective cohort study to examine the causes of heart disease [24] . Finally , we used data from 1 , 949 EA and 1 , 149 AA participants in the Multi-Ethnic Study of Atherosclerosis ( MESA ) , a prospective cohort study of the risk factors for atherosclerosis among middle age men and women aged 45–85 years [25] . This study was approved by the University of Arizona Human Subjects Protection Program ( Protocol number: 1300000659R001 ) . No patient consent was given as the data were analyzed anonymously . . Data was obtained from the database of Genotypes and Phenotypes ( dbGaP ) . Individuals with prevalent T2D at the baseline examination were excluded from this study . Prevalent T2D was defined as having a FG level > 125 mg/dL , a report of taking T2D medication , or a physician diagnosis . We also excluded individuals who reported using cholesterol medications , including statins , as these medications can influence blood lipids levels and other blood constituents , which may artificially reduce associations between TG and incident T2D . We also excluded individuals who had not fasted ( < 8 hours ) at the baseline exam . The samples drawn at the first visit of each study were processed and analyzed using standardized procedures and protocols , the details of which are described elsewhere [26–28] . We used baseline measurements of age , TG , HDL-C , LDL-C , BMI , FG and FI as covariates in the analysis . Height and weight were also measured at baseline , and body mass index ( BMI; kg/m2 ) was calculated . Incident T2D cases were identified at one of three follow-up visits in ARIC , seven follow-up visits in FHS , and four follow-up visits in MESA , based on FG , medication criteria , or physician diagnosis . In ARIC , time to incident type 2 diabetes was extrapolated based upon glucose values at the ascertaining visit and the previous visit , as previously described [29] . FHS participants were genotyped with the Affymetrix 500K SNP Array ( Affymetrix , Santa Clara , CA , USA ) . ARIC and MESA participants were genotyped with the Affymetrix Genome-Wide Human SNP Array 6 . 0 . Standard quality-control measures were employed prior to imputation . We used IMPUTE2 with the 1 , 000 Genomes data as a reference , to impute millions of genotypes in each study [30] . To assess genetic risk for elevated TG , we calculated two GRS . The first GRS was based on 31 single nucleotide polymorphisms ( SNPs ) identified in a large-scale meta-analysis of lipid levels [3] . The second GRS was based on 40 SNPs identified in a more recent meta-analysis of lipid levels [4] . All SNPs in the three studies had imputation quality scores ( ‘info’ ) > 0 . 6 . Weighted genetic risk scores were calculated by taking the sum of risk alleles for all SNPs and weighting each risk allele by its respective effect size . Risk alleles and effect sizes were defined according to the findings in the respective meta-analysis . Since TG , HDL-C , LDL-C , FG and FI levels were measured using different units in each study , we standardized them to a mean of 0 and standard deviation of 1 in each study so that we could combine the studies . Hazard ratios ( HR ) , defined as the ratio of hazard rates corresponding to different levels of TG genetic risk , were estimated using Cox proportional hazards regression models in each of the three studies and in the combined data . We considered the association of TG genes with incident T2D by conditioning on variables including age , sex , BMI , HDL-C , LDL-C , FG , FI , and TG . As mentioned above , phenotypic covariates were measured at the baseline examination in each respective study . Interactions were modeled as the product of TG and the TG GRS ( or TG SNP ) and included as covariates in the Cox proportional hazards regression model . We also tested a model in which additional interaction terms of both the TG and TG GRS with all other covariates were added to the model , based upon the recommendation of Keller [31] . We combined the studies into one ( one for EA and one for AA ) to conduct analyses , and accounted for the potential within-study dependence for the subjects from the same study by including a ‘frailty’ term for the study in the model . We also tested proportionality of hazard over time by including a time-dependent covariate consisting of the interaction of the logarithm of the time to event with TG and the TG GRS , and their interaction term . Among EA , there was not sufficient evidence ( p>0 . 05 ) to reject the null hypothesis of hazard proportionality over time in the TG GRS association or the TG-by-GRS interactions . Among AA , we found the same , except for the TG variable ( p = 0 . 04 ) . However , upon visual inspection and comparison of the Kaplan-Meier survival curves , we observe that the TG >-0 . 22 ( median ) and < = -0 . 22 curves look proportional . We suspect that the p-value<0 . 05 is attributed to the large sample size . For the SNP analysis , we considered a Bonferroni correction for 31 tests performed , resulting in an alpha of 0 . 0016 . To test the association of each TG GRS with TG , we first natural-log ( ln ) transformed TG separately in each study , then standardized it as described above . We also standardized each TG GRS in order to compare the association of each with TG . We used linear regression to test the association of each TG GRS with standardized ln ( TG ) in the set of combined studies , including age , sex , BMI and the random effect of study as covariates . All analyses were conducted with R Statistical Software [32] and SAS Software ( Cary , NC ) . Data for this studies was obtained from dbGaP through accession numbers: phs000007 . v23 . p8 , phs000280 . v2 . p1 , phs000209 . v10 . p2 .
|
An elevated triglyceride level is generally considered a risk factor for the development of type-2 diabetes . However , recent studies suggest , somewhat paradoxically , that genetic risk for elevated triglycerides may protect against type-2 diabetes . In this study , we test the relationship of triglyceride-associated genetic variants , collectively and individually , with incident type-2 diabetes across three prospective cohort studies comprised of European- and African-American participants . Our findings across studies , racial groups , and statistical models consistently demonstrate that triglyceride-increasing alleles are associated with decreased type-2 diabetes incidence . These genes therefore appear to both increase triglyceride levels and decrease type-2 diabetes risk . More work is needed to understand the physiological mechanism underlying these findings , and to determine the causal relationship between triglycerides and type-2 diabetes .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[] |
2015
|
Triglyceride-Increasing Alleles Associated with Protection against Type-2 Diabetes
|
Hereditary hearing loss is a clinically and genetically heterogeneous disorder . More than 80 genes have been implicated to date , and with the advent of targeted genomic enrichment and massively parallel sequencing ( TGE+MPS ) the rate of novel deafness-gene identification has accelerated . Here we report a family segregating post-lingual progressive autosomal dominant non-syndromic hearing loss ( ADNSHL ) . After first excluding plausible variants in known deafness-causing genes using TGE+MPS , we completed whole exome sequencing in three hearing-impaired family members . Only a single variant , p . Arg185Pro in HOMER2 , segregated with the hearing-loss phenotype in the extended family . This amino acid change alters a highly conserved residue in the coiled-coil domain of HOMER2 that is essential for protein multimerization and the HOMER2-CDC42 interaction . As a scaffolding protein , HOMER2 is involved in intracellular calcium homeostasis and cytoskeletal organization . Consistent with this function , we found robust expression in stereocilia of hair cells in the murine inner ear and observed that over-expression of mutant p . Pro185 HOMER2 mRNA causes anatomical changes of the inner ear and neuromasts in zebrafish embryos . Furthermore , mouse mutants homozygous for the targeted deletion of Homer2 present with early-onset rapidly progressive hearing loss . These data provide compelling evidence that HOMER2 is required for normal hearing and that its sequence alteration in humans leads to ADNSHL through a dominant-negative mode of action .
Targeted genomic enrichment and massively parallel sequencing ( TGE+MPS ) have revolutionized the field of hereditary hearing loss ( HHL ) by making comprehensive genetic testing for deafness a clinical reality and by accelerating the discovery of novel deafness-causing genes [1 , 2] . To date over 80 genes have been causally implicated in non-syndromic hearing loss ( NSHL; Hereditary Hearing Loss Homepage ) . The proteins encoded by these genes are involved in a broad array of molecular and cellular mechanisms essential for the development and maintenance of normal auditory function [3] . At the centerpiece of this intricate system are the outer and inner hair cells ( OHCs , IHCs ) —key structures in the mechanotransduction process by which mechanical stimuli are translated into electrical impulses [4] . The precise and efficient tuning of OHCs and IHCs is closely linked to their anatomical integrity and the coordinated movement of their apical stereocilia . The Homer proteins are scaffolding proteins crucial to many intracellular signaling cascades; their function underpins a variety of neuronal processes ranging from calcium homeostasis and cytoskeletal organization to synaptic plasticity associated with learning and memory [5 , 6] . Homer proteins are encoded by three genes; HOMER1 , 2 , and 3 ( MIM 604798 , MIM 604799 and MIM 604800 , respectively ) that are translated into multiple isoforms as a result of alternative splicing [7] . All isoforms share an N-terminal conserved EVH1 ( Ena VASP Homology 1 ) domain , which binds proline-rich regions of target proteins . The long isoforms ( HOMER1b and c , HOMER2 , HOMER3 ) additionally have a coiled-coil ( CC ) region and leucine zipper motifs in their divergent C-termini [8] . The CC region is required for homo/hetero-multimerization to form tetrameric hubs ( in which the CC domains align in a parallel fashion ) and for interaction with Rho family GTPase proteins like CDC42 ( MIM 116952 ) [9–11] . Although the short isoforms lack CC domains and therefore do not form multimers , like their longer counterparts they bind target proteins through their EVH1 domain . This interaction regulates the activity of proteins involved in Ca2+ signaling complexes including metabotropic glutamate receptors ( mGluR ) [12] , inositol tri-phosphate receptors ( IP3R ) [8] and transient receptor potential canonical channels ( TRPC ) [13] . Homer proteins also regulate basal cytosolic calcium via an interaction with the plasma membrane calcium reuptake pump , PMCA [14 , 15] . Aberrant Homer signaling has been associated with several developmentally related neurological syndromes including Fragile X syndrome , epilepsy , addiction , schizophrenia , neuropathic pain and Alzheimer’s disease [16–19] . To this list we now add autosomal dominant NSHL ( ADNSHL ) .
The family segregating the deafness-causing mutation in HOMER2 ( MIM 604799; RefSeq NM_004839 ) is a multi-generational kindred of European descent ( Fig 1A ) . Pure tone audiometric evaluation of affected members showed bilateral post-lingual progressive hearing loss that segregated as an autosomal dominant trait; bone conduction thresholds excluded conductive hearing impairment . Clinical examination was negative for any findings consistent with syndromic hearing loss and also ruled out autoimmune phenotypes . Hearing impairment had a typical onset in the first decade of life in the high frequencies , with significant subsequent progression of hearing loss over all frequencies . To evaluate progression at each frequency , we performed linear regression analyses of threshold on age [20] . The resulting annual threshold deterioration ( ATD ) was 1 . 2 to 1 . 6 dB per year ( Fig 1B , S1 Fig ) . The age-related typical audiogram ( ARTA ) derived from these data confirmed the down-sloping audiometric configuration and demonstrated fairly similar progression across all frequencies . Our initial strategy was to screen one family member ( III . 10 ) for pathogenic variants in known deafness-causing genes using a deafness-specific TGE+MPS panel ( OtoSCOPE ) [21] . Plausible pathogenic variants were excluded ( S1 Table ) . Whole exome sequencing ( WES ) was therefore completed on three affected family members ( II . 2 , IV . 5 , and IV . 10 ) and variants were filtered according to guidelines detailed in Materials and Methods section [22] . The resultant final candidate variant lists included 163 , 150 and 158 nucleotide changes for these three individuals , respectively ( S1 Table ) . Of these variants , only four were shared amongst the three sequenced exomes ( S2 Table ) and only one segregated with the ADNSHL phenotype in the extended family ( Fig 1A and 1C ) . The segregating variant , c . 554G>C; p . Arg185Pro in HOMER2 on chromosome 15q25 . 2 , is a novel non-synonymous change that substitutes a highly conserved arginine for a proline , a substitution that is predicted to be pathogenic and disease-causing by Polyphen2 , LRT , SIFT and MutationTaster ( S3 Table ) . HOMER2 belongs to the homer family of post-synaptic density scaffolding proteins and is expressed as two isoforms , HOMER2 isoform 1 ( NM_004839 , 343 aa ) and isoform 2 ( NM_199330 . 2 , 354 aa ) , which differ by 11 amino acids . The p . Arg185Pro variant lies in the CC domain ( Fig 1D ) . A lysine is found at the orthologous position in HOMER1 and HOMER3 , which like arginine is a basic polar amino acid that confers similar chemico-physical properties to the protein . Proline , in contrast , is non-polar and is predicted to alter the conformational structure of the CC domain or affect its ability to multimerize and/or interact with partner proteins . Homer2 is extensively expressed in the CNS throughout development [23] . It is also expressed in skeletal muscle , heart , liver , spleen , lung and kidney . To investigate and define its expression pattern in inner ear , we immuno-labeled whole mount P2 mouse cochlea with HOMER2 antibody . Organ of Corti expression was particularly enriched in the tips of stereocilia of both IHCs and OHCs ( Fig 2 ) , consistent with data by Hertzano and colleagues using cell sorting and RNASeq to identify Homer2 enrichment in the sensory cells of P0–P1 mice [24] . This expression pattern suggests a role for HOMER2 in hair bundle function , formation , development or maintenance . To evaluate whether the p . Arg185Pro mutation affects subcellular localization , we transfected HEK293T and COS7 cells with cMYC-tagged HOMER2WT proteins or FLAG-tagged HOMER2p . Arg185Pro . Both wild-type ( WT ) and mutant proteins distributed in a diffuse manner in the cytoplasm with no obvious differences in localization patterns ( S2 Fig ) . A similar pattern of distribution was observed when both WT and mutant constructs were cotransfected . These data indicate that the p . Arg185Pro mutation does not alter subcellular localization of HOMER2 . To assess the impact of the p . Arg185Pro mutation in HOMER2 in vivo we used the zebrafish model . Zebrafish hair cells share similarities with their mammalian counterparts in morphology and function . They reside inside the otic vesicles and in the neuromasts of the lateral line system , a sensory system on the surface of fish important for sensing propulsion through water , capturing prey , or avoiding predators and obstacles . The zebrafish homer2 ( NP_001018470 . 1 ) is 67% identical to human HOMER2 and is expressed mostly in the developing musculature although there is faint expression in the otic capsule at 24 hours post-fertilization ( hpf ) [25] . We used morpholino antisense oligonucleotides ( S4 Table ) to induce altered splicing and protein truncation . Although knockdown of homer2 altered neither ear size nor morphology ( p>0 . 5 ) ( S3 Fig ) , injection of in vitro synthesized mRNA encoding HOMER2 P185-mutant RNA ( P185RNA ) resulted in significantly smaller ear size in larvae as compared to injections with WT HOMER2 ( wtRNA ) ( p<0 . 001 ) ( Fig 3A i-iii and 3B; S4 Fig ) . In addition , the number of kinocilia detected per neuromast was decreased ( p = 0 . 03 ) ( Fig 3C i-iii and 3D ) . We also noted that P185RNA-injected embryos exhibited shorter kinocilia , an observation that needs to be confirmed with more detailed methods . These results show that HOMER2 plays an essential role in the normal development and/or maintenance of hair cells in the zebrafish inner ear and that the p . Arg185Pro mutation has a dominant-negative effect on this process . Dominant-negative activity has been previously validated in vivo for another homer protein , HOMER1 . Its long isoforms , HOMER1b and 1c , are constitutively expressed while the short isoform , HOMER1a is an immediate early gene product that is rapidly and transiently induced by high synaptic activity [26] . Because HOMER1a lacks the CC domain , it cannot form multimers but it can still competitively bind the target proteins of HOMER1b and 1c , suggesting that relative expression of each of these homer proteins is critical [27–29] . The importance of this dynamic balance has been validated by the observation that overexpression of Homer1a in a mouse impedes normal development and through a dominant-negative effect leads to significant defects in motor coordination and learning , with increased levels of fear-associated behavior and anxiety [30] . The role of Homer2 has been thoroughly studied in the murine brain and pancreas where it functions to decrease the intensity of Ca2+ signaling by reducing signaling by G protein-coupled receptors ( GPCRs ) [31] . Mice homozygous for the targeted deletion of homer2 ( Homer2-/- mice generated by deletion of exon 3 [32] ) show extensive behavioral and neurochemical similarities to cocaine or alcohol-sensitized animals , and demonstrate Homer2 involvement in appetitive pathways underlying responses to those drugs and their induced behavioral/cellular neuroplasticity within the nucleus accumbens [33 , 34] . Studies in Homer2-/- Homer3-/- mice show upregulation of cytokine expression and an increase in effector-memory T cells leading to an autoimmune-like pathology , indicating that Homer2 negatively regulates T cell activation [35] . To evaluate cochlear function of Homer2-/- mice , we measured auditory brainstem responses ( ABR ) , an electrophysiological hearing test that reflects the activity of afferent auditory neurons downstream of IHCs . We examined hearing in 2- , 4- and 8-week-old Homer2-/- , Homer2+/- and WT Homer2+/+ mice and observed no differences between age-matched WT and Homer2+/- animals ( P>0 . 2 ) , however Homer2-/- mice showed progressive hearing loss ( Fig 4 ) . At P14 Homer2-/- mice had slightly elevated broad band click ABR thresholds as compared to Homer2+/- mice [Homer2-/- ( n = 11 ) ; 67 . 3±2 . 56 verse Homer2+/- ( n = 17 ) ; 61 . 20±2 . 50 db SPL] . This difference progressively increased ( [P28 Homer2-/- ( n = 10 ) ; 82 . 0±5 . 40 db SPL vs Homer2+/- ( n = 6 ) ; 59 . 20±2 . 71 db SPL]; [P56 Homer2-/- ( n = 5 ) ; 100 . 0±0 . 0 db SPL vs Homer2+/- ( n = 9 ) ; 55 . 6±1 . 76 db SPL] ) ( Fig 4A ) . Tone bursts showed that P14 Homer2-/- mice had profound deafness at 32 kHz and that the rate of hearing deterioration at 8 kHz was dramatic ( P14 , 55 . 5±2 . 28 dbSPL; P28 , 73 . 0±6 . 84; P56 , 87 . 0±1 . 22 dbSPL ) ( Fig 4B-4D ) . We assessed OHC function using distortion product otoacoustic emissions ( DPOAEs ) and observed no significant differences in DPOAE thresholds between Homer2+/- and WT mice ( P>0 . 05 ) . In Homer2-/- mice , significant decreases in mid- to high-frequency DPOAE levels were seen at P14 and P28 that culminated in profound deafness at all frequencies by P56 ( Fig 4E-4G ) . To determine whether the auditory deficit in Homer2-/- mice was secondary to hair cell loss we analyzed whole mount preparations of the organ of Corti at P56 . No differences in IHCs and OHCs were observed in any animals regardless of genotype indicating the absence of Homer2 does not impair hair cell formation and development , and that the hearing loss in Homer2-/- mice is not due to hair cell death ( Fig 4H ) . Whether the hearing deficit in these mice is due to abnormal hair bundle morphology remains to be elucidated . We additionally investigated spiral ganglion morphology and found no obvious indication of spiral ganglion degeneration in Homer2-/- mice as compared to their WT littermates ( S5 Fig ) . In aggregate , our data show that HOMER2 is required for normal hearing . Its complete absence in mice leads to early onset progressive hearing loss starting at the high frequencies and rapidly involving all frequencies . The recessive phenotype exhibited by null alleles of Homer2 makes it a strong candidate for autosomal recessive hearing loss due to loss of function in humans as well . The absence of a disease phenotype in Homer2+/- mice suggests that haploinsufficiency does not cause hearing loss . Together with results obtained from zebrafish experiments , these data strongly suggest that the p . Arg185Pro mutation in HOMER2 exerts its effect through a dominant-negative mechanism on wild-type protein by either inhibiting multimerization or competing for other partner proteins . In defining the effect of the p . Arg185Pro mutation in HOMER2 at a molecular level , we believe two hypotheses warrant consideration . One hypothesis posits that HOMER2 exerts its function by regulating actin dynamics in stereocilia through its interaction with CDC42 , a highly conserved small GTPase of the RHO family that fine-tunes actin-turnover ( S6 Fig ) . HOMER2 is known to couple with and regulate CDC42 through its CDC42-binding domain ( CBD ) within the CC domain ( Fig 1D ) [10] . In HeLa cells , CDC42 induces the formation of filopodia-like protrusions , while overexpression of HOMER2 suppresses this phenotype [36] . In the cochlea , CDC42 localizes to stereocilia membranes and its targeted deletion in murine HCs leads to their degeneration and results in progressive hearing loss particularly at the high frequencies , a phenotype similar to the human HOMER2 mutant phenotype and the Homer2-/- murine phenotype [36] . A second hypothesis focuses on the role of HOMER2 in cytoplasmic Ca2+ control . Several studies have shown that HOMER2 regulates a number of Ca2+ handling proteins including TRPC and PMCA channels . Two TRPCs—TRPC3 ( MIM 602345 ) and TRPC6 ( MIM 603652 ) —are expressed in both sensory neurons and cochlear hair cells and are required for normal function . Their targeted deletion in mice causes significant dysregulation of Ca2+ re-entry that leads to hearing impairment and vestibular deficits [37] . While both proteins are potentially interacting partners for HOMER2 , to date only an interaction with TRPC1 ( MIM 602343 ) has been established [13] . However , a recent study has demonstrated a critical role for Homer2 in modulating PMCA activity by regulating the duration of the Ca2+ signaling in parotid acinar cells [38] . This finding suggests a possible role for Homer2 in cytosolic Ca2+ clearance to balance TRPC-mediated Ca2+ influx and PMCA-mediated Ca2+ extrusion . A suitable interacting partner of HOMER2 may be the PMCA2 pump ( MIM 108733 ) , which represents the only system for clearance of Ca2+ from hair cell stereocilia [39] . While both of these hypotheses are attractive , further functional studies are needed to identify the partner proteins of HOMER2 in inner ear and investigate the effect of the p . Arg185Pro mutation on these interactions . In summary , we have identified HOMER2 as essential to normal auditory function and have shown that the p . Arg185Pro HOMER2 mutation causes ADNSHL through a dominant-negative mechanism of action , thus expanding the phenotypic spectrum associated with Homer protein dysfunction .
A five-generation family of European descent segregating bilateral post-lingual progressive ADNSHL was ascertained for this study ( Fig 1A ) . After obtaining written informed consent from all participants with approval by the Institutional Review Board of the University of Iowa , clinical examination of the subjects was completed to exclude any additional and/or syndromic findings . Blood samples were obtained from 19 family members . Pure tone audiometry was performed according to current standards to determine air conduction thresholds at 0 . 25 , 0 . 5 , 1 , 2 , 3 , 4 , 6 and 8 kHz . Bone conduction thresholds were determined at 0 . 5 , 1 , 2 and 4 kHz in some patient to exclude conductive hearing impairment . After validating binaural symmetry , the binaural mean air conduction threshold ( dB Hearing Level , HL ) at each frequency was used for further analyses . An arbitrary value of 130 dB HL was used to indicate out-of-scale measurements . Linear regression analyses of threshold on age were used to evaluate progression of hearing impairment at individual frequencies . These analyses comprised both individual longitudinal data derived from serial audiograms and overall cross-sectional last-visit data . Progression was considered significant if the 95% confidence interval for slope did not include zero for two or more frequencies . Progression was expressed in dB-per-year and designated Annual Threshold Deterioration ( ATD ) . Cross-sectional regression data conformed to individual longitudinal regression data . Regression data from the last-visit thresholds were used to derive Age-Related Typical Audiograms ( ARTA ) , which show expected thresholds by decade steps in age [20] . OtoSCOPE v1 was used to evaluate all known genetic causes of NSHL ( including the non-syndromic mimics like Usher Syndrome ) in one affected individual ( III . 10 ) , as previously described [21 , 40] . Whole exome capture was performed with the Agilent SureSelectXT Human All Exon V4 ( Agilent Technologies , Santa Clara , CA ) according to the manufacturer’s protocol . All enriched libraries were sequenced on the Illumina HiSeq 2000 ( Illumina , Inc . , San Diego , CA ) using 100bp paired-end reads . Data analysis was performed on a local installation of Galaxy using the Burrows-Wheeler Alignment ( BWA ) for read mapping to the reference genome ( hg19 , NCBI Build 37 ) , Picard for removal of duplicate reads , GATK for local re-alignment and variant calling , and ANNOVAR and a custom workflow for variant annotation . Variant filtering was based on: quality ( >10X ) ; minor allele frequency ( MAF<0 . 0005 ) as reported in the 1000 Genomes Project database and the National Heart , Lung , and Blood Institute ( NHLBI ) Exome Sequencing Project Exome Variant Server ( EVS ) . Variants were annotated for conservation ( GERP and PhyloP ) and predicted pathogenicity ( PolyPhen2 , SIFT , MutationTaster and LRT ) . Variants were then filtered based on coding effect ( non-synonymous , indels and splice-site variants ) ; heterozygosity and allele sharing amongst the three sequenced affected individuals ( II . 2 , IV . 5 , and IV . 10 ) . Sanger sequencing was completed in all family members to confirm segregation of c . 554G>C; p . Arg185Pro in HOMER2 gene ( MIM 604799; RefSeq NM_004839 ) using primers HOMER2-6F: 5’-ATGGGAGAGGCAGCAAGTCT-3’ and HOMER2-6R: 5’-AGACCCACCTGCCAGCTTAC-3’ . Cochleae from Balb/c mice were harvested at P2 , locally perfused , fixed in 4% paraformaldehyde for 30min , and rinsed in PBS . Tissues were microdissected into cochlear and saccule subsets and stored at 4°C in preparation for immunohistochemistry . Following infiltration using 0 . 3% Triton X-100 and blocking with 5% normal goat serum , we incubated the tissues in rabbit HOMER2 polyclonal primary antibody ( NB100-98712 , Novus Biologicals , Littleton , CO ) diluted 1:1000 in PBS overnight at 4°C . Specificity of HOMER2 antibody was confirmed by staining whole mount cochlea from Homer2-/- mice ( S7 Fig ) . Subsequently , a secondary antibody Alexa-Fluor 568 Goat anti-rabbit ( Life Technologies , Carlsbad , CA , USA; 1:1000 ) was applied for 1h . Alexa-Fluor 488 phalloidin ( Life Technologies , Carlsbad , CA , USA; 1:500 ) was added for 15min to selectively visualize F-actin . We used anti-neurofilament 200 monoclonal primary antibody ( N0142 , Sigma-Aldrich , Saint Louis , MO ) and Alexa-Fluor 488 Goat anti-mouse as a secondary antibody to visualize spiral ganglions neurons . Whole-mount tissues were mounted in ProLong Gold Antifade Reagent ( Life Technologies , Carlsbad , CA , USA ) . Confocal images were collected using Leica TCS SP5 confocal microscope ( Leica Microsystems Inc . , Bannockburn , IL , USA ) and analyzed in LSM 5 Image Browser and Adobe Photoshop . Transfected HEK293T and COS7 cells were fixed in 4% paraformaldehyde in 0 . 1 M PBS ( pH 7 . 4 ) ; cells were permeabilized with 0 . 1% Trition-X100 . Fixed cells were incubated with primary antibody at room temperature in PBS for 1 . 5hrs . The following primary antibodies were used: monoclonal Anti-FLAG ( Sigma-Aldrich , St . Louis , MO , USA; 1:400 ) and Anti-cMYC ( Sigma-Aldrich , St . Louis , MO , USA; 1:400 ) . Secondary antibody incubation was for 1hr at room temperature . Secondary antibodies used: Alexa-Fluor-488 goat anti-mouse ( Invitrogen , Grand Island , NY , USA; 1:500 ) to stain FLAG-tagged HOMER2 p . Arg185Pro and Alexa-Fluor-568 goat anti-rabbit ( Invitrogen , Grand Island , NY , USA; 1:500 ) to stain cMYC-tagged HOMER2WT . F-actin was immuno-stained with Alexa-Fluor-647-phalloidin ( Invitrogen , Grand Island , NY , USA;1:500 ) . Cells were mounted in SlowFade Gold Antifade Reagent with DAPI ( Life Technologies , Grand Island , NY , USA ) . Images were taken using a Zeiss LSM 510 with ZEN 2009 confocal microscope ( Zeiss , Pleasanton , CA , USA ) . The Gateway PLUS shuttle clone for HOMER2 ( AF081530 . 1 ) was ordered from GeneCopoeia ( GeneCopoeia Inc , Rockville , MD , USA ) . QuickChange Site-Directed Mutagenesis Kit ( Stratagene , Cambridge , UK ) was used for site-specific mutagenesis to introduce the P185 mutation ( S4 Table ) . The mutant expression plasmid was sequence verified . Both full length open reading frames for WT protein HOMER2WT and mutant HOMER2p . Arg185Pro were cloned into the expression vectors pCMV-Tag3 ( cMYC-tagged ) and pCMV-Tag2 ( FLAG-tagged ) , respectively ( Agilent technologies , Santa Clara , CA , USA ) . All constructs were verified by sequence analysis . HEK293T cells and COS7 cells ( ATCC , Manassas , VA , USA ) were grown in Dulbecco's Modified Eagle's Medium ( DMEM ) supplemented with 10% FBS ( Life Technologies , Carlsbad , CA , USA ) . Cells were incubated in a 5% CO2-humidified incubator at 37°C . Cells were grown on Poly-L-Lysine Coated coverslips ( Corning , Tewksbury , MA , USA ) . Clonal cells were obtained by transfection with Transit-LT Transfection Reagent ( Mirus Bio , Madison , WI USA ) using cMYC-tagged HOMER2WT and FLAG-tagged HOMER2p . Arg185Pro plasmid constructs according to the manufacturer’s instructions . Zebrafish embryos were raised at 28 . 5°C as described [41] . All animal procedures were approved by the University of Iowa Office of Animal Resources ( OAR ) principle for the care and use of laboratory animals and the Institutional Animal Care and Use Committee ( IACUC ) . Antisense morpholino oligonucleotides ( MOs ) were designed to block the exon/intron splice junctions between exon 1 and intron 1 ( MO i1e1 5′- GGTACACATGTATCTGTCTGACCTT-3′ ) or intron 3 and exon 4 ( 5′-CGCAATGAAAACTGTAAACACTCTT-3′ ) of homer2 ( ENSDART00000124088 ) and bought from Gene Tools ( Philomath , OR , USA ) . They were injected at 2 . 2 mg/ml along with 1 mg/ml p53 MO ( 5’-GCGCCATTGCTTTGCAAGAATTG-3’ ) . A standard control MO ( 5’-CCTCTTACCTCAGTTACAATTTATA-3’ ) was used for injection of negative controls along with p53 MO . The efficacy of homer2 knockdown by each morpholino was assessed by RT-PCR analysis with the following primer sets: forward ( 5’- GGTTCCCGCCAGTAAACAG-3’ ) and reverse ( 5’-GTTTGAGCTCCGTCTTCAGG-3’ ) , which amplified the region between exons 1 and 12; B-actin primers were forward ( 5′-GAGATGATGCCCCTCGTG-3' ) and reverse 5'-GCTCAATGGGGTATTTGAGG-3' ) . MO i1e1 morpholino was used for all subsequent experiments . For in vivo mRNA synthesis , HOMER2WT and HOMER2p . Arg185Pro plasmids were transferred into the expression vector pCS2+ with Gateway LR Clonase according to the manufacturer’s instructions ( Life Technologies , Carlsbad , CA , USA ) . This cDNA was used as a template for HOMER2 capped mRNA synthesis using an Ambion mMESSAGE mMACHINE SP6 kit ( Applied Biosystems , Foster City , CA , USA ) , and the product was tested for quality and yield by electrophoresis and spectroscopy ( NanoDrop Thermo Scientific , Waltham , MA , USA ) before injection . Microinjection was performed at the one- to two-cell stage using a microinjection system consisting of a SZX9 stereomicroscope ( Olympus , Tokyo , Japan ) and an IM300 Microinjector ( Narishige , Tokyo , Japan ) . Overexpression of injected mRNA was assessed by quantitative PCR with the following primer sets: forward ( 5’-GACCCCAACACCAAGAAGAA-3’ ) and reverse ( 5’CACTGTGTTGGCTCTGCTGT-3’ ) . Primers for B-actin were forward ( 5’-CGCGCAGGAGATGGGAACC-3’ ) and reverse ( 5’-CAACGGAAACGCTCATTGC-3’ ) . At 72 hours post fertilization ( hpf ) , live larvae were submerged in 3 μM FM1-43 FX ( Invitrogen , Grand Island , NY , USA ) for 30 sec . They were then rinsed X3 in embryo media ( ddH2O with 5 . 03 mM NaCl , 0 . 17 mM KCl , 0 . 33 mM CaCl2 , 0 . 33 mM MgSO4 , 0 . 1% w/v methylene blue ) and fixed in 4% paraformaldehyde . Before viewing , fish were rinsed X3 in PBS and viewed in 75% glycerol , 25% PBS with a Zeiss 700 confocal microscope . We focused on neuromasts that reside around the surface of the otic vesicle: o1 , ml1 , ml2 , o2 and io4 . Z-stacks were prepared using the max intensity z-projection function in ImageJ ( NIH , Stapleton , New York City , USA ) . The morphous and structure of otic vesicles was observed in live larvae at 72hpf with a Leica MZFIII3 light microscope after anesthetizing with Tricaine . Images were analyzed with ImageJ ( NIH ) . Mouse studies were carried out in accordance with University of Iowa Office of Animal Resources ( OAR ) principle for the care and use of laboratory animals and the Institutional Animal Care and Use Committee ( IACUC ) . Mice were culled using methods approved by the American Veterinary Medical Association ( AVMA ) Guidelines for the Euthanasia of Animals . The knockout Homer2-/- mice were donated by Paul F . Worley at John Hopkins University ( Baltimore , Maryland , USA ) . These mice have a neomycin cassette inserted into exon 3 of Homer2 abolishing gene expression [31] . The Homer2 -/- colony was maintained on a C57BL/6J background . Mice were genotyped by PCR , as previously described ( S8 Fig and S4 Table ) [31] . Hearing thresholds were measured by click and tone burst ( 8 , 16 , and 32 kHz ) ABR and DPOAE in Homer2-/- , Homer2+/- and WT mice at two ( P14 ) , four ( P28 ) and eight ( P56 ) weeks . At least 23 animals were tested at each time point . Mice were anesthetized using intraperitoneal Ketamine/Xylazine at 0 . 1ml/10g body weight . Reference , ground and earth electrodes were placed subcutaneously just posterior to the tested ear ( left ear ) , anterior to the contralateral ear and at the vertex of the head , respectively . ABRs were performed using an experimental setup and testing protocol , as described [42] . Briefly , clicks and tone-bursts were delivered to the testing ear through a plastic acoustic tube . ABRs were measured using an Etymotic Research ER10B+ probe microphone ( Etymotic Research , Elk Grove , IL , USA ) coupled to two Tucker-Davis Technologies MF1 multi-field magnetic speakers ( Tucker-Davis Technologies , Alachua , FL , USA ) . Click and tone-burst stimuli were presented and recorded using custom software running on a PC connected to a 24-bit external sound card ( Motu UltraLite mk3 , Cambridge MA , USA ) . A custom-built differential amplifier with a gain of 1 , 000 dB amplified acoustic ABR responses . Output was passed through 6-pole Butterworth high-pass ( 100 Hz ) and low-pass ( 3 kHz ) filters and then to a 16-bit analog-to-digital converter ( 100 , 000 sample/s ) . Responses were recorded using standard signal-averaging techniques for 500 or 1000 sweeps . Hearing thresholds ( db SPL ) were determined by decreasing the sound intensity by 5 and/or 10 db decrements and recording the lowest sound intensity level resulting in a recognizable and reproducible ABR response wave pattern . Maximum ABR thresholds were capped at 100 db SPL . DPOAEs were measured unilaterally ( left ear ) using an experimental setup and testing protocol , as described [42] . In brief , DPOAE levels were elicited using two primary tone stimuli , f1 and f2 , with sound pressure levels of 65 and 55 db SPL , respectively , with f2/f1 = 1 . 22 . A custom plastic ear probe was inserted into the ear canal and DPOAE amplitudes were measured at f2 frequencies at 4000 , 5657 , 8000 , 11314 , 16000 , 22627 and 32000 Hz and plotted after subtraction of noise floor amplitude . IPA ( Ingenuity Systems , Mountain View , CA , USA ) was used to map interactions between genes involved in NSHHL ( 86 genes ) and HOMER2 and CDC42 . Networks were created from user-specified seed molecules by searching the knowledge base for molecules that are known to biologically interact with the seeds and connecting them . Networks are displayed graphically as nodes ( genes/gene products ) and edges ( the biological relationships between the nodes ) . IPA computes a score for each network according to the fit of all significant genes . A detailed description is given in the online repository ( http://www . ingenuity . com ) . All results are displayed as mean ± standard error of the mean ( mean ± SEM ) . Statistical analyses were performed using one-way ANOVA ( Zebrafish data ) or one-way ANOVA with post hoc T-test analysis using GraphPad Prism 6 ( La Jolla , CA , USA ) for ABR and DPOAE data . P-values < 0 . 05 were assigned as significant . The URLs for data presented herein are as follows: 1000 Genomes; http://www . 1000genomes . org Hereditary Hearing Loss Homepage , http://hereditaryhearingloss . org MutationTaster , http://www . mutationtaster . org NHLBI Exome Sequencing Project Exome Variant Server; http://evs . gs . washington . edu/EVS/ Online Mendelian Inheritance in Man ( OMIM ) , http://www . omim . org/ PolyPhen-2 , http://genetics . bwh . harvard . edu/pph2/ RefSeq , http://www . ncbi . nlm . nih . gov/RefSeq SIFT , http://sift . jcvi . org/
|
The most frequent sensory disorder worldwide is hearing impairment . It impacts over 5% of the world population ( 360 million persons ) , and is characterized by extreme genetic heterogeneity . Over 80 genes have been implicated in isolated ( also referred to as ‘non-syndromic’ ) hearing loss , and abundant evidence supports the existence of many more ‘deafness-causing’ genes . In this study , we used a sequential screening strategy to first exclude causal mutations in known deafness-causing genes in a family segregating autosomal dominant non-syndromic hearing loss . We next turned to whole exome sequencing and identified a single variant—p . Arg185Pro in HOMER2—that segregated with the phenotype in the extended family . To validate the pathological significance of this mutation , we studied two animal models . In zebrafish , we overexpressed mutant HOMER2 and observed inner ear defects; and in mice we documented robust expression in stereocilia of cochlear hair cells and demonstrated that its absence causes early-onset progressive deafness . Our data offer novel insights into gene pathways essential for normal auditory function and the maintenance of cochlear hair cells .
|
[
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
HOMER2, a Stereociliary Scaffolding Protein, Is Essential for Normal Hearing in Humans and Mice
|
Tetherin , an interferon-inducible membrane protein , inhibits the release of nascent enveloped viral particles from the surface of infected cells . However , the mechanisms underlying virion retention have not yet been fully delineated . Here , we employ biochemical assays and engineered tetherin proteins to demonstrate conclusively that virion tethers are composed of the tetherin protein itself , and to elucidate the configuration and topology that tetherin adopts during virion entrapment . We demonstrate that tetherin dimers adopt an “axial” configuration , in which pairs of transmembrane domains or pairs of glycosylphosphatidyl inositol anchors are inserted into assembling virion particles , while the remaining pair of membrane anchors remains embedded in the infected cell membrane . We use quantitative western blotting to determine that a few dozen tetherin dimers are used to tether each virion particle , and that there is ∼3- to 5-fold preference for the insertion of glycosylphosphatidyl inositol anchors rather than transmembrane domains into tethered virions . Cumulatively , these results demonstrate that axially configured tetherin homodimers are directly responsible for trapping virions at the cell surface . We suggest that insertion of glycosylphosphatidyl inositol anchors may be preferred so that effector functions that require exposure of the tetherin N-terminus to the cytoplasm of infected cells are retained .
Cells have evolved numerous defense measures to inhibit the replication of infectious agents . In animal cells , sensing of viruses by pattern recognition receptors leads to interferon production and signaling , which induces the expression of hundreds of interferon-stimulated genes ( ISGs ) in infected and bystander cells [1]–[3] . Among these are several classes of autonomously acting proteins ( the APOBEC3 proteins , TRIM5 proteins , tetherin and SAMHD1 ) . These proteins are popularly termed “restriction factors” , and are considered to comprise an intrinsic immune system [4] or a specialized arm of conventional innate immunity . Recent efforts have revealed that these proteins directly inhibit the replication of viruses via remarkably divergent and elegant mechanisms of action [5] , [6] . Tetherin ( also known as BST-2 , CD317 , or HM1 . 24 ) is a type II membrane glycoprotein whose expression is strongly upregulated by type I interferon in most cell types . Tetherin expression causes the physical entrapment of nascent mature enveloped virions at the cell surface [7]–[11] . Structurally , tetherin comprises of a short N-terminal cytosolic tail , a single pass transmembrane helix , an extracellular domain that is predominantly alpha helical [12]–[15] , and has three extracellular cysteine residues stabilizing parallel homodimer formation via disulphide bridges . Tetherin is also modified at its C-terminus by a glycosylphosphatidylinositol ( GPI ) membrane anchor [16] , [17] . A few pieces of evidence suggest that tetherin acts directly and autonomously to trap virions at the cell surface . First , trapped virions can be liberated from the cell surface by treatment with the protease subtilisin A , indicating that protein is an essential component of the tethers [18] . In such experiments , tetherin fragments can be found in subtilisin-liberated virions [19] . Second , inactive tetherin proteins in which one of the two membrane anchors is removed are efficiently incorporated into virions [19] . Third , fluorescent and electron microscopic analyses demonstrate that tetherin is localized at sites of virion entrapment [19]–[22] . Fourth , an artificial tetherin protein assembled from heterologous protein domains that have similar configuration but no primary sequence homology to tetherin , recapitulates tetherin function [19] . Taken together these findings suggest that ( i ) the biological activity of tetherin can be ascribed to its overall configuration rather than its primary sequence and ( ii ) tetherin does not require specific cofactors or the recognition of specific viral components to cause virion entrapment . These findings are difficult to reconcile with complex models in which tetherin might act as a virion sensor to induce other factors that have tethering activity . Rather , they are more easily explained by the idea that tetherin acts autonomously and directly to trap virions , simply as a consequence of being incorporated into the lipid envelope of virions as they bud through cell membranes . Consistent with these arguments , tetherin exhibits antiviral activity against a broad spectrum of enveloped virions whose proteins have essentially no sequence homology [23]–[30] . Another argument in favor of the notion that tetherin acts rather nonspecifically to trap enveloped virions arises from the mechanisms that viruses have evolved to evade tetherin action . Rather than acquiring viral protein sequence changes that might enable escape from interaction with tetherin , viral proteins have instead adapted to gain interaction with , and thereby antagonize , tetherin . For example , the HIV-1 accessory protein Vpu interacts with the tetherin transmembrane domain [31]–[36] , and employs surface downregulation [37]–[41] and degradation [32] , [42]–[45] to antagonize tetherin . Additionally , the SIV Nef proteins [46]–[49] , the KSHV K5 protein [27] , [50] , and the HIV-2 Env [38] , [51] , SIVMAC Env [52] and Ebola Env [53] , [54] proteins have adapted to counteract tetherin proteins in their hosts by targeting different portions of the tetherin cytoplasmic tail or ectodomain . One question that remains incompletely addressed is the precise molecular mechanisms by which tetherin exerts its antiviral activity . As discussed above , a preponderance of the evidence support a direct tethering mechanism , wherein tetherin dimers infiltrate the lipid envelope of assembling particles [19]–[21] . However , while previous biochemical analyses [19] and structural studies [12]–[15] indicate that tetherin forms a rod-like structure with membrane anchors at either end , the configuration adopted by the tetherin protein during entrapment is unknown . For example , because the membrane anchors are spatially separated from each other , it is possible that one pair of anchors partitions into the lipid envelope of assembling particles , while the other pair remains rooted in the plasma membrane of the infected cell ( axial configuration , Figure 1A ) . In this configuration , each tetherin dimer could potentially link viral and cell membranes in either “polarity” , i . e . with N-termini inserted into either the infected cell or the assembling particle . Other obvious possibilities by which entrapment might be achieved would be via the separate partitioning of dimerized tetherin molecules into virion and cell membranes ( equatorial configuration , Figure 1B ) or the non-covalent oligomerization of tetherin dimers that have both pairs of anchors embedded in either virion envelopes or cell membranes ( Figure 1C ) . Because a protected , β-mercaptoethanol-sensitive , dimeric amino-terminal tetherin fragment can be recovered from virions that have been liberated by protease treatment , it appears that at least some trapped virions are infiltrated by both N-termini of a parallel tetherin homodimer , favoring the models shown in Figure 1A or 1C [19] . Moreover , a tetherin variant that lacks a GPI anchor preferentially localizes to sites of viral budding , suggesting that the tetherin N-terminus provides the dominant driving force for infiltration into budding virions [19] . These results have also been supported by other studies involving super-resolution microscopy [22] . Nevertheless , it remains a challenge to establish if any of the aforementioned configurations are adopted during the retention of virions , or whether the contribution of any one outweighs that of the others . Herein , we have employed quantitative biochemical experiments and engineered tetherin proteins to demonstrate conclusively that tetherin acts directly to trap virions and to elucidate the mechanisms of virion entrapment . Specifically , we placed epitope tags and cleavage sites for the site-specific protease Factor Xa at strategic positions in the tetherin molecule . Virions that were tethered at the cell surface by these modified tetherin proteins were liberated upon specific protease treatment and analyzed . Our results demonstrate that tetherin dimers trap virions by adopting the axial configuration ( Figure 1A ) , with either transmembrane domains or GPI anchors capable of infiltration into assembling particles . Quantitative analyses suggested that , on an average , a few dozen tetherin dimers are involved in trapping each virion and that there is a ∼3–5 fold preference for a tetherin orientation in which the GPI anchored C-terminus rather than the transmembrane domain is inserted into a tethered particle . Taken together , our biochemical experiments constitute the most compelling evidence to date that tetherin is directly responsible for trapping virions at the cell surface and that this is achieved using axially positioned tetherin homodimers , that are primarily configured with their GPI anchored C-termini inserted into virions .
In this study , we endeavored to develop biochemical assays to unequivocally determine whether tetherin acts as a direct tether in trapping virions , and to determine the configuration of tetherin dimers that are engaged in virion entrapment . A variety of approaches , including hydropathy analyses , fusion with reporter enzymes , or the insertion of target sites for proteases , antibodies and chemical modifiers , have been used to deduce membrane protein topology [55] . For example , the insertion of Factor Xa cleavage sites into hydrophilic loops has proven to be useful in such analyses [56] , [57] . We adapted these approaches by engineering modified human tetherin proteins that carried ( i ) single cleavage sites for Factor Xa and ( ii ) epitopes such as hemagglutinin ( HA ) and FLAG tags positioned either N- or C-terminal to the Factor Xa site ( Figure 1D ) . Previous experiences with modified tetherin proteins led to the expectation that these alterations should have no or modest effects on antiviral activity [7] , [19] . Initial experiments in which Factor Xa sites alone were incorporated into Tetherin resulted in proteins that were somewhat refractory to proteolysis ( unpublished observations ) . Hence , we reasoned that the introduction of flexible linkers into its primary sequence might facilitate access to the cleavage site , and increase the efficiency of proteolysis . Therefore , we generated a panel of proteins in which we inserted five and eight GGGGS peptide linker units into the extracellular domain of tetherin , either N-terminal ( at amino acid 50 ) or C-terminal ( at amino acid 157 ) to the predicted coiled-coil domain . The GGGGS peptide is predicted to be unstructured because the glycine residues impart flexibility , and the polar serine residue permits hydrogen bonding to the solvent [58] , [59] . Among the panel of linker modified tetherin proteins , we determined that the proteins with eight linker units C-terminal to the coiled-coil ( C8 , Figure 1D ) and five linker units N-terminal to the coiled-coil ( N5 , Figure 1D ) were expressed at comparable levels to WT tetherin ( Figure 2A ) . Note that Tetherin is heterogeneously glycosylated , and because the cells were lysed in non-reducing buffer , the tetherin proteins migrated primarily as a smear of dimeric species [19] ( Figure 2A ) . To examine the antiviral activity of the linker-modified tetherin proteins , we co-expressed an HIV-1 proviral plasmid ( HIV-1 ( WT ) ) or its Vpu-deficient counterpart ( HIV-1 ( ΔVpu ) ) along with varying amounts of plasmids expressing WT tetherin or one of the modified tetherin proteins . Hereafter , WT tetherin ( Figure 1D ) refers to a previously described construct that harbors an HA epitope tag at amino acid 155 in the extracellular domain , but retains the antiviral activity of the untagged , endogenous protein [7] . As expected , WT tetherin potently inhibited the release of HIV-1 ( ΔVpu ) in a dose-dependent manner , while only marginally affecting the release of HIV-1 ( WT ) ( Figure 2B , C ) . Importantly , the C8 and N5 tetherin proteins were only modestly impaired in their antiviral activity compared to WT tetherin as determined by infectious virion yield and extracellular particulate CA protein measurements ( Figure 2B , C ) and the levels of cell-associated Gag protein were unaffected by the expression of the tetherin proteins ( Figure 2C ) . Thus , the insertion of linker sequences into tetherin was well tolerated with little effect on antiviral activity . We next programmed the C8 and N5 tetherin proteins with single Factor Xa cleavage sites . The rationale was that these proteins ( referred to hereafter as C8Fac and N5Fac respectively , Figure 1D ) would differ in the relative ordering of the HA epitope tag and the protease site . Thus , the epitope tag is positioned N-terminal to the protease site in the C8Fac protein , whereas it is positioned C-terminal to the protease site in the N5Fac protein . In addition to the C8Fac and N5Fac proteins that carried only one epitope tag , we also appended the N-terminus of the N5Fac construct with three tandem FLAG epitope tags . This manipulation results in FLAG and HA epitope tags flanking the protease site ( Flag N5Fac , Figure 1D ) . The use of three FLAG tags in tandem reportedly enhances signal intensity by ∼10–20-fold [60] . Analysis of the antiviral activity of the Factor Xa site-modified tetherin proteins revealed that the C8Fac and N5Fac proteins were only slightly impaired in activity relative to WT tetherin , while the Flag N5Fac protein was nearly indistinguishable in antiviral activity to WT tetherin ( Figure 2B , C ) . The C8Fac and N5Fac proteins were expressed at slightly lower levels than the C8 and N5 proteins respectively ( Figure 2A ) , and were proportionately impaired in antiviral activity ( Figure 2B , C ) . Interestingly , despite harboring more tags as compared to any of the other modified tetherin proteins , the Flag N5Fac protein was virtually as potent as WT tetherin , and expressed at levels indistinguishable from WT tetherin . Vpu antagonized all modified tetherin proteins and restored the yield of extracellular virions ( Figure 2B , C ) . Thus , all modified tetherin proteins mimicked the biological activity and Vpu sensitivity of WT tetherin . We next generated a panel of 293T cells that stably expressed the epitope-tagged Factor Xa-cleavable tetherin proteins . The levels of cell surface tetherin in these stable cell lines was assessed by flow cytometry using a monoclonal antibody that recognizes the extracellular region of human tetherin . Importantly , the surface expression levels of the WT and modified tetherin proteins were quite similar to each other , varying over a 2 . 5-fold range ( mean fluorescent intensities were 6200 , 15000 , 8800 , and 12000 for WT , C8Fac , N5Fac and Flag N5Fac tetherin proteins , respectively ) and were only 1 . 5 to 3-fold greater than that of the endogenous protein in HeLa cells , a prototype tetherin-positive cell line ( mean fluorescent intensity = 5000 , Figure 3A ) . Additionally , we verified that the engineered tetherins exhibited antiviral activity in the stable cell lines using single-cycle HIV-1 replication assays ( Figure 3B ) . As expected , both the WT and the modified tetherin proteins inhibited the release of virions from infected cells , but did not affect cell associated Gag protein expression ( Figure 3B ) . Also , the expression of Vpu reversed the inhibitory effect of the modified tetherin proteins ( Figure 3B ) . Our previous studies have employed a protease “stripping” assay [18] , [61] in which a relatively nonspecific protease ( subtilisin A ) was used to demonstrate that tetherin causes virions to become entrapped on cell surfaces by a protein based tether . The logic underpinning the assay described herein was that if the tetherin protein itself functions as the direct tether , then treatment of cell surfaces with a specific protease ( Factor Xa ) would trigger the release of virions , only when tetherin was programmed with a Factor Xa cleavable site ( Figure 4A ) . Moreover , cleavage should result in partitioning of the epitope-tagged proteolytic fragments either into the liberated virions or the infected cells . Because the epitope tags were strategically positioned relative to the protease site , topological information could be deduced about tetherin in its functional state ( Figure 4B ) . However , because we expect that only a minority of the tetherin molecules on the cell surface would actually be involved in tethering virions , only fragments that are found in virions should be regarded as informative with respect to tetherin topology during virion entrapment . Note that if tetherin adopts the equatorial configurations depicted in Figure 1B or 1C then we would not expect Factor Xa cleavage to result in virion release , because the cleavage sites are positioned outside the region of tetherin-tetherin interaction , in the rod like portion of the molecule . Indeed , the cleavage sites are positioned in artificially introduced sequences whose insertion did not perturb tetherin function ( Figure 2 , 3 ) . Conversely , if tetherin adopts the axial configuration in virion tethers , then Factor Xa cleavage should result in virion release . Moreover , if as depicted in Figure 4B , the HA-tagged proteolytic fragments partition with virions that are liberated from Factor Xa-treated , C8Fac-expressing cells , it would suggest that tetherin dimers exist with their N-termini inserted into the interior of the virion . Conversely , if HA-tagged proteolytic fragments partition with virions that are liberated from the N5Fac cell line , we would deduce that tetherin dimers exist with their GPI anchors embedded in the virion membrane . If , however , tetherin dimers adopt both polarities , then HA-tagged proteolytic fragments would be observed in virions liberated by Factor Xa from both C8Fac and N5Fac expressing cell lines . We first investigated the utility of this approach using cell lines expressing the single epitope tagged C8Fac and N5Fac tetherin proteins . Cells were infected with single-cycle , Vpu-deficient HIV-1 , and constitutively released particles were harvested from culture supernatants . Thereafter , the monolayer of cells was treated with Factor Xa , and then the cell lysates and any liberated virions were also harvested . As before , infected tetherin-negative control cells constitutively released comparatively high levels of virions into the culture supernatant , while virion yield from cells expressing WT , C8Fac or N5Fac tetherin proteins was substantially reduced ( Figure 4C ) . The levels of HIV-1 Gag expression in cell lysates were uniform ( Figure 4C ) . Incubation in Factor Xa cleavage buffer alone resulted in the release of only low levels of pelletable CA from tetherin-deficient cells . This may have represented virion particles that were constitutively released during incubation , or virions that were loosely adhered to the cell surface ( Figure 4C ) . Even lower levels of particles were released from cells expressing the WT , C8Fac or N5Fac tetherin proteins that were incubated in Factor Xa cleavage buffer alone . Strikingly however , Factor Xa treatment of the C8Fac and N5Fac resulted in the release of substantial amounts of particulate CA ( Figure 4C ) . Crucially , Factor Xa treatment of tetherin-negative or WT tetherin expressing cells did not increase particle release over the low background levels that were observed in the absence of protease , underscoring the strict requirement for a Factor Xa-cleavable tetherin in Factor Xa-induced virion release ( Figure 4C ) . Notably , proteolytic fragments of tetherin were observed in virions released by Factor Xa from both C8Fac and N5Fac expressing cells and these virion-associated fragments were consistent with the incorporation of tetherin dimers therein . These dimers were the only tetherin species that were detectable on non-reducing SDS PAGE gels ( Figure 4C ) . Because tetherin is intrinsically heterogeneous , due to variable glycosylation as well as dimer formation , it was difficult to assess the extent of Factor Xa cleavage in cell lysates ( Figure 4C , center panel ) , or to unambiguously demonstrate that only cleaved tetherin fragments were present in Factor Xa liberated virions ( Figure 4C , bottom panel ) . Therefore we treated cell and virion lysates with PNGase-F and repeated the western blot analyses under reducing conditions . We observed that the HA-tagged proteolytic fragments ( predicted molecular weights of ∼20 . 8 kDa and ∼17 kDa for C8Fac and N5Fac respectively ) could be resolved from the full-length molecules ( ∼24 . 7 kDa and ∼23 . 8 kDa for C8Fac and N5Fac respectively ) ( Figure 4D ) . This analysis revealed that about half of the cell-associated C8Fac and N5Fac protein was cleaved by Factor Xa that was applied to the cell surface . The incomplete cleavage may have been due to the intracellular localization of a fraction of the tetherin protein . As expected , no proteolysis of the WT tetherin protein was observed ( Figure 4D ) . Notably , only the cleaved tetherin protein was found in PNGase-F-digested virion lysates , consistent with the notion that tetherin cleavage by Factor Xa was necessary for virion release in this assay . To assess the efficiency of tetherin cleavage and virion release by Factor Xa , we compared the levels of virion released from C8Fac and N5Fac expressing cell lines following treatment with Factor Xa or with subtilisin A ( Figure 4E ) . Similar amounts of virions were released by the site-specific and non-specific proteases . This result suggested that tetherin cleavage and virion release caused by Factor Xa was quite efficient . It also suggested that it was unlikely that a significant fraction of virions are retained using alternative configurations of tetherin ( Figure 1 ) in which virion release might be resistant to Factor Xa treatment . Overall , these results strongly suggested that tetherin traps virions by adopting the axial configurations depicted in Figure 1A . Moreover , because HA-tagged proteolytic fragments from both C8Fac and N5Fac tetherin proteins partitioned with virions these data suggested that both polarities depicted in Figure 1A are adopted by tetherin during virion entrapment . To estimate the number of tetherin dimers that were involved in the entrapment of a single virion , we used a quantitative western blotting approach and PNGase-F-digested virion lysates to measure the relative number of CA and HA epitopes associated with virions that had been tethered by the C8Fac and N5Fac proteins , and then released by Factor Xa cleavage . First , we generated an appropriate internal standard protein to enable relative quantitation . This standard consisted of a fusion protein that comprised the HIV-1 p24CA protein , appended at its C-terminus with three tandem FLAG tags and an HA epitope tag . Thus , this single protein included each of the epitopes that we planned to probe , at a stoichiometric ratio of 1∶1∶1 and could be used as a standard to compare the relative numbers of HA and CA epitopes in tethered virions liberated from C8Fac and N5Fac expressing cells . Specifically , serial dilutions of cell lysates expressing the HA-Flag-CA protein were run on SDS-PAGE gels , blotted onto membranes and probed with antibodies against CA and HA . The band intensities were analyzed using a LiCOR Odyssey scanner ( Figures 5A , B ) , and regression analysis was performed over the linear range of signal intensities ( Figures 5A , B ) . Dilutions of the PNGase-F-treated virion lysates recovered from C8Fac and N5Fac expressing cells that also yielded band intensities in the linear range of the assay were resolved on the same gel as the standard , and the relative amounts of CA and HA epitope in each samples were deduced by interpolation using the standard curves ( Figures 5A , B ) . HIV-1 virions have been reported to contain between 1000–5000 copies of the Gag protein , of which only a fraction contribute to core formation [62]–[66] . We calculated our estimates of tetherin dimers per virion based on the extremities of this range ( Table 1 ) . Thus , if each virion contains 1000 CA protein molecules , we estimate that 16±5 dimers of the N-terminus of C8Fac and 71±26 dimers of the C-terminus of N5Fac tetherin dimers were associated with a single tethered virion ( Table 1 ) . Conversely , if a single virion contains 5000 CA epitopes , then we estimate that 80±25 dimers of the N-terminus of C8Fac and 355±130 dimers of the C-terminus of N5Fac tetherin dimers were associated with a single tethered virion . Thus these numbers suggested a preference ( ∼4 to 5-fold ) for the insertion of the GPI-anchored tetherin C-terminus , rather than the N-terminal transmembrane domain into virions . Note that the larger number of HA tags associated with virions in the case of N5Fac cannot be explained by differences in tetherin expression levels . In fact , there were lower levels of N5Fac on cell surfaces ( MFI = 8800 , Figure 3A ) as compared to the C8Fac protein ( MFI = 15000 , Figure 3A ) . The aforementioned experiments indicated that tetherin directly tethers HIV-1 particles in an axial configuration ( Figure 1A ) and suggested that both polarities , with either N- or C- termini inserted into virions contribute to antiviral activity . However , it was possible that the two different estimates for the numbers of tetherin molecules inserted into virions with each polarity might reflect intrinsic differences in the properties of the two different tetherin molecules used ( C8Fac and N5Fac ) . Therefore , we quantitated tetherin insertion into virions in a second set of experiments employing a single tetherin species with two different epitope tags on either side of the Factor Xa cleavage site ( Flag N5Fac , Figure 1D , Figure 6A ) . Additionally , we have previously found that virions that accumulate on the surface of cells as a result of tetherin action can sometimes be tethered to each other as well as to the cell surface . This scenario could be the result of virion assembly at sites on the cell surface already occupied by trapped virions and would result in both ends of a tetherin molecule being associated with virions . These events would tend to reduce any indication that tetherin N-or C-termini are preferentially inserted into virion envelopes . Because the accumulation of virions should exacerbate this effect over time , we treated the surface of cells expressing Flag N5Fac with Factor Xa at predetermined time intervals following infection with HIV-1ΔVpu , and quantified HA- and FLAG-tagged proteolytic fragments in liberated virions . The HIV-1 Gag protein became detectable in infected Flag N5Fac-expressing cell lysates at ∼24 h after infection and levels progressively increased with time thereafter ( Figure 6B ) . Treatment of these infected cells with Factor Xa resulted in a time dependent increase in the amount of recovered virions ( Figure 6B ) . The Factor Xa site is positioned N-terminal to the sites of N-linked glycosylation as well as to the extracellular cysteines in the Flag N5Fac molecule ( Figure 1D , Figure 6A ) and so the Factor Xa cleavage of the 65–70 kDa dimeric , glycosylated Flag-N5Fac protein yields a cell associated dimeric , glycosylated ∼50–55 kDa αHA reactive species as well as a cell associated monomeric , nonglycosylated 10 kDa α-FLAG reactive species ( Figure 6B ) . Notably , both the dimeric glycosylated ∼50–55 kDa αHA reactive species and the 10kDa α-FLAG reactive species were observed in virions , and their levels in the virion fraction increased with time , approximately in parallel with the increasing yield of Factor Xa liberated virions ( Figure 6B ) . Notably , the N-terminal FLAG tagged fragment of Flag N5Fac was also found in virions in a form that was consistent with the formation of dimers . We hypothesize that this is because the tetherin cytoplasmic tail contains two cysteines that can form disulphide bonds in the interior of virions . Consistent with this idea , only the smaller of the two Flag tagged species was observed when virion lysates were subjected to SDS PAGE gel electrophoresis under reducing conditions ( Figure 6C ) . Additionally , the dimeric glycosylated ∼50–55 kDa αHA reactive species collapsed to a single ∼17 kDa band when samples were deglycosylated with PNGase and reduced ( Figure 6C ) . We used quantitative western blot analyses of PNGase-F-digested virion lysates to estimate the number of copies of HA- and FLAG-tagged proteolytic fragments per trapped virion ( Figure 6C , D ) . Again we used the FLAG-HA-CA protein as a standard to determine the relative numbers of HA , FLAG and CA epitopes in the virions liberated from Flag N5Fac expressing cells . Although tethered virions could be recovered from the surface of Flag N5Fac expressing cells beginning at 24 h after infection , we could not make reliable estimates of the HA and FLAG fragments at this time point , as they were present at levels that were close to the limit of detection . However , we could make reasonably robust estimates of the levels of incorporation of HA- and FLAG-tagged fragments into virions beginning at 32 h after infection . Importantly , the number of FLAG-tagged dimers that were estimated to be present in virions ( assuming 1000 CA molecules per virion ) tethered by Flag N5Fac ( 11±3 [at 32 h] to 16±6 [at 48 h] , Figure 6C , D , Table 2 ) correlated quite well with the number of HA-tagged dimers present in virions tethered by C8Fac ( 16±5 [at 48 h] , Table 1 ) . Similarly , the number of HA-tagged dimers in tethered virions recovered from the Flag N5Fac expressing cells ( 34±18 [at 32 h] to 55±28 [at 48 h] ) ( Figure 5D , Table 2 ) correlated quite well with the number of copies of HA-tagged dimers in tethered virions recovered from the N5Fac expressing cells ( 71±26 at 48 h ) ( Table 1 ) . Notably , we estimated that the virions liberated from Flag N5Fac expressing cells carried ∼3 to 4-fold more HA tags than FLAG tags , again suggesting that axially configured tetherin dimers infiltrate assembling particles , with a tendency to embed their C-termini rather than their N-termini in tethered virions . Also noticeable was a marginal trend for the appearance of increasing numbers of tetherin molecules per virion over time . This trend was not statistically significant and could be due to some unknown bias in the measurements . However , it is also possible that virions with smaller numbers of tetherin molecules are more readily released , leading to the preferential accumulation of virions with greater numbers of tetherin molecules on the surface of cells . Finally , to confirm that alternative Factor Xa-resistant configurations of Flag N5Fac tetherin were not responsible for retaining a significant fraction of virion particles , we compared the levels of virions released from Flag N5Fac expressing cells by Factor Xa or by subtilisin A treatment . Similar amount of particles were released by the two proteases , suggesting that axially configured Flag N5Fac tetherin molecules were the major form responsible for virion retention ( Figure 6E ) .
We devised a biochemical approach to probe tetherin molecules that have infiltrated virions at the cell surface , with the goal of elucidating the configuration adopted by tetherin during virion entrapment . This approach was based on two previous findings . First , a non-specific protease , subtilisin , could be used to liberate tethered particles from the infected cell's surface [18] , [61] . Second , the primary sequence of tetherin can be drastically altered while retaining biological activity [19] . Thus , we employed the site-specific protease Factor Xa to liberate virions trapped by tetherin molecules that were engineered to include its cleavage site . This manipulation gave the approach tight specificity and enabled the unequivocal demonstration that the tetherin protein itself is an essential component of virion tethers . Moreover , the use of a site specific protease to release tethered virions from cell surfaces enabled the preservation of epitope tags inserted into the tetherin ectodomain , allowing us to infer the organization of tetherin molecules in virion tethers . We could use a double epitope-tagged version of tetherin , as well as single epitope-tagged versions to analyze the incorporation of both N- and C-terminal proteolytic fragments into virions , and thereby determine tetherin configuration . Additionally , we constructed a protein standard and performed quantitative western blotting to estimate the numbers of tetherin dimers in each orientation that are associated with trapped virions . Because virions were efficiently liberated by Factor Xa treatment of N5Fac or C8Fac expressing cells , our data effectively exclude the “equatorial” configuration shown in Figure 1B , as cleavage of the tetherin peptide backbone in this context would leave intact the majority of the bonds holding the virion on the cell surface . Moreover , the fact that tetherin fragments found in virions liberated by Factor Xa were exclusively disulphide linked homodimers also constitutes strong evidence disfavoring this model . While our data do not completely discount the possibility that tetherin multimers adopt the equatorial configuration , with virions becoming trapped via hypothetical noncovalent dimer-dimer interactions ( Figure 1C ) this scenario appears unlikely for two reasons . First , such a configuration would not be expected to result in virion release upon Factor Xa cleavage , because dimer-dimer interactions would not be expected to be perturbed , particularly since the Factor Xa cleavage site is placed within a foreign spacer sequence whose insertion does not itself perturb tetherin function . Second , the scenario envisaged in Figure 1C would result in precisely equal numbers of tetherin N- and C termini being placed in tethered virions . We found that there were modestly , but clearly , more tetherin C-termini than N-termini in virions , arguing that tetherin N- and C-termini partition separately into virion and cell membranes . Overall our experiments indicated that tetherin homodimers adopt an axial configuration in their functional state , with a preference for the insertion of their GPI-anchored C-termini into virions during their entrapment at the surface of infected cells . Quantitative analysis indicated that an average of ∼80 to 400 tetherin dimers ( depending on how many CA molecules are assumed to be present in each virion ) associated with each tethered particle . Our findings do not discount the discount the possibility that higher order tetherin multimers , e . g . tetramers , might contribute to tethering , but if such complexes do exist , then they must involve non-covalent interactions between axially configured tetherin molecules and be arranged in such a way that all N-termini and in one membrane ( be it virion envelope or cell membrane ) and all C-termini are in the opposing membrane . Previous studies have not resolved the configuration adopted by tetherin during virion entrapment . For example , conflicting results have been obtained in studies where the release of virions was attempted by cleavage of the tetherin GPI anchor using phosphatidyl-inositol-specific phospholipase C ( PI-PLC ) . In one study , the efficiency of virion release induced by PI-PLC treatment was poor ( ∼20% compared to subtilisin ) [22] , while other studies indicated that PI-PLC treatment fails to liberate any virions [20] , [67] . Second , the failure of reducing agents to release virions would tend to suggest that the equatorial model shown in Figure 1B is incorrect [20] . However , this argument is somewhat confounded by the fact that tetherin molecules are twisted around each other in a dimer , and so breaking the disulphide bonds in an already-formed tether would not necessarily be expected to cause virion release . One caveat of our assay is that some tetherin dimers might infiltrate particles and yet be uninvolved in restriction . Thus , it is possible that the number of tetherin molecules that we measured to be associated with a virion might be greater than the number of molecules actually involved in virion entrapment . Indeed , previous studies have shown that low levels of complete tetherin molecules can be found in the small number of virions that are released from tetherin-positive cells [20] . However , to be uninvolved in restriction would require that both tetherin N- and C-termini were embedded in virions . If the numbers of tetherin dimers that were inserted into virions in this way was in excess of the numbers of tetherin dimers involved in tethering , with N- and C-termini partitioned separately into virion and cell membranes , then there would be little or no difference in the number of tetherin N- and C-termini found in virions . The fact that we do indeed observe a 3-to 5-fold excess C-termini in tethered virions , argues strongly that most of the tetherin molecules ( at least 65–80% ) that are tethered-virion associated , have their N- and C-termini separately partitioned into virion and cell membranes . Thus most tetherin molecules must be in the axial configuration with only their C-termini embedded in virions . Our estimates of the number of tetherin molecules that are associated with tethered virions are several-fold higher than those obtained using super-resolution microscopy approaches ( i . e . 4–7 dimers per virion ) [22] . At least three factors could account for this discrepancy . First , the microscopy studies use a tetherin-mEosFP fusion protein , that includes a bulky 230 amino acid ( ∼26 kDa ) protein at its N-terminus , appended to the otherwise short ( 21 amino acid ) native tetherin cytoplasmic tail . This could very easily reduce the numbers of tetherin molecules that associate with virions . Second , the estimates made in the microscopy studies correspond to groups of tetherin molecules present at the same location as clusters of Gag molecules that may not represent completely assembled virions . Thus , microscopy studies cannot determine whether the imaged tetherin molecules are in the act of restriction . Conversely , our estimates are based on bona fide tethered virions that are recovered from cells by specific cleavage of the tether . Finally , the cell lines that we used to derived our estimates modestly overexpressed tetherin ( 1 . 5- to 3- fold ) as compared to HeLa cells , which might have slightly elevated the numbers of tetherin molecules that were associated with virions . In previous studies [22] , transfected HeLa cells were used , and the levels of tetherin-mEosFP relative to preexisting endogenous tetherin , or the total ( endogenous plus exogenous ) levels of tetherin expression were not determined , which could lead to underestimates or overestimates of tetherin association with tethered virions . Given that virions are trapped not only at the cell surface , but are also linked to each other , it should be expected that both tetherin N- and C-termini would be found in virions . Most likely , the appearance of virions tethered to each other results from the assembly of a virion at a location on the plasma membrane already occupied by a tethered particle . This being so , our finding of a 3- to 5-fold preference for the insertion of C-termini rather than N-termini into virion membranes may represent an underestimate of the true preference . If this is the case , then one might expect that the apparent preference for the insertion of C-termini into virions would become less apparent over time as virion accumulate at the cell surface and the likelihood of a virion assembly at a site already occupied by a tethered virion increased . However , we did not observe such a trend , and thus it remains unclear whether the 3- to 5-fold preference for C-terminus insertion into virions is an accurate number , or an underestimate resulting from virion accumulation . The biophysical mechanism underpinning the apparent preference for the insertion of GPI-anchored C-termini over TM domain anchored N-termini into virions is unclear at present . Although it is not the predominant scenario , the tetherin N-terminal domain is clearly capable of being incorporated into virions . Indeed , a tetherin molecule lacking the GPI anchor is efficiently incorporated into released virions [19] . Moreover , it is the N- terminus that is targeted by Vpu to block tetherin incorporation into virions [19] , [33] . Perhaps the tetherin N-terminal domain acts as a sensor of membrane curvature , driving localization to assembly sites , but the GPI anchor diffuses more freely into virion membranes . Consistent with this idea , recent work indeed indicates that tetherin colocalizes better with HIV-1 Gag proteins that cause membrane curvature than those which do not [68] . There is potential biological utility in preferentially inserting GPI anchored tetherin C-termini rather than N-termini into virions . In such a scenario , the tetherin N-terminus remains available to the cytoplasm of the infected cell , from where it may execute important functions . For instance , virions trapped at the cell surface are internalized and degraded in lysosomes [61] , [69] . Moreover , human tetherin appears capable of initiating signaling cascades , particularly when it is engaged in tethering , and in some respects may act as a virion sensor [70]–[72] . Thus , the need to interact with the endocytic machinery and/or initiate signaling might favor a scenario in which tetherin dimers are oriented with their N-termini in the infected cell and their C-termini in the virion membrane .
Tetherin was transiently expressed using pCR3 . 1 ( Invitrogen ) based plasmids or stably expressed using pLHCX ( Clontech ) based retroviral vectors . A human tetherin protein internally tagged with an HA epitope at amino acid 155 and , expressed using pCR3 . 1 or LHCX vectors , has been described previously [7] . Eight copies of a peptide linker sequence , each comprising the amino acid sequence GGGGS , were inserted immediately C-terminal to the HA tag , to generate the C8 modified tetherin protein ( Figure 1D ) . Similarly , five GGGGS linker units were inserted immediately C-terminal to the tetherin transmembrane domain at amino acid position 50 , to generate the N5 modified tetherin protein . Because the BamHI recognition site ( GGATCC ) encodes a glycine and serine , we incorporated its sequence into the fourth and third linker units for the C8 and N5 proteins respectively . We then used these BamHI sites for the subsequent insertion of a Factor Xa cleavage site ( IEGR ) to generate the C8Fac and N5Fac proteins ( Figure 1D ) . Thereafter the Flag N5Fac protein was generated by inserting three copies of a FLAG epitope tag at the N-terminus of the N5Fac protein ( Figure 1D ) . The protein standard used for quantitative western blotting was generated by appending the C-terminus of HIV-1 p24 CA protein with three FLAG epitope tags and an HA epitope tag . Specifically , the p24 CA coding sequence was amplified from the proviral plasmid pNL4-3 using oligonucleotides that encoded the epitope tags , and inserted as an EcoRI-NotI fragment into the multiple-cloning site of pCRV-1 , a previously described hybrid expression vector [73] that is derived from pCR3 . 1 and from a highly modified HIV-1 provirus ( V1B ) . All mutagenesis was accomplished by using overlap-extension PCR . Human embryonic kidney ( HEK ) 293T cells and HeLa-TZM cells expressing CD4/CCR5 and a LacZ reporter gene under control of the HIV-1 LTR were maintained in Dulbecco's Modified Eagle Medium ( DMEM ) supplemented with 10% FBS and gentamycin ( 2 µg/ml , Gibco ) . HEK293T cells were transduced using pLHCX based retroviral vectors expressing genes of interest and selected with hygromycin ( 50 µg/ml ) ( MediaTech , Inc ) to generate cell lines expressing either the empty vector or epitope-tagged WT or modified tetherin proteins . The 293T cells stably expressing the modified tetherin proteins and HeLa cells were harvested in PBS plus 5mM EDTA , washed in FACS buffer ( PBS plus 2% BSA ) , and stained with PE anti-human CD317 ( tetherin ) antibody ( Biolegend ) . Dead cells were excluded by DAPI staining . All data were acquired on an LSR II flow cytometer ( Becton Dickinson ) , and data were analyzed with FlowJo software ( Tree Star ) . A HIV-1 proviral plasmid that expresses green fluorescent protein ( GFP ) in place of Nef has been described previously [74] . 293T cells were seeded in 10 cm plates at a concentration of 3×106 cells/plate and were cotransfected the following day using polyethylenimine ( PolySciences ) with 10 µg of wild-type ( HIV-1 ( WT ) ) or Vpu-deficient ( HIV-1 ( ΔVpu ) ) GFP reporter plasmids , along with 1 µg of a VSV-G expression plasmid . The culture medium was replaced the following day . At 48 hours post transfection , the culture supernatants were harvested , clarified by centrifugation at 3000 rpm , and filtered through a 0 . 2 µm PVDF membrane ( Millipore ) . The viruses were stored at -80°C . Infectious virus titers were determined by inoculating sub-confluent monolayers of 293T cells that were seeded in 96 well plates at 30 , 000 cells/well with 100 µl of serially diluted supernatants . At 48 hours post infection , the cells were dispersed with trypsin , fixed in 4% paraformaldehyde and analyzed by flow cytometry . 293T cells were seeded in 24-well plates at a concentration of 2×105 cells/well and were cotransfected the following day using polyethylenimine ( PolySciences ) with 350 ng of wild-type ( HIV-1 ( WT ) ) or Vpu-deficient ( HIV-1 ( ΔVpu ) ) proviral plasmids along with varying amounts of a Tetherin expression plasmid ( 25 ng to 100 ng ) and a plasmid expressing YFP ( 75 ng ) , to monitor transfection efficiency . In all transfection experiments , the total amount of DNA was held constant by supplementing the transfection with an empty expression vector . The culture medium was replaced the following day . At 48 hours post transfection , the culture supernatants were harvested , clarified by centrifugation at 3000 rpm , and filtered through a 0 . 2 µm PVDF membrane ( Millipore ) . Infectious virus yield was determined by inoculating sub-confluent monolayers of HeLa-TZM cells that were seeded in 96 well plates at 10 , 000 cells/well with 100 µl of serially diluted supernatants . At 48 hours post infection , β-galactosidase activity was determined using GalactoStar reagent , in accordance with the manufacturer's instructions ( Applied Biosystems ) . Physical particle yield was determined by layering 700 µl of the virion containing supernatant onto 1 ml of 20% sucrose in PBS followed by centrifugation at 20 , 000×g for 90 minutes at 4°C . Virion pellets were then analyzed by Western blotting . Cells ( HEK293T ) stably expressing WT or engineered tetherin proteins were infected with VSV-G-pseudotyped HIV-1 ( WT ) or HIV-1 ( ΔVpu ) GFP at 1 infectious unit per cell in 10 cm dishes . The inoculum was removed 6 h later . At 48 hours post transfection , the culture supernatants were harvested , clarified by centrifugation at 3000 rpm , and filtered through a 0 . 2 µm PVDF membrane ( Millipore ) . Physical particle yield was determined as outlined above . Simultaneously , the cells were washed with Factor Xa reaction buffer ( 20 mM Tris·Cl , pH 6 . 5; 50 mM NaCl; 1 mM CaCl2 ) and incubated with 50 µg of Factor Xa in 5 ml of Factor Xa reaction buffer for 2 hours at 37°C . Alternatively , the cells were washed with with subtilisin A buffer ( 10 mM Tris , pH 8 . 0; 1 mM CaCl2; 150 mM NaCl ) , and treated with 5 ml of 1 µg/ml of subtilisin A ( Sigma ) for 3 min at room temperature . Subtilisin treatment was stopped using DMEM containing 10% FCS , 5 mM PMSF , and 20 mM EGTA . Thereafter , the supernatants were centrifuged , filtered and virions pelleted as described above , and the cells were lysed for analysis of viral protein expression by Western blotting . Lysates of cell and liberated virions were denatured with 0 . 5% SDS at 100°C for 10 minutes and then treated with 1% NP-40 to neutralize the SDS . The lysates were incubated with ( or without ) 500 U of peptide-N-glycosidase-F ( New England Biosciences ) at 37°C for 3 hours . Thereafter , the reactions were quenched with SDS-PAGE loading buffer and the samples were analyzed with western blotting . Pelleted virions and cell lysates were resuspended in SDS-PAGE loading buffer , in the presence or absence of β-mercaptoethanol , and resolved on NuPAGE Novex 4–12% Bis-Tris Mini Gels ( Invitrogen ) in MOPS running buffer . Proteins were blotted onto nitrocellulose membranes ( HyBond , GE-Healthcare ) in transfer buffer ( 25 mM Tris , 192 mM glycine ) . The blots were then blocked with Odyssey blocking buffer and probed with mouse anti-HIV-1 capsid ( NIH ) , rabbit anti-HA ( Rockland ) , and mouse anti-FLAG ( Sigma ) primary antibodies . For quantitative western blotting , the bound primary antibodies were detected using fluorescently labeled secondary antibodies ( IRDye 800CW Goat Anti-Mouse Secondary Antibody , IRDye 680LT Goat Anti-Rabbit Secondary Antibody and IRDye 680LT Goat Anti-Mouse Secondary Antibody; LI-COR Biosciences ) . Fluorescent signals were detected using a LI-COR Odyssey scanner and quantitated with Odyssey software ( LI-COR Biosciences ) .
|
The cellular restriction factor , tetherin , prevents HIV-1 and other enveloped virus particles from being disseminated into the extracellular milieu by infiltrating their envelopes and by physically crosslinking them to the cell surface . It is known that tetherin consists of pairs of membrane anchors , situated at either end of a rod-shaped molecule , but how tetherin causes virion tethering has been difficult to unambiguously determine . In this work , we develop genetic and biochemical approaches to probe tetherin molecules that have infiltrated tethered virions . We show that tetherin adopts an “axial” configuration in its functional state , with a pair of membrane anchors situated at one end of the rod-like structure inserted into a tethered virion . While either end of the rod can be inserted into a virion , there is a preference for the insertion of its lipid ( glycosylphosphatidyl inositol ) modified carboxyl-terminus into virion envelopes . These studies demonstrate unequivocally that the tetherin molecule itself is directly responsible for trapping virions , and dissect the molecular mechanism underpinning its antiviral activity .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"immunodeficiency",
"viruses",
"virology",
"biology",
"microbiology"
] |
2013
|
Mechanism of HIV-1 Virion Entrapment by Tetherin
|
At the crossroad between the NF-κB and the MAPK pathways , the ternary complex composed of p105 , ABIN2 and TPL2 is essential for the host cell response to pathogens . The matrix protein ( M ) of field isolates of rabies virus was previously shown to disturb the signaling induced by RelAp43 , a NF-κB protein close to RelA/p65 . Here , we investigated how the M protein disturbs the NF-κB pathway in a RelAp43-dependant manner and the potential involvement of the ternary complex in this mechanism . Using a tandem affinity purification coupled with mass spectrometry approach , we show that RelAp43 interacts with the p105-ABIN2-TPL2 complex and we observe a strong perturbation of this complex in presence of M protein . M protein interaction with RelAp43 is associated with a wide disturbance of NF-κB signaling , involving a modulation of IκBα- , IκBβ- , and IκBε-RelAp43 interaction and a favored interaction of RelAp43 with the non-canonical pathway ( RelB and p100/p52 ) . Monitoring the interactions between host and viral proteins using protein-fragment complementation assay and bioluminescent resonance energy transfer , we further show that RelAp43 is associated to the p105-ABIN2-TPL2 complex as RelAp43-p105 interaction stabilizes the formation of a complex with ABIN2 and TPL2 . Interestingly , the M protein interacts not only with RelAp43 but also with TPL2 and ABIN2 . Upon interaction with this complex , M protein promotes the release of ABIN2 , which ultimately favors the production of RelAp43-p50 NF-κB dimers . The use of recombinant rabies viruses further indicates that this mechanism leads to the control of IFNβ , TNF and CXCL2 expression during the infection and a high pathogenicity profile in rabies virus infected mice . All together , our results demonstrate the important role of RelAp43 and M protein in the regulation of NF-κB signaling .
Depending on specific stimuli , the transcription factors of the NF-κB family are major regulators of cellular physiology . They regulate apoptosis , cell survival , proliferation and immune response , which requires each step of the cell signaling cascade to be highly regulated [1] . The NF-κB family is constituted of 3 proteins with a transactivation domain ( TAD ) : RelA ( p65 ) , cRel and RelB and 3 proteins lacking a TAD: p105/p50 ( NF-κB1 ) , p100/p52 ( NF-κB2 ) and RelAp43 , a sixth member of the NF-κB family and splicing variant of RelA [2] . All these proteins share a Rel Homology Domain ( RHD ) involved in dimerization , DNA- and IκB binding . The NF-κB pathway is divided in a canonical pathway , which mostly relies on RelA-p105/p50 dimers , and a non-canonical pathway involving RelB-p100/p52 dimers . Upon activation of the cascade , the IκB kinase ( IKK ) complex is phosphorylated and induces the phosphorylation of IκB proteins , such as IκBα or IκBβ , leading to their ubiquitination and degradation by the proteasome . Similarly , p100 and p105 , acting both as precursors and inhibitors , are phosphorylated and cleaved into maturated p52 and p50 upon IKK activation . Therefore , NF-κB dimers are released from their inhibitors and free to translocate into the nucleus to regulate the expression of their target genes . Interestingly , a pool of p105 is also forming a complex with the A20-binding inhibitor of NF-κB activation 2 ( ABIN2 ) and the Tumor progression locus 2 ( TPL2 ) . Both ABIN2 and TPL2 interact with the C-terminal ( C-ter ) half of p105 and not with p50 [3] . TPL2 , stabilized and inhibited by p105 , is a key MAP3K involved in the activation of the immune response mediated by the MAPK pathway through downstream kinases MEK1/2 and ERK1/2 . ABIN2 is essential to maintain steady-state TPL2 levels , so that in ABIN2 depleted cells , the activation of downstream ERK is reduced under stimulation [4] . TPL2 induces the NF-κB pathway without prior stimulation in a p105-dependant manner [5] , increasing the production of p50 and its translocation in the nucleus together with RelA , thereby activating NF-κB-responsive reporter genes . Interestingly , TPL2 phosphorylates p105 at different sites compared to IKK and is autophosphorylated in unstimulated cells [6] . It has been suggested that TPL2 promotes the full processing of p105 by the proteasome , releasing independent NF-κB dimers while maintaining the overall rate of p50 production . Indeed , TPL2 could modulate the NF-κB pathway while being associated to a restricted pool of p105 forming the core of high-molecular-weight regulatory complexes [3 , 7] . Finally , upon IKK activation , p105 is phosphorylated by the regulatory IKKγ/NEMO subunit and ABIN2 is released from the complex together with TPL2 , which once liberated , can activate the ERK/MAPK pathway . This ternary complex then establishes an intersection step between both NF-κB and MAPK pathways , notably involved in the regulation of IFNß , CXCL2 , TNF , CCL5 or CXCL10 [8 , 9] , constituting a target of choice for immune escape by viruses such as Lyssavirus , the agents of rabies . Lyssaviruses are already known to interfere with several signaling pathways [10] . One of the 5 proteins coded by the lyssavirus genome , the phosphoprotein ( P ) inhibits both the IRFs [11] and the JAK-STAT pathways [12] through a mechanism conserved across the Lyssavirus genus [13] . Another viral protein , the matrix protein ( M ) , is a small pleiotropic protein forming oligomers and involved in various functions such as structural organization of the viral particle or potentially in the formation of viral inclusions [14 , 15] . First involved in the dysregulation of cell homeostasis and induction of cell death [16–18] , the M protein has been also shown to be associated with the inhibition of the NF-κB pathway through an interaction with RelAp43 [2 , 19] . Interestingly , only the M protein of field rabies virus ( RABV ) isolates such as the Thailand strain ( MTha ) was shown to interact with RelAp43 , compared to attenuated viral strains such as the vaccinal PV or SAD virus ( MSAD ) . Recently , 4 residues of the M protein ( position 77 , 100 , 104 , 110 ) were shown to be involved in the interaction of MTha with the C-ter part of RelAp43 and linked to the control of the expression of RelAp43-dependant genes [19] . Here we further characterize the modulation of the NF-κB network mediated by RelAp43 and the role of MTha in the formation of RelAp43-p50 dimers through its interaction with a complex formed by RelAp43 , NF-κB1 ( p105/p50 ) , ABIN2 and TPL2 ( RNAT ) . Interestingly , although MTha , MSAD and MTh4M ( a defective mutant ) are interacting with the complex , we can distinguish several profiles of interaction . All M proteins can interact with TPL2 , but only MTha has a strong interaction for both RelAp43 and ABIN2 . Thereby only the MTha protein and Tha virus are able to affect NF-κB signaling leading to the control of the host response to the infection . Taken together , our data bring new insight in the NF-κB pathway and particularly on the p105-ABIN2-TPL2 complex , shedding light on the role of the matrix protein on cell signaling and on viral immune evasion .
Using RelAp43 as bait in a tandem affinity purification ( TAP ) assay associated with a label-free quantitative proteomics approach , we identified 329 proteins and 49 of them were shown to be significantly interacting–directly or not—with RelAp43 ( S1 Fig ) . For 38 of these proteins , the interaction with RelAp43 was observed to be modulated ( 5 decreased and 33 increased ) in the presence of MTha . The network of RelAp43 interactome was build based on hits identified after RelAp43 purification ( S2 Fig ) , revealing highly intra- and inter-connected clusters of protein . Besides the NF-κB pathway , RelAp43 appears to form complexes with proteins involved in gene expression regulation ( mRNA splicing , nucleoplasm ) , protein expression regulation ( translation , ubiquitin-proteasome ) or the regulation of protein localization ( cytoskeleton , nuclear transport ) . As expected from previous work on RelAp43 [2] , in absence of MTha our approach led us to identify all known NF-κB proteins except RelA , as well as IκBα and the two other major members of the IκB proteins: IκBß and IκBε ( in blue in Fig 1A left ) . The lack of RelA identification can be explained ( 1 ) by the common RHD of RelA and its overexpressed variant RelAp43 which are undistinguishable and ( 2 ) by the presence of only 2 Lys and 1 Arg in the 207 residues constituting the specific C-ter sequence of RelA , which strongly impairs its capacity to be detected by mass spectrometry ( MS ) conversely to the specific C-ter part of RelAp43 ( S3 Fig ) . Nevertheless , p100/p52 , and p105/p50 were quantified as some of the most important partners of RelAp43 ( Fig 1A left panel , 1B ) . In addition to the 49 proteins significantly interacting with RelAp43 , many other proteins including 5 NF-κB / IκB proteins were identified but failed the statistical test ( Fig 1A ) , most likely due to the low intensity of detected peptides , and could nonetheless be confidently considered in further investigation ( S2 Fig ) . While the interaction of cRel and p105/p50 with RelAp43 is not modulated by MTha , IκBß interaction with RelAp43 is reduced by a factor of 3 in the presence of MTha ( Fig 1A right panel , 1B ) . Furthermore , IκBε , p100/p52 , RelB and IκBα exhibited an increased interaction with RelAp43 by a factor of 3 to 7 in the presence of MTha , RelB and IκBα becoming significantly more present . Hence , we can observe that MTha induced a strong reorganization of the interaction profiles of RelAp43 with the NF-κB and IκB proteins . Indeed , MTha led to an increased interaction between RelAp43 and RelB and p100/p52 , part of the non-canonical pathway and as well as a change in its association with the regulatory proteins IκBα , IκBß and p100 . Moreover , we identified 2 proteins that are known to form a regulatory complex with p105: ABIN2 and TPL2 ( S2 Fig ) , forming an intersection step between the NF-κB and the MAPK pathways [3] . The volcano plot ( Fig 1A left panel ) shows that aside from p105/p50 and p100/p52 , ABIN2 was one of the most significant proteins interacting with RelAp43 in the absence of MTha . However , ABIN2-RelAp43 interaction is reduced tenfold in the presence of MTha ( falling under the cutoff of 5% FDR ) conversely to p105/p50 which interaction remained similar ( Fig 1B , S2 ) . It is worth to note that the interaction of TPL2 with RelAp43 , was found to be increased by a factor of 4 in the presence of MTha ( Fig 1 right panel , 1B ) . Therefore , we decided to focus our work on the regulatory role of ABIN2 in a RNAT complex , and on deciphering how the activity of this complex could be modulated by the M protein of Tha virus . As TPL2 is known to regulate the processing of p105 and activation of NF-κB [5] , and as ABIN2 inhibits NF-κB pathway and stabilizes TPL2 , we looked in the presence of MTha for an evidence of the modification of RelAp43 interactions with p105 and p50 . To this aim , we used the samples purified by TAP and analyzed by MS . In the whole cell extract controls prior to the TAP ( input ) , the quantity of p50 revealed by western blot is equal or slightly higher than that of p105 ( Fig 1C ) regardless of the presence of RelAp43 and MTha . In the absence of MTha ( output ) , the quantity of p50 is inferior to the quantity of p105 when p105/p50 is purified alongside with RelAp43 . Interestingly , the quantity of p50 co-purified with RelAp43 is significantly higher in the presence of MTha , similar to the quantity of p105 . Altogether these results showed a role of MTha in the modulation of the composition of a RNAT complex , resulting in the formation of RelAp43-p50 dimers . To further confirm the existence of a RNAT complex , we used protein-protein interaction ( PPI ) assays such as bioluminescence resonance energy transfer ( BRET ) and protein-fragment complementation assay ( PCA ) . BRET enables identification in living cells of any interaction occurring below a 10 nm radius . STAT1 , a protein from the JAK-STAT pathway which is not interacting with p50 , was used as a negative control ( Fig 2A ) . As it interacts with RelAp43 but not with ABIN2 nor TPL2 ( Fig 2A ) , p50 was also used as a control . All negative controls were used to determine a threshold based on a three standard deviations of the mean ( 3SD ) . Hence , although the BRET signal of each interaction appeared to vary , with stronger signal for p105-RelAp43 , p50-RelAp43 and TPL2-ABIN2 ( netBRET>0 . 05 ) than p105-ABIN2 , p105-TPL2 , RelAp43-TPL2 and RelAp43-ABIN2 ( netBRET<0 . 05 ) , overall all combinations of interaction within the complex RNAT showed a positive result ( Fig 2A ) . We next study the interaction between the 4 RNAT partners using PCA ( Fig 2B–2G ) . In comparison to BRET , here only RelAp43-p105 and p105-TPL2 interactions ( stable logNLR above the 0 . 54 threshold ) were significant ( Fig 2B and 2C ) . This result can be explained by the high stringency required for the molecular complementation during PCA compared to the more flexible resonance energy transfer performed at nanometric distances with BRET . In order to study more precisely the hierarchical interactions within the RNAT complex , we used PCA in the context of another third overexpressed protein ( Fig 2B–2G ) . In these conditions , we found that the RelAp43-p105 interaction was not modified by ABIN2 nor TPL2 ( Fig 2B ) . Similarly , a high level of interaction was observed with p105-TPL2 ( Fig 2C ) , but in this case , it was significantly enhanced by RelAp43 ( logNLR of 1 . 99 instead of 1 . 67 with the control CAT ) . This is more in favor of a modification of the p105-TPL2 complex induced by RelAp43 which allows the generation of a higher luciferase signal . Again , ABIN2 had not effect on p105-TPL2 complex . Moreover , a stable interaction between ABIN2 and p105 or TPL2 , separately , could not be observed ( Fig 2D and 2E ) . However , ABIN2-p105 interaction ( Fig 2D ) was stabilised by RelAp43 ( logNLR = 0 . 65 ) and even more by TPL2 ( logNLR = 1 , 1 ) . In the case of the ABIN2-TPL2 interaction ( Fig 2E ) , it was only stabilized by p105 ( logNLR = 0 . 86 ) . Regarding RelAp43 , the collaboration of p105 was mandatory to stabilize its interaction with ABIN2 ( Fig 2F ) or TPL2 ( Fig 2G ) with a logNLR of 1 . 1 and 1 . 3 , respectively . This further establishes the crucial role of p105 , TPL2 and a second NF-κB protein such as RelAp43 , in the initiation of the RNAT complex . Altogether , the BRET results ( Fig 2A ) showing the strongest signal for ABIN2-TPL2 interaction ( netBRET = 0 . 093 ) and the PCA results ( Fig 2B–2G ) , demonstrate the participation of ABIN2 to the RNAT complex through its interaction with both TPL2 and p105 , p105 binding itself to RelAp43 , to form a quaternary RNAT complex . In order to confirm the role of rabies virus proteins in the modulation of the composition and activities of the RNAT complex and more specifically on interactions involving ABIN2 , we performed a PCA using 3 partners , as described in Fig 2B , but in the context of viral infection ( Fig 3 ) . To do so , we used the Tha and SAD viruses which M proteins were previously described as differentially targeting the NF-κB pathway [2] . In these conditions , we confirmed that the interaction between ABIN2 and RelAp43 is facilitated by p105 ( Fig 3A ) . However , the infection with Tha virus significantly ( p<0 . 05 ) decreased this interaction even in the presence of p105 , when compared to the mock infected cells ( logNLR dropped from 1 . 45 to 0 . 80 ) , confirming the results from MS . Regarding ABIN2 and TPL2 interaction ( Fig 3B ) , Tha virus infection reduced the interaction of the two proteins in absence of p105 compared to the control cells ( logNLR dropped from 1 . 11 to the level of the threshold set at 0 . 54 ) . This effect is even more striking in cells overexpressing p105 , where a strong and significant decrease of the interaction between ABIN2 and TPL2 is noticed ( from 1 . 52 and 1 . 34 in the mock , p<0 . 01 , and SAD , p<0 . 05 , infected cells to 0 . 80 logNLR with Tha , Fig 3B ) . In the case of ABIN2 and p105 , the interaction was decreased in Tha compared to mock infected cells , in absence ( 1 . 53 to 0 . 91 logNLR ) as well as in presence ( 1 . 92 to 1 . 18 logNLR ) of overexpression of TPL2 ( Fig 3C ) . Overall , this indicates that Tha virus can strongly disturb the interaction of ABIN2 with the other members of the RNAT complex . Next , we investigated the capacity of the M protein of Tha and SAD to interact with the RNAT complex using the BRET technology . Both STAT1 , which doesn’t interact with the M proteins , and the P protein , which doesn’t interact with the proteins RelAp43 , p105/p50 , ABIN2 or TPL2 , were used as negative controls to determine a 3SD threshold . Since the P protein can interact with phosphorylated STAT1 [20] , this pair was not considered as a negative control . None of the M proteins lead to a significant interaction ( higher than 3SD ) with STAT1 , p105 and p50 ( Fig 4A ) . However , MTha but not MSAD seems to interact with RelAp43 , which is expected [2] , but also with ABIN2 . Both , MTha and MSAD gave a strong positive signal with TPL2 ( netBRET > 0 . 05 ) . To further investigate the interactions between the M proteins and each of the members of the RNAT complex , we included a MTha protein mutated on the positions 77 , 100 , 104 , 110 ( MTh4M ) and showing a loss of interaction with RelAp43 ( Fig 4C ) [19] . Based on the absence of significant interaction with M proteins , p105 ( Fig 4B ) , p50 and STAT1 ( S4A Fig ) were used as negative controls . They did not show any significant BRET activity , even while using higher YFP:Nluc ratios ( S4B Fig ) , which are known to increase the potential efficiency of BRET [21 , 22] . At the opposite , the M proteins exhibited a strong increase of the BRET signal significantly above the controls with RelAp43 , TPL2 and ABIN2 ( Fig 4C–4E ) . Fig 4C confirmed the ability of interaction of MTha with RelAp43 ( netBRET ranging from 0 . 022 to 0 . 109 ) while both MTh4M and MSAD presented a very low netBRET if any ( ranging from 0 . 001 and 0 . 006 to 0 . 003 and 0 . 059 , respectively ) . Moreover , the observation of a significant difference at the 1:2 ratio , where the efficiency of BRET is low , is a strong evidence of this interaction . As expected , all M proteins interacted with TPL2 ( Fig 4D ) , although MTh4M showed for unknown reason a netBRET about half that of MTha and MSAD . More interestingly , Fig 4E highlights that MTha interaction with ABIN2 ( increasing from 0 . 042 to 0 . 154 of netBRET ) is strongly higher than that of MTh4M or MSAD ( increasing from 0 . 028 and 0 . 022 to 0 . 052 and 0 . 063 of netBRET , respectively ) . Overall , this shows that M protein interaction with the NF-κB pathway is more complex than previously described [2 , 19] . While we highlight that M proteins can also interact with TPL2 , a distinct interaction of the M protein from a field isolate virus with ABIN2—similarly to RelAp43—is shown for the first time . NF-κB signaling is involved in the induction of the immune response and its control by the M protein is already well established . While MSAD and MTh4M can induce NF-κB activity , MTha is able to strongly inhibit its activation [2 , 19] . Hence , we quantified the expression of various immunity-related genes regulated by NF-κB: IFNß , CXCL2 , TNF ( Fig 5A–5F ) , CCL5 and CXCL10 ( S5A–S5D Fig ) in the brain of mice from two genetic backgrounds [23 , 24] , at late stage of the infection by Tha or the isogenic Th4M virus mutated on the positions 77 , 100 , 104 , 110 of the M protein [19] . Tha and Th4M virus have similar replication rates in the brain of mice , both at 9 days after infection and at the experimental end point ( S6B Fig ) . IFNß , CXCL2 and TNF have already been shown to be modulated by the interaction of MTha with RelAp43 in cellulo [19] . After 9 days of infection , neither Tha nor Th4M induce the expression of IFNß , CXCL2 or TNF in BALB/c ( S6C–S6E Fig ) . However , at the experimental end point , both Th4M and Tha virus induce an increase of the expression of IFNß , CXCL2 and TNF in BALB/c mice ( Fig 5A–5C ) while almost no induction is observed in C57BL/6 mice infected by Tha virus compared to Th4M ( Fig 5D–5F ) . Indeed , Th4M virus induced a stronger increase of the expression of IFNß ( Fig 5A and 5D ) , CXCL2 ( Fig 5B and 5E ) and TNF ( Fig 5C and 5F ) compared with Tha ( 13- , 4- and 3 . 3-folds in BALB/c , respectively; 5 . 8- , 14- and 5 . 3-folds in C57BL/6 mice , respectively ) . In comparison , both Tha and Th4M viruses induce a strong expression of CCL5 and CXCL10 ( S5 Fig ) and no differences between the two viruses could be observed . Therefore , if CCL5 is overexpressed under SAD infection compared to Tha infected cells [19] , this pattern seems to be due to a higher stimulation of the immune response than an inhibition by MTha and its control together with that of CXCL10 must be strictly restricted to the N protein of pathogenic Ni virus [25] . Interestingly , late infection symptoms ( corresponding to experimental end point ) appeared in Tha infected mice within 10 days post-infection ( dpi ) regardless of the genetic background ( Fig 5G ) , while appearance of the late infection symptoms in Th4M infected mice was delayed by 4 to 8 days ( median of survival at 14 and 17 dpi in BALB/c and C57BL/6 , p<0 . 05 , respectively ) . Overall and although some differences exist between the two genetic background [23 , 24] , those results suggest that Tha virus could exert a repression of the expression of inflammatory related genes such as IFNß , CXCL2 and TNF in the brain of infected mice at the late stage of the infection compared with Th4M while not affecting CCL5 and CXCL10 . Moreover , the mutation on the positions 77 , 100 , 104 , 110 of the M protein of Tha virus , which leads to a stronger immune response to the infection , can also be correlated with a higher survival rate of the infected mice .
The capacity of the NF-κB pathway to control various responses derives from the differential regulation of a wide range of target genes through the formation of a broad NF-κB dimer repertoire [26] . A splicing variant of RelA , RelAp43 , was recently involved in the control of the immune response during rabies virus ( RABV ) infection [2 , 19] and forms complexes with all the NF-κB and IκB proteins . Amongst them , the regulatory proteins p105/p50 and p100/p52 are the most significant NF-κB proteins interacting with RelAp43 . While active NF-κB dimers are mainly composed of RelA-p50 dimers , RelA is also involved in the regulation of p100/p52 . The regulatory p100 precursor serves as an inhibitor of p65/RelA as RelA-p100 dimers are not active after TNF stimulation [27] and RelA-p52 dimers have been shown to be part of several signaling pathways [28] . Therefore , RelAp43 could act as an important competitor of RelA in both canonical and non-canonical pathways , according to the physiological context . In addition , all the partners of the ternary complex p105-ABIN2-TPL2 , regulating the activation of downstream NF-κB and MAPK pathways [3] are associated to RelAp43 within a close range ( < 10 nm ) . The mapping of PPIs showed TPL2 to be mainly responsible for the interaction of ABIN2 with the complex while co-immunoprecipitation experiments showed that ABIN2 preferentially forms a larger ternary complex including TPL2 and also p105 [3] . Interestingly , the C-ter region of TPL2 involved in the interaction with the processing inhibitory domain ( PID ) of p105 ( which also mediates p105 dimerization ) is also involved in the interaction with the 194–250 region of ABIN2 ( Fig 6A ) [3] . Hence , both ABIN2 and p105 interact with one singular domain of TPL2 . Investigating close proximity interactions , we shed a new light on the specific role of the different proteins in the complex ( Fig 6A ) . ABIN2 is the member of the complex exhibiting the weaker interaction with the others . Overall , this suggests that p105 and TPL2 might form together a region including the C-ter of TPL2 and the PID of p105 securing a stable interaction with ABIN2 . Finally , RelAp43 interacts with p105 most likely through the RHD [29] and helps its association with TPL2 and therefore of ABIN2 with the complex . All observed and reported interactions are summarized in S1 Table and Fig 6A . Interestingly , RelAp43 and ABIN2 are also able to interact with the help of p105 but without the overexpression of TPL2 . This suggests two possibilities . RelAp43 could allow this interaction without TPL2 as it exhibited a small signal of interaction between ABIN2 and p105 . Another explanation could be that the endogenous TPL2 is sufficient to bring ABIN2 to the RNAT complex . Altogether , this suggests that RelAp43 affects p105 and stabilize the formation of the complex . It is likely that while p105 is mandatory to bring ABIN2 to NF-κB dimers , several members of the NF-κB family can form a complex with p105-ABIN2-TPL2 [5 , 30] . Whether the capacity of p105 to form a complex with TPL2 and ABIN2 increases solely when dimerized with RelAp43 and its specific C-ter part or whether any other NF-κB proteins can contribute remains to be determined . However , crystallization of RelA-p50 dimers [29] suggests that the RelAp43 C-ter region should be nearby the PID region of p105 and therefore , close to the suggested pocket of interaction between p105 , ABIN2 and TPL2 ( Fig 6A ) . Here we confirmed that ABIN2 forms a stable interaction within close proximity only with both p105 and TPL2 . Moreover , while it has been suggested so far that the p105-ABIN2-TPL2 complex is separated from the main NF-κB pathway and uses a separate pool of p105 [3] , here we show that the assembly of this ternary complex can be favored by a second NF-κB protein such as RelAp43 and that all proteins interact below 10 nm of distance of each other as shown by BRET . The central role in cell homeostasis of the NF-κB pathway makes it a target of choice for viruses such as RABV to control the immune response and help it to silently spread within the host [10] . In the case of pathogenic lyssaviruses , such as the Thailand strain , the M protein interacts with RelAp43 and disturbs the homeostasis within the NF-κB dimers . MTha could lock RelAp43-dependent NF-κB dimers with IκBα in an inactive state , as well as favor its interaction with the non-canonical pathway . In the case of RelB , it has been well established that the formation of RelA-RelB dimers inhibit both RelA and RelB respective pathways by inhibition of the DNA binding [31–33] . Hence , it could be interesting to investigate the DNA binding capacity of a RelAp43-RelB dimer , as the enhancement of the interaction between RelAp43 and RelB by MTha could lead to a differential regulation of the canonical and non-canonical pathway . Although firstly described as having a most likely direct interaction with RelAp43 , here we show the capacity of MTha , but also in a lower extent of MSAD , to interact with a cluster of NF-κB signaling proteins ( S1 Table ) . Indeed , if the capacity of the M protein to interact with TPL2 is conserved in both attenuated and pathogenic viruses , only the M protein of a pathogenic virus seem able to significantly interact with both RelAp43 and ABIN2 . The M protein is then likely to first dock on TPL2 , and interact afterward with ABIN2 and/or RelAp43 , two properties that are lost in attenuated viruses . Therefore , this questions the effect of the single interaction of M with TPL2 during the infection by attenuated virus on cell signaling . However , as the capacity of MTha to disturb the NF-κB pathway was shown to depend on RelAp43 expression [19] , the interaction of MTha with only ABIN2 and TPL2 is not enough to disturb it . MTha significantly enhances the interaction of many proteins with RelAp43 complexes ( Figs 1 and S2 ) and interacts with 3 out of the 4 proteins of a single complex: RelAp43 , ABIN2 and TPL2 . Even if the M protein is known for having pleiotropic properties , the specific interaction of MTha with such a wide range of host proteins is highly unlikely . As the M protein of lyssaviruses is known to oligomerize [34] and form super structures within the cell [15] , we suggest that MTha can act as a scaffolding agent for the formation of high molecular weight complexes , targeting several cellular function at the same time ( S2 Fig ) . Additionally , a region of the M protein ( positions 33–36 ) was described as sharing the features of a proline-rich motif ( PRM ) and interacts with the position 107–112 on a second M protein and forming a hydrophobic cleft [34] . PRMs are common motifs amongst host protein-protein interactions and their hydrophobic properties could facilitate the binding to many host proteins . At this stage we observed that ( 1 ) the MS quantification of ABIN2 purified with RelAp43 is reduced by 10 fold in presence of MTha , which goes against the general trend of MTha enhancing interactions , ( 2 ) the MTha-ABIN2 interaction is strikingly stronger than that of MTh4M or MSAD and , ( 3 ) the Tha virus has a higher propensity to weaken the interaction of ABIN2 with RelAp43 , TPL2 and p105 compared to the SAD virus and mock infected cells . This leads strongly to the hypothesis that MTha disturb the RNAT complex , weakening the capacity of interaction between ABIN2 and the hypothesized region of interaction formed by p105 , TPL2 and maybe RelAp43 ( Fig 6A ) . Interestingly , our results demonstrate a switch of the interaction of the M protein between a pathogenic and a vaccinal strain . Further investigation should also focus on ABIN2 and its role as an inhibitor of RIP1 , upstream of NEMO ( which were not identified by MS ) and the NF-κB pathway [35] . A working hypothesis is that ABIN2 could be recruited to polyubiquitin chains when it is released from activated TPL2 , restricting the activation of innate immune signaling networks [36] . Therefore , as we have shown a strong interaction of MTha to ABIN2 , MTha could interfere with the NF-κB signaling at several levels through its interaction with ABIN2 . Furthermore , ABIN2 has been linked to the ESCRT pathway [37] , and its interaction with the M protein could be related to RABV budding [38] . While destabilizing the RNAT complex leading to the exclusion ABIN2 , MTha controls downstream NF-κB and MAPK pathways . Indeed , MTha facilitates the formation of RelAp43-p50 NF-κB dimers lacking a TAD which can regulate the expression of NF-κB targets [19] . To do so , MTha either induce directly the processing of p105 into p50 in a RelAp43 dependent manner , or modify the homeostasis of NF-κB dimers [5] . This remains to be clarified . It is worth to mention that the production and/or release of p50 from the cytoplasm and translocation of an active NF-κB dimer ( not including RelA ) was observed under infection by the laboratory strain CVS [39] . Together with our results , it shows that p50 could have an important role in rabies virus infection that should be further investigated . In parallel , MS results suggest that TPL2 is being stabilized without ABIN2 , implying that MTha could lock TPL2 on p105 ( Fig 6 , S1 Table ) . Yet , the implications on TPL2 remain to be investigated using time sensitive approaches to determine if and in which order MTha destabilizes , liberates or blocks TPL2 as well as the effects on downstream activity on MAPK ERK1/2 and MAPK-dependent transcription factors [3] . Hence , in addition to P protein control of IRF and JAK-STAT pathways during RABV infection [10] , M appears to have a central function in cell signaling inhibition and modulation of innate immune response through the control of TPL2 , a key regulator of NF-κB and MAPK pathways . It is worth to note that TPL2 regulates also IRF proteins , leading to a strong induction of IFNβ during vesicular stomatitis virus ( VSV ) infection [40] and appears downstream of JAK-STAT pathway , holding an essential role in antiviral host defense against influenza virus infection [41] . Therefore , the perturbation of TPL2 signaling by M could be potentially implicated in a much further control of the host response , which could be conserved across pathogenic and attenuated viruses . The NF-κB pathway is involved in the expression of several genes of the immune and inflammation response to the infection such as IFNβ , CXCL2 and TNF [8 , 42] . The modulation of their expression by Tha virus depending on the capacity of the M protein to interact with RelAp43 [19] , as well as ABIN2 , is here confirmed in Tha infected mice . Interestingly , IFNβ , CXCL2 and TNF expression control by the M protein is observed in the late stages of the infection in vivo , further corroborating the modulation by MTha of the immune response observed in cellulo [19] . Further , it would be interesting to explore the effect of MTha hijacking of the NF-κB signaling on other NF-κB-dependent genes , possibly involved in RABV pathogenesis . Additionally , the control of the expression of such genes mainly by the NF-κB signaling pathway , but also potentially in synergy with other pathways , correlates with the incubation and pathogenicity of RABV . Hence , through the regulation of the NF-κB pathway by the M protein , Tha virus is able to fine-tune the host response to the infection . RelAp43 and the matrix protein of pathogenic rabies virus are known to interact and modulate the innate immune response [2 , 19] . However , we show here that the mechanisms involved are far more complex than initially thought . MTha , and to a lower extent MSAD , can interact with RelAp43 and other proteins involved in NF-κB as well as MAPK signaling . While introducing a new insight in the NF-κB network and especially in a p105-regulating complex , we gathered for the first time evidences suggesting that a pathogen , rabies virus , can use one of its proteins ( the M protein ) to modulate the interaction of ABIN2 with p105 and TPL2 . Thereby , the M protein leads to the control the NF-κB pathway in order to modulate the inflammation in response to RABV infection .
Human carcinoma epithelial cells ( HeLa , ATCC CCL2™ ) , human epithelial kidney cells ( HEK-293T/17 , ATCC CRL-11268™ ) are part of the collection of our laboratory and were cultured as previously described [17] . Virus infection was performed in 6- or 96-well plate dishes during indicated times at 37°C and using different viruses at a multiplicity of infection ( MOI ) of 1 . Thailand virus , referred as Tha ( isolate 8743THA ) , is a field strain of RABV isolated in Thailand from a human bitten by a dog ( EVAg collection , Ref-SKU: 014V-02106 ) . SAD-B19 virus ( SAD ) is a vaccine strains of RABV ( EVAg collection , Ref-SKU: 014V-02283 ) . A recombinant Tha virus mutated on the positions 77 , 100 , 104 , 110 of the matrix protein ( Th4M ) was used as previously described [19] . The coding sequence ( CDS ) of CAT , MTha , MSAD , MTh4M , RelA and RelAp43 were obtained from pcDNA3 . 1N-V5-dest plasmids in the laboratory [2] . The CDS of NFκB1 ( p105/p50 ) was given by R . Weill ( Institut Pasteur ) . The CDS of TPL2 and ABIN2 were amplified by PCR from cDNA obtained from HeLa cells . The CDS were compared to the RefSeq of their respective variant number 1: NM_005204 . 3 and NM_024309 . 3 ( NCBI ) . Genes were inserted using In-Fusion technology ( Clontech ) in various plasmids . pKmyc vector was a gift from Ian Macara ( Addgene plasmid #19400 ) to add the c-Myc tag in N-ter position . Gaussia Luciferase-Based Protein Complementation Assay ( PCA ) plasmids to add a Glu1 or Glu2 tag in N- or C-ter position of the insert were a gift from D . Gerlier [43] . A modified version of the pEGFP-C1 plasmid ( Promega ) into p3xFLAG-EGFP-C1 , presenting a 3xFLAG tag upstream to the eGFP[44] was a gift from F . Thierry ( Institute of Medical Biology , Singapore ) . It adds a FG ( 3xFLAG-eGFP ) tag in N-ter position . The pEYFP-C1/N1 plasmids ( Promega ) were commercially purchased , adding a eYFP tag in N- or C-ter position . The pEYFP-C1/N1 plasmids were modified by switching the eYFP to the Nano-Luciferase ( Nluc ) to obtain new pNluc-C1/N1 plasmids , adding a Nluc tag in N- or C-ter position . All sequences were controlled by sequencing , using Sanger technology . All constructions are summarized in S2 Table . Finally , the plasmid pGL4 . 50 ( Promega ) expressing the Firefly Luciferase was used as control . HeLa cells were plated onto 75 cm2 dishes with 2 . 106 cells per dish in 15 mL of medium ( 10 dishes per condition ) . After 24h , cells were transfected with 6 μg of FG or FG-RelAp43 plasmids and 4 μg of V5-CAT or V5-MTha plasmids using Lipofectamine 2000 ( Invitrogen ) . Cell pellets were lysed 24h later in a FLAG Buffer ( 150 mM Tris-HCl , 300 mM NaCl , 1% Triton-100X ) . 30 mg of protein extract were incubated overnight at 4°C on anti-FLAG-M2 beads ( Sigma ) in FLAG Buffer . Protein complexes were then eluted with 3xFLAG purified peptides ( Sigma ) and incubated for 1 h with GFP-Trap_A ( Chromotek ) in a GFP Buffer ( 50 mM Tris-HCl pH 7 . 4 , 150 mM NaCl , 0 . 2 mM EDTA ) . Proteins were finally eluted directly using loading sample buffer ( Invitrogen ) and heated at 95°C for 10 min . After centrifugation , a quality control of the eluted proteins was performed by western blot . Each condition was done in 3 biological replicates . dx . doi . org/10 . 17504/protocols . io . jeqcjdw [PROTOCOL DOI] . Protein samples were loaded on a SDS-PAGE gel ( 4–12% gradient , Biorad ) . After the electrophoretic migration the gel was stained with Coomassie Blue R-250 ( Biorad ) and each lane was cut into 10 gel bands . Gel slices were washed twice with 100 mM ammonium bicarbonate for 15 min , followed by 100 mM ammonium bicarbonate/acetonitrile ( 1:1 ) for 15 min . After reduction and alkylation , proteins were digested by 0 . 5 μg of modified sequencing grade trypsin ( Promega , Madison , WI , USA ) in 10 mM ammonium bicarbonate overnight at 37°C . Resulting peptides were extracted from the gel by incubation in 50 mM ammonium bicarbonate for 15 min , and three times in 5% formic acid ( FA ) and 50% acetronitrile ( ACN ) for 15 min . All extractions were pooled and dried down in a vacuum concentrator , and further resuspended in 2% acetonitrile , 0 . 1% FA before injection . Trypsin-digested peptides obtained for all gel slices were analyzed separately by nanoLC-MS/MS using an UltiMate 3000 RSLC ( Dionex , Amsterdam , The Netherlands ) coupled to an LTQ-Orbitrap Velos mass spectrometer ( Thermo Fisher scientific , Bremen , Germany ) . Five μL of each sample were loaded on a C18 pre-column ( 300 μm inner diameter × 5 mm; Dionex ) at 30 μL/min in 2% ACN , 0 . 1% FA . After 4 min of desalting , the pre-column was switched online with the 15 cm capillary column ( 75 μm diameter filled with 3 μm Reprosil-Pur Basic C18-HD resin ) ( Dr . Maisch GmbH , Ammerbuch-Entringen , Germany ) equilibrated in 98% solvent A ( 2% ACN , 0 . 1% FA ) and 2% solvent B ( 80% ACN , 0 . 08% FA ) . Peptides were eluted using a 2 to 55% gradient of solvent B during 30 min at 300 nL/min . The LTQ-Orbitrap Velos was operated in data-dependent acquisition mode with the XCalibur software . Survey scan MS were acquired in the Orbitrap in the 300–2000 m/z range with the resolution set to a value of 60 , 000 at m/z = 400 . The 10 most intense ions per survey scan were selected for collision-induced dissociation ( CID ) , and resulting fragments were analyzed in the linear trap ( LTQ ) . Dynamic exclusion was employed within 20 s and repeated during 30 s to prevent repetitive selection of the same peptide . Raw files were processed with Maxquant [45] ( v . 1 . 4 . 1 . 2 ) and the Human Swiss-Prot FASTA database ( 20 , 240 proteins ) concatenated with 4 recombinant proteins ( FG , FG-RelAp43 , V5-CAT and V5-Mtha ) was used . Andromeda searches [46] were performed choosing trypsin as specific enzyme with a maximum number of 2 missed cleavages . Possible modifications included carbamidomethylation ( Cys , fixed ) , oxidation ( Met , variable ) and Nter acetylation ( variable ) . The mass tolerance in MS was set to 20 ppm for the first search then 6 ppm for the main search and 0 . 5 Da for MS/MS . Additional peptides were identified by the “match between run” option with a maximal retention time window of 1 min . Five amino acids were required as minimum peptide length and 1 unique peptide was required for protein identification . A false discovery rate ( FDR ) cutoff of 1% was applied at the peptide and protein levels . MaxLFQ , Maxquant’s label-free quantification ( LFQ ) algorithm was used to calculate protein intensity profiles across samples [47] . A minimum peptide ratio count of 2 was required for LFQ calculation ( S2A Fig ) . For statistical and bioinformatics analysis , as well as for visualization , Perseus , which is part of Maxquant , was used [48] . The “proteinGroup . txt” file was processed as described in Tyanova et al . 2016 . Two valid LFQ values out of three were required for a confident quantification across all replicates . Protein LFQ intensities were logarithmised and missing values imputed by values simulating noise around the detection limit . For pairwise comparison and identification of interacting proteins , t-test statistics were applied with a permutation-based FDR set to 5% and a S0 of 2 [49] . Protein-protein interaction networks for genes of interest were obtained using STRING v10 [50] . Interactions were determined with the following sources: experimentally determined , automated text mining and database annotations ( minimum score: 0 . 9 ) . Finally , the network was visualized with Cytoscape [51] and clusters were annotated using DAVID v6 . 8 [52 , 53] . Western blot analysis was performed using NuPAGE gels ( Invitrogen ) . Protein transfer on nitrocellulose membrane was performed using iBlot transfer system ( Invitrogen ) , as indicated by provider . Membranes were saturated for 1 h in PBS-Tween 0 . 1% with 5% non-fat dried milk . Immunoblotting procedure consisted in overnight incubation with indicated primary antibody diluted in 5% dried milk PBS-Tween , washed three times for 5 min in PBS-Tween , then incubated 1 h with indicated HRP conjugated secondary antibody . The following antibodies were used: mouse a-V5 antibody ( Invitrogen ) ; mouse a-FLAG M2 antibody ( Sigma ) , mouse a-p50 antibody ( SantaCruz ) , HRP-linked a-mouse antibody and HRP-linked a-rabbit antibody ( GE Healthcare ) . Blots were revealed by chemiluminescence and exposure to X-ray films or with an imager ( Amersham ) for different time to avoid saturation . HEK-293T cells were plated in 96-well plates with 25 000 cells per well in 100 μL of culture medium . After 24h , cells were transfected with Lipofectamine 2000 with 2 plasmids , each of them expressing either Glu1 or Glu2 ( N- or C-ter tagged ) recombinant protein , a plasmid expressing a cMyc-tagged protein and a control plasmid expressing the Firefly Luciferase . If specified , cells were infected 3 hours before transfection . After cell lysis , the Gaussia and Firefly activity were measured separately using respectively the Renilla Luciferase kit and the Firefly Luciferase kit ( Promega ) . After a first normalization with the Firefly activity , Normalized Luminescence Ratio ( NLR ) was obtained from split Gaussia luciferase’s activity using the following formula ( 1 ) , as previously described [54] . The PCA efficiency for each couple ( Glu1 or Glu2 in N- or C-ter ) was assessed in a prior experiment and the best combinations were selected further investigation . dx . doi . org/10 . 17504/protocols . io . jekcjcw [PROTOCOL DOI] . HEK-293T cells were plated in 384 plates with 3000 cells per well in 50 μL of medium and transfected 3h later using FuGENE6 ( Promega ) with a total of 25 ng of plasmids expressing YFP- and Nluc-tagged proteins at a ratio of 1:1 . When specified , cells were transfected with different ratios of DNA ( from 1 YFP:2 Nluc to 4 YFP:1 Nluc ) per condition . Direct bioluminescence from the donor ( Nluc ) and the acceptor ( YFP , noted NlucY ) was measured 48h later using the Wallac 1420 VICTOR 3V multilabel plate reader ( PerkinElmer ) . Next , the energy transfer between the Nluc and YFP ( BRET ) was calculated according to the following formulas ( 2 and 3 ) and normalized to a YFP-Nluc linked recombinant protein [55] . All combinations were assessed and combinations producing more than 0 . 05 of netBRET ( with an N-ter tag if possible ) were selected for further investigation . A threshold of specific interaction was determined using the mean+3SD ( simplified 3SD ) of the negative controls within each experiment . dx . doi . org/10 . 17504/protocols . io . jepcjdn [PROTOCOL DOI] . All mice experiments were performed in accordance with guidelines of the European and French guidelines ( Directive 86/609/CEE and Decree 87–848 of 19 October 1987 ) and the Institut Pasteur Safety , Animal Care and Use Committee , and approved by the French Administration ( Ministère de l’Enseignement et de la Recherche ) under the number O522-02 . All animals were handled in strict accordance with good animal practice . Three-weeks old BALB/c or C57BL/6 ( Charles River ) were infected by intramuscular injection of 1000 focus-forming units ( FFU ) and monitored over 21 days . Mice were sacrificed at 9 days post infection or upon the apparition of late infection symptoms ( humane end point ) . The infection was confirmed by RT-qPCR . Total RNA was isolated using trizol . Reverse transcription was performed on 1 , 2 mg of RNA using Superscript II ( Invitrogen ) with 2 pmol of oligodT primers ( Fermentas ) in a final volume of 20 μL . Transcription analysis was performed on 100 ng of total RNA using Taqman Power SYBR Green ( Applied Biosystems ) in a 7500 instrument ( Applied Biosystems ) and Quantitect primers ( Qiagen ) , following manufacturer instruction . Relative quantification was performed using GAPDH gene as endogenous control gene . Results were analyzed using 7500 SDS software v2 ( Applied Biosystems ) . Multiple comparisons of data were performed by ANOVA using the GraphPad Prism software .
|
Rabies virus is a recurring zoonosis responsible for about 60 , 000 deaths per year . A key feature of rabies virus is its stealth , allowing it to spread within the host and escape the immune response . To do so , rabies virus developed several mechanisms , including a thorough interference with cell signaling pathways . Here , we focused our attention on the molecular aspects of rabies virus escape to the NF-κB pathway through the interaction between the M protein and the NF-κB protein RelAp43 . Monitoring close range interactions , we found that RelAp43 plays an important role in the stabilization of the p105-ABIN2-TPL2 complex , which is essential in the regulation of both NF-κB and MAPK pathways , and we brought a new insight on the dynamics within the host protein complex . These results were confirmed in living cells and in mice . Overall , our data suggest that rabies virus interference with the p105-ABIN2-TPL2 complex is a cornerstone of its stealth strategy to escape the immune response .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Material",
"and",
"methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"luciferase",
"protein",
"interactions",
"pathology",
"and",
"laboratory",
"medicine",
"enzymes",
"pathogens",
"protein",
"interaction",
"networks",
"immunology",
"microbiology",
"enzymology",
"viruses",
"rna",
"viruses",
"network",
"analysis",
"mapk",
"signaling",
"cascades",
"rabies",
"virus",
"computer",
"and",
"information",
"sciences",
"proteins",
"medical",
"microbiology",
"gene",
"expression",
"microbial",
"pathogens",
"oxidoreductases",
"proteomics",
"immune",
"response",
"biochemistry",
"signal",
"transduction",
"lyssavirus",
"cell",
"biology",
"viral",
"pathogens",
"genetics",
"biology",
"and",
"life",
"sciences",
"cell",
"signaling",
"organisms",
"signaling",
"cascades"
] |
2017
|
Regulation of NF-κB by the p105-ABIN2-TPL2 complex and RelAp43 during rabies virus infection
|
Prompt therapy with high-dose intravenous benzylpenicillin for a prolonged period is critical for neurosyphilis patients to avoid irreversible sequelae . However , life-threatening neutropenia has been reported as a complication of prolonged therapy with high doses of benzylpenicillin when treating other diseases . This study aimed to investigate the incidence , presentation , management and prognosis of benzylpenicillin-induced neutropenia in treating neurosyphilis based on a large sample of syphilis patients in Shanghai . Between 1st January 2013 and 31st December 2015 , 1367 patients with neurosyphilis were treated with benzylpenicillin , 578 of whom were eligible for recruitment to this study . Among patients without medical co-morbidities , the total incidence of benzylpenicillin-induced neutropenia and severe neutropenia was 2 . 42% ( 95% CI: 1 . 38–4 . 13% ) and 0 . 35% ( 95% CI: 0 . 06–1 . 39% ) , respectively . The treatment duration before onset of neutropenia ranged from 10 to 14 days , with a total cumulative dose of between 240 and 324 megaunits of benzylpenicillin . Neutropenia was accompanied by symptoms of chills and fever ( 5 patients ) , fatigue ( 2 patients ) , cough ( 1 patient ) , sore throat ( 1 patient ) , diarrhea ( 1 patient ) and erythematous rash ( 1 patient ) . The severity of neutropenia was not associated with age , gender or type of neurosyphilis ( p>0 . 05 ) . Neutropenia , even when severe , was often tolerated and normalized within one week . A more serious neutropenia did not occur when reinstituting benzylpenicillin in patients with mild or moderate neutropenia nor when ceftriaxone was used three months after patients had previously experienced severe neutropenia . Benzylpenicillin-induced neutropenia was uncommon in our cohort of patients . Continuation of therapy was possible with intensive surveillance for those with mild or moderate neutropenia . For severe neutropenia , it is not essential to aggressively use hematopoietic growth factors or broad-spectrum antibiotics for patients in good physical condition after withdrawing anti-neurosyphilis regimen . We did not see an exacerbation of neutropenia in patients with the readministration of benzylpenicillin .
Neutropenia is a condition marked by an absolute neutrophil count ( ANC ) below 1 . 5×109/L in adults [1] , which can be further categorized as mild ( 1×109/L≤ANC<1 . 5×109/L ) , moderate ( 0 . 5×109/L≤ANC<1×109/L ) and severe type ( ANC<0 . 5×109/L ) [1 , 2] . There are many causes including drug-induced neutropenia [2 , 3] . Benzylpenicillin-induced neutropenia , a complication of prolonged therapy with high doses , has been well documented when treating infective endocarditis , leading some patients to withdraw necessary treatment and even undergo insidious life-threatening sepsis [4–7] . Syphilis has returned to china with a vengeance in the 21st century [8 , 9] . The epidemiology of neurosyphilis ( NS ) has largely mirrored that of early infective syphilis [10] . Prompt therapy of NS is critical for avoiding irreversible sequelae such as general paresis and tabes dorsalis [11] . The current recommended regimen is high-dose intravenous benzylpenicillin ( 18 to 24 megaunits daily ) for a prolonged period ( 10 to 14 days ) [12 , 13] . It is worth considering how to balance the benefit of treating NS with benzylpenicillin and harm if drug-induced neutropenia arises . We analyzed the clinical data of NS patients during three continuous years in order to investigate the incidence , presentation , management and prognosis of benzylpenicillin-induced neutropenia in order to provide helpful experience for other regions with a high burden of syphilis .
This retrospective study was approved by the medical ethics committee of the Shanghai Skin Disease Hospital , and conducted according to the principles expressed in the Declaration of Helsinki at the Sexually Transmitted Disease Institute of the Shanghai Skin Disease Hospital from January 1 , 2013 to December 31 , 2015 . We recruited NS patients who ( 1 ) underwent their first therapy of high-dose intravenous benzylpenicillin , ( 2 ) did not have a recent history of other infections ( e . g . : viral , bacterial , protozoal ) , ( 3 ) denied a past and family medical history of autoimmune diseases , underlying hematological diseases , nutritional deficiencies , splenic sequestration or congenital leukopenia , ( 4 ) did not receive chemotherapy , radiotherapy , immunotherapy , oral /intravenous /intramuscular usage of antibiotics , or other new medications in the past three months [14 , 15] , ( 5 ) had no history of alcohol abuse , and ( 6 ) had negative HIV status . Patients were excluded if they were under 18 years of age or had a pre-treatment complete blood count ( CBC ) outside the normal reference range . Written informed consent was obtained before the laboratory test and NS treatment for clinical care and research . NS was defined as having ( 1 ) any stage of syphilis , ( 2 ) a reactive cerebrospinal fluid-venereal disease research laboratory ( CSF-VDRL ) , and/or ( 3 ) an elevated CSF-protein ( >50 mg/dL ) or pleocytosis ( >10 white blood cells/μL ) in the absence of other known causes of the abnormalities [12 , 13] . Neutropenia was further categorized as mild ( 1×109/L≤ANC<1 . 5×109/L ) , moderate ( 0 . 5×109/L≤ANC<1×109/L ) and severe ( ANC<0 . 5×109/L ) as indicated above [1 , 2] . A CBC , urinalysis , routine stool studies for infection and occult blood , biochemical profile , electrolytes , chest radiography and electrocardiograph were performed in all patients before benzylpenicillin therapy . CBC monitoring was performed every other day for mild or moderate neutropenia and every day for severe neutropenia until the value normalized . Other essential tests , including blood culture , sputum culture , biochemical profile , or virus antibody , were also performed when neutropenia occurred . The NS treatment regimen was 4 megaunits of benzylpenicillin as a freshly prepared bolus and slow infusion intravenously every 4 hours for 14 days [12 , 13] . Clinical data were recorded in terms of age , gender , diagnosis , cumulative dose of benzylpenicillin , days to onset of neutropenia , accompanying symptoms when ANC nadir occurred , clinical management , recovery time and readministration of benzylpenicillin . All data were independently double-coded with Epidata software ( version 3 . 1; Denmark ) , then transferred into SPSS software ( version 18 . 0; Chicago , IL , USA ) for analyses . Descriptive statistics were used to calculate median , percentage , and incidence with 95% confidence interval ( CI ) . A chi-square test ( p<0 . 05 indicating statistical significance ) was applied to analyze the potential factors associated with neutropenia . The continuous variable "age" was categorized into two subgroups , including age<55 years and age≥55 years . Multivariate logistic regression was used to further identify factors independently associated with neutropenia when significant factors were found by chi-square test .
A total of 1 , 367 NS patients were treated with a standard regimen of benzylpenicillin during the study period , 613 of whom received treatment for the first time . Of these , 578 patients underwent repeat CBC during the treatment and were included according to the study criteria . Fourteen patients , all of whom had prior normal CBCs , had a repeat ANC below 1 . 5×109/L . The median age of these patients was 55 years ( range: 27 to 79 ) . Nine were male , and 12 had neurologic complications with a diverse spectrum of diagnoses , including syphilitic meningitis and parenchymatous neurosyphilis . ( Fig 1 , Tables 1 and 2 ) The total incidence of benzylpenicillin-induced neutropenia was 2 . 42% ( 95% CI: 1 . 38–4 . 13% , 14/578 ) among this cohort of patients with NS . Mild neutropenia was observed in 1 . 56% ( 95% CI: 0 . 76–3 . 04% , 9/578 ) , moderate neutropenia in 0 . 52% ( 95% CI: 0 . 13–1 . 64% , 3/578 ) , and severe neutropenia in 0 . 35% ( 95% CI: 0 . 06–1 . 39% , 2/578 ) of patients . The severity of neutropenia had no association with age , gender or type of neurosyphilis ( p>0 . 05 ) . The multivariate logistic regression was not carried out since no significant factors were found by chi-square test . For the majority ( 13/14 ) of patients , the duration of treatment before onset of neutropenia ranged from 10 to 14 days , and the cumulative dose of benzylpenicillin varied from 240 to 324 megaunits . A single patient received 120 megaunits over five days of treatment . The range of nadir total white blood cell ( WBC ) counts was 0 . 60 to 3 . 69×109/L , with nadir ANC from 0 . 04 to 1 . 49×109/L . Three patients had concurrent thrombocytopenia . The accompanying symptoms were chills and fever ( 38 . 5–39 . 7℃ , 5 patients ) , fatigue ( 2 patients ) , cough ( 1 patient ) , sore throat ( 1 patient ) , diarrhea ( 1 patient ) and erythematous rash ( 1 patient ) . Blood , sputum and throat swab cultures did not reveal an infectious etiology ( e . g . : bacterium and fungus ) among the febrile patients . ( Table 2 ) One patient with syphilitic meningitis , ocular and otic syphilis ( case 12 ) had an itchy rash on the trunk on day 4 and fever ( maximum 38 . 5℃ ) on day 5 . Repeat laboratory testing revealed that his ANC declined to 1 . 13×109/L . Thus , antihistamine and methylprednisolone ( 40mg daily ) were commenced instead of benzylpenicillin . His symptoms and CBC count normalized on day 8 . Subsequently , an alternative regimen of intravenous ceftriaxone ( 1 . 0 g every 12 hours for 15 days ) ( 13 ) was reinstituted uneventfully three months later . ( Table 2 ) Another two patients ( case 1 with general paresis and case 13 with syphilitic meningitis ) did well until they had fever , and repeat CBC revealed thrombocytopenia and severe neutropenia ( ANC of case 1: 0 . 04×109/L; ANC of case 13: 0 . 21×109/L ) near the end of therapy . Benzylpenicillin was discontinued . Despite mild symptoms , their fever in the context of a severe neutropenia caused a high level of concern for underlying life-threatening infection , and both patients were transferred to the emergency department . The results of bone marrow examination , Coombs' test , and cytomegalovirus and rubella virus IgM antibodies were not significant . They were given symptomatic relief and supportive treatment rather than human granulocyte colony-stimulating factor , glucocorticosteroid or other prophylactic broad-spectrum antibiotics . Both patients' CBCs returned to normal within four and five days , respectively , after withdrawing benzylpenicillin . Initiation of intravenous ceftriaxone did not induce neutropenia three months later in either patient . ( Table 2 ) The other 11 patients with mild or moderate neutropenia finished the 14-day therapy with close monitoring of CBC and observation for sequelae of neutropenia . None experienced any severe complication of therapy and all had recovery of a normal ANC within seven days . Some of these patients received a second round of therapy with benzylpenicillin three months later and either had no neutropenia or experienced similar neutropenia without symptoms . ( Table 2 )
The neutrophil is the most abundant WBC in the peripheral blood and plays a critical role in preventing infections as part of the innate immune system [16] . It has been documented that the offending medications associated with severe neutropenia are methimazole , ticlopidine , clozapine , sulfasalazine , trimethoprim-sulfamethoxazole and dipyrone in descending order of likelihood [17–19] . The hematologic complication of hypersensitivity to penicillin is rare , with an overall acute neutropenia of 2 . 4 to 15 . 4 cases per million populations over the last 20 years [14] . Based on this large clinical dataset , we concluded a total incidence of 2 . 42% among healthy NS patients receiving benzylpenicillin . It is noteworthy that the likelihood of acute neutropenia caused by intravenous benzylpenicillin for NS is much higher than that by penicillin for other diseases [14] . Penicillin agents are thought to be able to cause granulopoiesis inhibition [20 , 21] , and benzylpenicillin-induced neutropenia is dose related more than a pure immunological reaction [15 , 22] . As indicated earlier , the duration of beta-lactam therapy prior to the start of neutropenia always exceeded 15 days [21] . We saw cases of neutropenia caused by benzylpenicillin within 14 days probably due to the higher daily dose used for NS than for other diseases . In the 1980s , Al-Hadramy and his colleagues [6] summarized 28 reported cases of benzylpenicillin-induced neutropenia for diseases such as infective endocarditis , bowel obstruction , cellulitis , gangrenous appendix , pneumonia , hemangioma , septic arthritis , and pleural empyema . Therein , 71% patients developed neutropenia after taking 200 megaunits or more , and neutropenia developed in 82% of patients on treatment for two or more weeks , which is consistent with our findings of neutropenia being associated with high-dose and prolonged treatment [6] . Some studies have proposed the hypothesis that genetic and epigenetic modifications predispose an individual to idiosyncratic drug sensitivity [23 , 24] . The genetic susceptibility might be associated with an increased risk of neutropenia induced by high-dose benzylpenicillin which needs to be further investigated . According to previous reports , acute neutropenia was often well tolerated and normalized rapidly [2] . In our study , fever accompanied by general malaise was the first and often the only manifestation in patients . No patients experienced life-threatening complications . Withdrawing benzylpenicillin rapidly led to a recovery in the patient who had an ANC of 0 . 04×109/L at nadir but no other high-risk symptoms . Even though potential antibody cross-reactivity existed , we found that it was relatively safe when benzylpenicillin was reinstituted in patients with mild or moderate neutropenia , and ceftriaxone in patients with severe neutropenia , three months later . Previous research has identified that older age ( >65 years ) , septicemia or shock , metabolic disorders such as renal failure , and an ANC under 0 . 1×109/L were poor prognostic factors associated with drug-induced neutropenia [14] . Thus , in patients with these factors , the empirical use of hematopoietic growth factors , glucocorticosteroid and/or broad-spectrum antibiotics may positively impact the prognosis [18] . Among the 14 NS patients with acute neutropenia in this study , none had metabolic disorders , severe infections , septicemia or septic shock . No patients were given hematopoietic growth factors or broad-spectrum antibiotics , even though three patients were older than 65 years , and one patient had an ANC of 0 . 04×109/L . We also found the severity of neutropenia had no significant association with age , gender or the type of NS . Syphilis is far from eradicated , especially in the resource-limited areas worldwide , and it can affect any part of the neuraxis at any stage of infection [25 , 26] . There is a growing consensus that NS patients can benefit from regular benzylpenicillin therapy , and high-dose benzylpenicillin is of proven efficacy at the early stage of NS [27] . Here , we outlined benzylpenicillin-induced neutropenia as a complication of NS treatment . Some limitations should be acknowledged . First , due to limited published data on when to obtain surveillance CBCs during treatment , we arranged the first repeat CBC on day 10 unless any clinical symptom occurred beforehand . Thus , asymptomatic neutropenia may have been present in the two patients ( case 8 and 9 ) earlier than day 10 . Second , prompt NS therapy was limited to patients whose other medical conditions ( e . g . uncontrolled hypertension or diabetes ) were stable in order to minimize risk of therapy . Meanwhile , HIV co-infected patients were not included in the analysis because of possible confounding of leukopenia caused by HIV . These factors might limit the generalizability of our findings . In conclusion , benzylpenicillin-induced neutropenia was well tolerated in our cohort of patients with mild or moderate type . It also normalized rapidly without aggressive management for those with severe neutropenia after withdrawing anti-neurosyphilis regimen . We did not see an exacerbation of neutropenia in patients with the readministration of benzylpenicillin .
|
High-dose intravenous benzylpenicillin is an effective treatment for neurosyphilis although it can cause potentially life-threatening drug-induced neutropenia . We investigated the incidence , presentation , management and prognosis of benzylpenicillin-induced neutropenia among neurosyphilis patients treated over a three year period at the Shanghai Skin Disease Hospital . We recruited 578 patients with neurosyphilis who received benzylpenicillin ( 4 megaunits intravenously every 4 hours for 14 days ) according to strict study criteria . For patients without medical co-morbidities , the total incidence of benzylpenicillin-induced neutropenia was 2 . 42% ( 95% CI: 1 . 38–4 . 13% ) . The incidence of mild , moderate and severe neutropenia was1 . 56% ( 95% CI: 0 . 76–3 . 04% ) , 0 . 52% ( 95% CI: 0 . 13–1 . 64% ) , and 0 . 35% ( 95% CI: 0 . 06–1 . 39% ) , respectively . The duration of therapy given before the onset of neutropenia ranged from 10 to 14 days , and cumulative doses of benzylpenicillin varied from 240 to 324 megaunits . The accompanying symptoms were tolerated and often normalized within one week under close monitoring of blood counts . Therefore , benzylpenicillin can be continued with surveillance in the presence of mild or moderate neutropenia . Aggressive management is not essential for patients with severe neutropenia in good physical condition after withdrawing anti-neurosyphilis regimen . We did not see an exacerbation of neutropenia in patients with the readministration of benzylpenicillin .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"blood",
"cells",
"inflammatory",
"diseases",
"urology",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"body",
"fluids",
"pathology",
"and",
"laboratory",
"medicine",
"maternal",
"health",
"blood",
"counts",
"immunology",
"tropical",
"diseases",
"throat",
"treponematoses",
"bacterial",
"diseases",
"women's",
"health",
"signs",
"and",
"symptoms",
"sexually",
"transmitted",
"diseases",
"neglected",
"tropical",
"diseases",
"infectious",
"diseases",
"infectious",
"diseases",
"of",
"the",
"nervous",
"system",
"white",
"blood",
"cells",
"animal",
"cells",
"antenatal",
"care",
"neutropenia",
"diagnostic",
"medicine",
"blood",
"cell",
"biology",
"anatomy",
"meningitis",
"fevers",
"genitourinary",
"infections",
"physiology",
"neurology",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"neck",
"syphilis"
] |
2017
|
Neutropenia induced by high-dose intravenous benzylpenicillin in treating neurosyphilis: Does it really matter?
|
Rab proteins are small GTPases that act as essential regulators of vesicular trafficking . 44 subfamilies are known in humans , performing specific sets of functions at distinct subcellular localisations and tissues . Rab function is conserved even amongst distant orthologs . Hence , the annotation of Rabs yields functional predictions about the cell biology of trafficking . So far , annotating Rabs has been a laborious manual task not feasible for current and future genomic output of deep sequencing technologies . We developed , validated and benchmarked the Rabifier , an automated bioinformatic pipeline for the identification and classification of Rabs , which achieves up to 90% classification accuracy . We cataloged roughly 8 . 000 Rabs from 247 genomes covering the entire eukaryotic tree . The full Rab database and a web tool implementing the pipeline are publicly available at www . RabDB . org . For the first time , we describe and analyse the evolution of Rabs in a dataset covering the whole eukaryotic phylogeny . We found a highly dynamic family undergoing frequent taxon-specific expansions and losses . We dated the origin of human subfamilies using phylogenetic profiling , which enlarged the Rab repertoire of the Last Eukaryotic Common Ancestor with Rab14 , 32 and RabL4 . Furthermore , a detailed analysis of the Choanoflagellate Monosiga brevicollis Rab family pinpointed the changes that accompanied the emergence of Metazoan multicellularity , mainly an important expansion and specialisation of the secretory pathway . Lastly , we experimentally establish tissue specificity in expression of mouse Rabs and show that neo-functionalisation best explains the emergence of new human Rab subfamilies . With the Rabifier and RabDB , we provide tools that easily allows non-bioinformaticians to integrate thousands of Rabs in their analyses . RabDB is designed to enable the cell biology community to keep pace with the increasing number of fully-sequenced genomes and change the scale at which we perform comparative analysis in cell biology .
Intracellular compartmentalisation is found in all cellular lifeforms , yet eukaryotes have evolved extensive membranous compartments unique to this domain of life . Protein trafficking pathways accomplish the movement of cellular components like proteins and lipids between the cellular compartments . These essential pathways play house-keeping roles , such as transport of proteins destined for secretion to the plasma membrane via the secretory pathway , or recycling of membrane receptors via the endocytic pathway . In addition , they play a variety of specialised roles , such as bone resorption in osteoclasts , pigmentation in melanocytes and antigen presentation in immune cells . Malfunction of protein trafficking components leads to a large number of human diseases , ranging from hemorrhagic disorders and immunodeficiencies to mental retardation and blindness [1]–[4] , as well as cancer [5]–[9] . Furthermore , protein trafficking pathways are frequently exploited by human pathogens to gain entry and survive within host cells [10]–[13] . The endomembrane system accounts for a large fraction of the protein coding sequences in eukaryotic genomes [14] , and a plethora of data on molecules and interactions in different model organisms is available . However , it is unclear how these data map across organisms , and how general the mechanisms characterised in single species are . To answer these question we need to understand the evolution of the protein trafficking pathways and organelles . An evolutionary framework for protein trafficking is particularly important given the overwhelming accumulation of genomes , many from pathogenic organisms . Their comparative analysis can distinguish conserved from taxon-specific machineries , with clear practical applications . For example , conservation of genes led to the discovery of novel components and mechanisms in ciliogenesis [15] , whereas the presence of taxon-specific pathways allowed the identification of Fosmidomycin as a potential antimalarial drug [16] . Studying the evolution of protein trafficking is essential to understand the origins of eukaryotes . Comparative genomics and phylogenetics have established that the Last Eukaryotic Common Ancestor ( LECA ) already had a complex membrane trafficking system [17] including most types of extant molecular components [18] . These are believed to have expanded by duplication and specialisation giving rise to the full diversity of organelles and trafficking pathways observed today ( see [17] for a detailed description of this evolutionary scenario ) . Rabs are central regulators of protein trafficking . They are small GTPases that work as molecular switches to regulate vesicle budding , motility , tethering and fusion steps in vesicular transport [19] . Most recently the authors of [20] also linked Rabs to membrane fission . They recruit molecular motors to organelles and transport-vesicles , coordinate intracellular signalling with membrane trafficking , organise distinct sub-domains within membranous organelles and play a critical role in the definition of organelle identity ( recently reviewed in [21] ) . Rab subfamilies localise to distinct cellular locations , and regulate trafficking in a pathway- , organelle- and tissue-specific manner . This makes them ideal markers for the majority of trafficking-processes and compartments . Among trafficking-associated proteins , the Rab family expanded most in evolution [17] , [22] , suggesting that it provided the primary diversification element in the evolution of trafficking [22] . An important feature of the Rab family is that Rab orthologs tend to perform similar functions even in divergent taxa . For example , the mouse Rab1 has been shown to be able to functionally replace its ortholog YPT1 in yeast [23] . Hence assigning a Rab to a known and functionally described subfamily , e . g . Rab1 , is a strong functional prediction , i . e . functioning in the early secretory pathway in the case of Rab1 . Together with the ability to classify them into subfamilies based on sequence alone , this allows to establish the presence or loss of pathways and organelles solely based on the annotation of the Rab repertoire—a procedure we subsequently refer to as Rab profiling . Previously , we defined criteria to identify and classify Rab proteins [24] , which have been used as a basis for detailed manual analysis of the Rab families in a variety of organisms [25]–[33] . However , manual identification of Rab repertoires is tedious and time-consuming and not compatible with the deluge of fully sequenced eukaryotic genomes that new sequencing technologies are generating . We thus need to develop methods that enable the automated annotation of Rab proteins . Several characteristics of the Rab family make this a challenging bioinformatics problem . First , there is a strong non-specific signal from GTPase motifs spread throughout the protein sequence [34] , which makes it hard to distinguish Rabs from other small GTPases . Second , the Rab family is large due to extensive duplication in several branches of the eukaryotic tree ( e . g . [28] , [29] ) . Together with high sequence similarity amongst Rabs this causes difficulties to correctly classify Rabs into subfamilies and to further discern yet unseen subfamilies . Lastly , any automated scheme has to respect and perpetuate as much as possible the current naming conventions , despite any inconsistencies stemming from the decentralised nature of scientific discovery and the huge bias of existing annotations towards Ophistokonts . This requires a flexible , learning scheme both able to cope with the contingency of the field and to easily incorporate new naming consensuses . Here , we overcame these problems and developed an automated bioinformatic pipeline for the identification and classification of Rabs . We termed our pipeline the ‘Rabifier’ , which we describe , validate and benchmark . Using our tool , we cataloged nearly 8 . 000 Rabs from 247 genomes covering the major taxa of the eukaryotic tree , which we make available along with our pipeline at www . RabDB . org . Based on this comprehensive dataset of Rab proteins , we describe and analyse the evolution of Rabs . We found a highly dynamic family undergoing frequent taxon-specific expansions and losses . We extend the Rab repertoire previously reported to have been present in the LECA , identify the changes in the Rab family that accompanied the emergence of multicellularity and show that neo-functionalisation best explains the emergence of new human Rab subfamilies .
We implemented a bioinformatics pipeline to identify and classify Rab GTPases in any set of protein sequences independently of taxonomical information , which we term ‘Rabifier’ . The Rabifier proceeds in two major phases , which are schematised in Figure 1 . First , it decides whether a protein sequence belongs to the Rab family , i . e . that it is not a Ras , a Rho , etc . , and in the second phase it classifies the predicted Rab sequence into a Rab subfamily ( e . g . Rab1 ) . We describe the rationale for this procedure below—technical details are given in Materials and Methods and Text S1 . Phase 1 ( Figure 1A ) , which classifies protein sequences to the Rab family , proceeds in three stages . First , we check that the protein has a G-protein family domain . As the presence of such a domain can be decided with near certainty , this step drastically reduces the number of candidate Rabs while not excluding any real Rab . In order to do so , we align the sequence against a profile Hidden Markov Model ( HMMs ) [35] describing the known GTPase structures , as provided by the Superfamily database [36] . Secondly , we search for local sequence similarity by performing a BLASTp [37] query against an internal reference set of manually curated GTPases and discard the protein if it is most similar to a GTPase other than a Rab . At this stage of the workflow , the majority of non-Rab sequences has already been rejected ( see Figure 1C , where the number of sequences that transition between these phases is shown for M . brevicollis and for a database of 247 genomes described below ) . However , small GTPases are so similar to each other that a residual amount of false positives still remains undetected . We remove them in the third stage , where we scan the sequence for the presence of at least one of five characteristic RabF motifs defined in [24] . If no motif is found , it is concluded that the protein cannot be a Rab and rejected . Remaining sequences are all assigned to the Rab family at an individual confidence level computed for each Rab . The confidence score is derived from the combination of the individual statistics generated by the three stages according to a procedure described in Text S1 . The second phase ( Figure 1B ) proposes a classification into one of the Rab subfamilies present in our internal reference set , or suggests no similarity to any of those . It proceeds in two stages . First , we test whether the Rab respects a 40% identity cut-off to its BH that prevents assignment of too disparate sequences to any of the pre-defined subfamilies . If the cut-off is met , a classification is proposed , if not , the Rab is classified as belonging to the undetermined subfamily RabX . The use of a 40% threshold is supported in Figure S1 , and has previously been employed for example in [30] . The actual subfamily classification is based on the computation of a likelihood score for each of the subfamilies in our reference set . Intuitively , the protein is classified as belonging to the highest scoring subfamily , however , all scores are kept and thus provide an estimate of the relative uncertainty associated with each call . Like the Rab family score generated in the first phase of the Rabifier , the computation integrates output statistics from different tools , namely from local alignments via BLAST and from alignments using reverse Ψ-BLAST ( RPS-BLAST [38] ) . Similar to HMMs , RPS-BLAST compares a sequence against a summary of a set of sequences , in our case summaries of all sequences in our reference set belonging to a single Rab subfamily , and measures how likely the input belongs to any the subfamilies . This way we take information from all sequences in the internal reference set into account . For details on the procedure check Materials and Methods and Supplementary Methods Text S1 . Any new methodology has to be validated . Ideally this is based on a test data set fulfilling three requirements: the test data is correctly and comprehensively annotated with those features the tool automatically detects , it is large enough to provide robust statistics , and it covers the entire range of possible inputs the tool might encounter in its real-world application , at best even respecting the expected proportions of worst- to best-case inputs . In our case , no dataset is available which fulfils the three requirements simultaneously: Rab repertoires are only available for a limited number of organisms which are not evenly distributed across eukaryotic phylogeny , and whose annotation was manually performed by different groups , hence may be inconsistent or even incorrect ( in some cases a ‘correct’ , i . e . consensual , classification might not even exist ) . In the absence of a suitable validation dataset , we opted to validate the Rabifier against the manually curated Rab families of three organisms representing distinct worst case scenarios for the Rabifier ( Figure 2A–C , see Table S1 for a list of all sequences used ) . This ensures that the validation is meaningful , as it provides a strict lower bound on the expected performance in everyday use . First , we chose the Excavate Trypanosoma brucei [32] , which is one of the most distantly related organism to our reference sequences , which are dominated by Ophistokonts ( an unranked scientific classification sometimes also called ‘Fungi/Metazoa group’ ) . The second is Entamoeba histolytica [30] , a Unikont from the phylum of Amoebozoa that is thus marginally closer to the sequences that dominate our reference database , but has a heavily expanded and diverse Rab repertoire which makes it challenging to assign Rab subfamilies . The third organism , Monosiga brevicollis from the class of Choanoflagellates , was chosen as a representative of a phylum ( Choanozoa ) for which no information on the Rab family is available yet . In this third case , we compare the automated predictions against a manual analysis we performed in this study ( Figure 2E ) , and which we will discuss below . The first aspect we assessed is the ability of the Rabifier to distinguish Rabs from other GTPases ( summarised in Figure 2A ) . We present the Rabifier with the set of GTPases from the above organisms and count how often we miss a Rab ( false negative—FN ) , and how often we incorrectly classify a non-Rab as a Rab ( false positive—FP ) . For T . brucei , we correctly classified 101 out of 102 GTPases as being a Rab or not , 292 out of 295 in E . histolytica and finally all 125 GTPases in M . brevicollis . Altogether , we have no FP and 4 FN , which means that for this particular set of genomes we make correct decisions about whether a protein is a Rab in 99 . 2% of the cases with no differences amongst the organisms . In order to understand the sources of the misannotations at family level , we inspected the false negatives individually . The Rabifier disagrees with the manual curation of [32] in T . brucei for TbRabX3 , a RabL2-like protein , that is counted as a false negative . We explicitly added RabL2 sequences to our negative data set as we do not consider these proteins as members of the Rab family ( see Materials and Methods ) . The remaining disagreements between the Rabifier and the manual annotations are three false negative proteins in E . histolytica in which we cannot find any detectable RabF motif , and one protein which has no similarity to any member of our reference dataset of small GTPases . We conclude that these proteins are likely misclassified in [30] , and hence that the above failures of the Rabifier to identify Rabs are artificially introduced by our validation procedure . Secondly , we established the accuracy by which a given Rab sequence is assigned to the right subfamily ( summarised in Figure 2A ) . Concretely , for those sequences which were correctly identified as Rabs , we checked whether the proposed subfamily agreed either with the public annotation or our own one for M . brevicollis . We distinguished between two operating modes of the Rabifier: a normal one which does not consider the confidence levels the Rabifier attributes to its classifications , and a high-confidence mode which accepts only the high-confidence annotations above a certain confidence threshold , whereas those below are classified as belonging to the undetermined subfamily RabX . Ignoring the information provided by the classification confidence , we correctly called 16 out of 17 Rabs for T . brucei , 59 out of 91 in E . histolytica and 20 out of 25 for M . brevicollis , leading to an overall fraction of 71 . 4% correct decisions ( 79 . 7% on average per organism ) . However , if one defines a threshold below which a classification is systematically considered as belonging to the undefined subfamily RabX , the accuracy can be substantially improved . To illustrate this , Figure 2B displays the distribution of scores associated to correct and wrong calls , which shows that wrong calls clearly have lower confidence scores on average . In order to test for all possible thresholds exploiting this difference , we performed a ROC curve analysis presented in Figure 2C . This machine learning technique allows to summarise and quantify the classification performance for all thresholds ( Area Under the Curve ( AUC ) [39] , here 0 . 94 ) , and enables to objectively choose a threshold providing an optimal TP/FP-tradeoff . Here , we opted for 0 . 4 , which we propose as a default choice for the interpretation of the Rabifier's results . Yet , the use of this threshold is not fixed as it may vary depending on the dataset , and can be freely modified by users of the Rabifier . The consequences of applying a cutoff on the classification accuracy are quantified by the inlay in Figure 2C: only trusting calls with confidence higher or equal to 0 . 4 greatly reduces the amount of misclassified Rabs from non-human subfamilies and improves the overall accuracy to 90% ( 92 . 01% on average per organism ) . In summary , we conclude that our workflow is able to correctly discern Rabs from other GTPases . Furthermore , calls both at family and subfamily level have an associated confidence score which correctly captures uncertainty in the decision . Relying on the information provided by the confidence level , the Rabifier suggests correct subfamilies around 90% of the time even in difficult and phylogenetically isolated cases . After having established the correctness of our procedure , we wished to assess the improvement it represents over possible alternative large-scale approaches in an objective manner . This excludes benchmarking against methods for example based on phylogenetic trees , as reasoning over them is difficult to automate and not feasible for thousands of sequences . We chose to compare the Rabifier to the Conserved Domain Database at the NCBI [40] , the only resource we are aware of that specifically scores for RabF motifs . To this end , we implemented an alternative decision scheme which given a protein retrieves the protein name and CDD domain annotation of its BH in the NCBI protein database . Note that if the protein is in the NCBI database , the BH retrieves the protein itself . As for the choice of genome , the Rabifier has to be benchmarked against an organism whose Rab family has not been manually curated , as our alternative procedure would simply retrieve that annotation . Moreover , an organism from a taxon which is both close to Metazoa and for which no information on the Rab family exists best ensures an unbiased measurement . These requirements are met by the Choanoflagellate M . brevicollis , which we analysed ourselves and is thus an ideal candidate for a direct comparison . The results of this experiment are detailed in Figure 2D ( see also Table S1 ) . As above , we distinguished between the ability to discern Rabs from other GTPases and to actually propose the correct subfamily for a given Rab . First , while the Rabifier achieved 100% accuracy in separating Rabs from other GTPases in M . brevicollis , the alternative strategy—although not introducing false positives—misses 8 of 25 Rabs leading to an overall drop in sensitivity . On top of these 8 sequences , the Rabifier correctly suggests subfamilies for 4 further proteins wrongly classified by the alternative strategy , leading to an overall difference of 12 sequences correctly classified only by the Rabifier . Thus , our annotation pipeline represents a significant improvement over currently available large scale approaches , both in terms of sensitive identification of Rabs and especially with regards to the difficult automatic classification of Rabs into subfamilies . In order to make our pipeline useful to the cell biology community interested in Rabs , we provide access to the Rabifier in form of a web tool ( Figure 3A ) . Via the graphical interface users can submit up to five protein sequences at a time , and the classifications generated by our workflow are returned together with their associated degree of confidence . We envisage users who want to quickly generate hypotheses about one or a few candidate proteins . Users wishing to classify more sequences are encouraged to contact us . We emphasise that the Rabifier works without need for phylogenetic information about the input , hence any set of protein sequences can be submitted . In addition , we generated a database of nearly 8 , 000 classified Rab sequences in 247 eukaryotic genomes , which we make publicly available at www . RabDB . org ( Figure 3A ) together with basic browsing and visualisation tools . Our database is built on top of the Superfamily database [41] ( September 2009 release ) , which allows us to follow its release cycle and include predictions for all newly sequenced genomes contained therein . Figure 3B details the phylogenetic distribution of genomes in RabDB and the number of Rabs we predict in each of those eukaryotic branches . The correctness of the content in www . RabDB . org is not manually confirmed systematically . However , we constantly inspect and manually curate the generated predictions and update our internal reference database accordingly . Furthermore , we provide users the possibility to notify us of a potential mis-annotation found in the database such that we can correct the classification of the Rab in question . These measures further enhance the expected quality of future releases of www . RabDB . org . As can be noticed from Figure 3B , the Rabifier detected a large number of Rabs not belonging to any subfamily represented in our reference set , i . e . most subfamilies which have been described before . By definition these sequences show no similarity to any functionally characterised Rab , hence a bioinformatic annotation is not possible . However , in order to structure the space of new sequences and provide a starting point to study this yet unexplored diversity , we clustered these Rabs with respect to their sequence identity and propose several hypothetical Rab subfamilies ( see Material and Methods for details ) . The result of this procedure is shown in Figure 4 , which details the amount of hypothetical subfamilies according to the breadth of their occurrence ( see Figure S7 for an overview of the amount of Rabs falling into each of these classes ) . We integrated these new subfamilies both in our database , where they can be browsed with help of the visualisation tools we provide , and in the online version of the Rabifier . Note that in addition to these new hypothetical subfamilies we still find hundreds of Rabs that we cannot group with others . Those may result from erroneous gene models in less well curated genomes , represent cases where our simple clustering procedure failed , or indeed be bona fide singletons . A detailed phylogenetic analysis may be required to resolve these cases which is out of the scope of this study . A dataset of 8 , 000 Rabs allows us to take a global view of the Rab sequence space , and to address previously inaccessible questions . Here , we investigate the patterns of Rab repertoire expansion in the eukaryotic tree ( Figure 5 ) . Expansion of certain protein families has been found to correlate with organismal complexity [42] . The anecdotal evidence of Rab profiles in different organisms suggests at least three possible scenarios: a conserved core of Rabs present in all organisms; tinkering with a core of subfamilies by taxon- or species-specific expansions of existing subfamilies; a major variation of the Rab machinery with taxon- or species-specific Rab repertoires . We asked whether any such scenario is apparent for the Rab family across the eukaryotic tree , or if different ones predominate in different branches . We observe a tremendous heterogeneity in the sizes of Rab repertoires , ranging from 5 to several hundreds of Rabs in Encephalitozoon cuniculi and Trichomonas vaginalis respectively . Genomic analyses have shown a general trend for more and larger families in bigger genomes [43] , [44] . In the case of Rabs , linear regression over all taxa reveals that genome size explains roughly 60% of the observed variance in numbers of Rabs in an organism ( Figure S2 ) . However , due to the current bias in fully sequenced genomes towards Ophistokonts ( compare Figure 3B ) , it is unclear whether these numbers will remain as such in the future . We find that closely related organisms tend to have similar Rab repertoires in size , but at the level of phyla we encounter marked differences indicating taxon-specific adaptations . For example , although Ciliophora and Apicomplexa belong to the same superphylum ( Alveolata ) , these sister phyla show very different repertoires , highly expanded in the first case , and streamlined in the second . The smaller Rab repertoires in Apicomplexan genomes , mostly dominated by intracellular parasites , may be due to secondary gene loss , similar to that reported in bacterial intracellular parasites and endosymbionts [45] and in the obligate intracellular parasitic Microsporidia [45] . Another example of reduction of Rab repertoires is observed in the fungal branch , as we reported previously [26] and now confirm based on an extended set of 103 genomes . It is noteworthy that Fungi are Unikonts , a taxon which comprises Metazoa and Amoebozoa , i . e . branches that appeared to have suffered independent expansions of their Rab repertoires [24] , [30] . We observe large expansions in Diplomonadida/Trichomonadida , Ciliophora and Amoebozoa . Much of these expansions are accounted for by species-specific subfamilies ( see Figure 4 ) . This demonstrates that there is frequent invention of new Rabs , perhaps in a taxon-specific manner—a hypothesis that will have to await broader sampling of the genomes space to be tested in most taxa . On the other hand , inspection of Figure 5 reveals that for those Rabs that can be classified , different subfamilies expanded in each branch of the tree . For example , Rab7 forms the largest subfamily in Diplomonadida/Trichomonadida and Amoebozoa , whereas Ciliophora's most expanded subfamily is Rab2 . This suggests that these are independent expansions , which has already been observed for example within the Rab5 subfamily [26] , [46] . Note that we repeated these analyses for different confidence cutoffs and observed no significant consequences on the broad picture . In summary , the global evolution of Rab repertoires is highly dynamic with frequent taxon-specific subfamily expansions , gain of new Rabs and losses . Hence , we observe a scenario where a core set of Rabs tends to be universally conserved , and can coexist in different taxa with subfamily expansions and/or taxon- or species-specific Rabs . It is clear that no unique path to cellular complexity and specialisation exists , implying that any conclusion about the evolution of Rabs in a given taxon is not necessarily true for other eukaryotic taxa . The systematic identification and classification of Rab repertoires in multiple branches of the eukaryotic tree of life allows the establishment of a phylogenetic profile for each Rab subfamily . As Metazoa and Fungi are the most extensively sampled and best annotated groups , we profiled human subfamilies ( Figure 6 ) and determined their likely time of origin ( Figure 7 ) . For a detailed analysis of fungal Rabs see [26] . We further established the direction of duplication , i . e . from which Rab subfamily another emerged by duplication and subsequent divergence , by crossing their likely time of origin with a phylogenetic tree of the human Rab family . We reasoned that for two closely related Rabs , the one that is present in more taxa is likely the ancestral one . Since all Rabs are by definition paralogs and especially the deeper evolutionary relationships are unclear , we restricted the inference of direction of duplication to well supported branches . Here , we define well supported branches as those with bootstrap support higher than 58% in a tree of human Rabs , which is chosen to include the branch between Rab5 and Rab22 as their association is commonly accepted [47]–[51] . As further support , we note that all branches selected according to this criterion are also present in the tree of mouse Rabs we present below , however , in general 58% is not a strong branch support and should not be used indiscriminately on trees of other Rabs . Based on a 58% cutoff , one obtains directed duplication scenarios for a number of subfamilies as summarised in Figure 7 . We term subfamilies with a clear origin as ‘derived’ . This analysis suggests new candidates for ancestral Rabs . Previously Rab1 , 2 , 4 , 5 , 6 , 7 , 8 and Rab11 [17] , Rab18 [31] , [52] , Rab21 [30] , [53] as well as Rab23 and 28 [32] could be mapped to more than one major branch of the eukaryotic tree , making them likely candidates to be present in the LECA . Our results support these assignments and reveal a new set of proteins that can be found in two or more basal eukaryotic taxa , namely Rab14 , 32 and RabL4 . Applying the same parsimony argument as previous studies suggests that these Rabs were part of the ancestral set of Rab in the LECA . Are these putative ancestral Rabs an artefact due to incorrect assignments or convergent evolution ? We validated the automated subfamily classification by phylogenetic trees , and could not disprove their annotation ( Figures S4 A–C ) . The possibility of convergent evolution is however harder to rule out . Regardless , an organism with 15 Rabs is not surprising and comparable with some unicellular eukaryotes [32] , [33] , and free living fungi frequently have less [26] . It is remarkable that with every new analysis the LECA appears to become increasingly more complex [54] . On functional grounds , mapping these Rabs to the LECA is plausible . RabL4 , also known as IFT27 , plays a role in ciliogenesis as part of the Intra Flagella Transport ( IFT ) machinery [55] . Flagella are believed to be ancestral characters , present in the LECA [56] , [57] . Rab32 regulates transport to the pigmented/secretory granules [58] , an animal-specific function , but it has also been claimed to have a mitochondria-related function [59] , [60] . The known function of Rab14 in phagosome maturation and a recycling step at the TGN [61] , [62] is less clearly ancestral , but it may lend support for a phagotrophic LECA as previously proposed [63] . In summary , our results support the claim that the LECA had a highly complex endomembrane system , and that secondary Rab losses have been dominant in the evolution of the major eukaryotic taxa [17] . The emergence of multicellularity is one of the major transitions in evolution [64] , which happened independently multiple times ( see [65] for a recent review ) . There are several critical features necessary for the evolution of multicellular organisms , for example mechanisms for cell adhesion , cell polarity and inter-cellular communication . Little is known about how protein trafficking has evolved during this transition . We take advantage of our extensive annotation of the Rab family to derive the Rab complement prior to and after the emergence of multicellularity in Metazoa . Monosiga brevicollis belongs to the Choanozoa , the closest unicellular relatives of Metazoa . The genome of this organism was only recently sequenced [66] , and in the context of the validation of the Rabifier we conducted a detailed analysis of its Rab family . The phylogenetic tree in Figure 2E reveals a relatively large Rab family with nearly no subfamily expansions ( see also Figure 5 ) , i . e . mostly with a single member per subfamily ( only Rab32 has two members ) . This is also observed in simpler animals like D . melanogaster and C . elegans [52] , suggesting that larger subfamilies observed in mammals represent taxon-specific duplications . Secondly , we observe several organism-specific Rabs , which we labeled MbRabX . Consistent with results from the last section , the “invention” of new Rabs is a recurrent feature in multiple branches of the tree of life ( e . g . [28] , [30] , [32] , [52] ) . We observed the emergence of three novel sub-families , Rab9 , 22 , 29 , none playing ‘animal-specific’ roles . The function of Rab29 is unknown , but Rab9 and Rab22 both appear to be involved in late endocytic traffic [49] , [50] , [67] , [68] . Surprisingly , the genome of M . brevicollis codes for proteins previously believed to be specific to multicellular organisms , for example Cadherins [66] , [69] . In animals , trafficking of the cell adhesion molecules Integrins and Cadherins is regulated by Rab4 , 5 , 11 , 21 and 25 [70]–[73] , and Rab5 and 7 [74] , [75] , respectively . Interestingly , these Rabs are also found in M . brevicollis , and—with the exception of Rab25—are all likely ancestral proteins . That highlights that complex new functions , as are for example the regulation of Cadherin and Integrin and ultimately cell adhesion , can be gained without inventing new subfamilies . Our analysis revealed 14 Rab subfamilies that emerged at the base of Metazoa ( Figure 7 ) . Surveying the currently known functions of these animal-specific subfamilies suggests roles mainly in regulated secretion ( Rab3 [76]–[79] , Rab26 [80] , Rab27 [79] , [81]–[83] , Rab33 [79] , Rab37 [79] , [84] , Rab39 [85] ) , trafficking from ( Rab10 [86] ) and to the Golgi ( Rab43 [87] ) and more generally localisation at the Golgi ( Rab30 [88]–[90] , Rab33 [91] , Rab34 [92] , Rab43 [93] ) . Hence , our analysis suggests that the appearance of animals cooccurred with an important expansion and specialisation of the secretory pathway . Gene duplication is a frequent mode of gene gain in eukaryotes . This is well illustrated by the expansion of the Rab family in emergence and evolution of Metazoa . Following gene duplication , the most common fate for one of the duplicates is accumulation of mutations up to the point of pseudogenisation . In the alternative case , the retention of both duplicates has been explained by different theoretical scenarios , recently surveyed in [94] . Most prominently , either divergence results in gain of a beneficial new function ( neo-functionalisation ) by one of the duplicates , or disruption of complementary parts of the function in each of the genes leaves both paralogs indispensable to perform the original function ( sub-functionalisation ) . As discussed in [94] , those models predict distinct types and strengths of selective forces acting on the two duplicates allowing to test and distinguish amongst putative scenarios . Namely , while in both neo- and subfunctionalisation the new copy indistinguishably evolves neutrally , detecting purifying selection acting on the original copy is an indication of neo-functionalisation , whereas relaxed purifying or neutral selection is suggestive for sub-functionalisation . In the case of Rabs , Figure 6 shows that the original copy is conserved and keeps its identity as the original subfamily , whereas the new copy initiates a distinct subfamily defined by a discernible level of sequence divergence . We interpret this pattern as evidence that the mode by which the Metazoan Rab family expands is most probably neo-functionalisation rather than subfunctionalisation . To gain further insights into the nature of the gain of function , we asked whether the derived Rab subfamilies show differences in tissue-specificity that could hint at the type of newly evolved functions . To this end , we investigated tissue-specificity in expression of Rabs in mouse tissues and cell lines ( Figure 8 ) by means of PCR ( see Materials and Methods ) . We also analysed publicly available microarrays ( Figures S5 , S6 ) which overall corroborate the trends described in the following . First , we observed that all ancestral Rabs are widely expressed ( i . e . in all tested tissues ) , most probably performing general functions required in all tissues . Similarly , Rabs that predate the advent of multicellularity are also broadly expressed , a general phenomenon that has been described for genes which emerged prior to multicellularity [95] . Second , for the derived subfamilies in which a clear directionality of duplication could be established ( see Figure 7 ) , we detected a trend for an increase in tissue specificity , i . e . a reduction in number of tissues in which the Rab is expressed relative to its progenitor subfamily . For example , Rab34 is expressed in all tissues investigated but the liver , whereas the derived Rab36 is only expressed in lung and brain . Thirdly , at no time we observe complementary expression , i . e . a pair of subfamilies which have opposite tissue specificities . Overall , these observations are strong indications that derived subfamilies are retained for a new tissue-specific functions , different from or at least complementing the progenitor ones . Thus , our results support a neo-functionalisation model explaining the retention of novel Rab sub-families in Metazoa . This model makes several predictions about expression patterns of Metazoan Rabs for which we could not derive expression data . Concretely , Rab41 which we only find in primates and dolphin is expected to show a restricted tissue expression , as its origin from Rab6 is statistically well supported . Rab29 is expected to be ubiquitously expressed despite its clear origin from Rab32 as it predates the evolution of multicellularity , a prediction at least supported by our microarray-based analysis ( Figure S5 ) . One notable observation is that the tested mouse tissues express an unexpectedly high number of distinct Rabs . This is also observed in individual cell lines , which indicates that it is not an artefact from multiple cell types mixed in the tissue . While it is clear that Rabs are expressed at different levels [96] ( see also Figure S6 ) , our results from a more sensitive method than microarrays reveal that the tissue-specific Rabs may be more widely expressed than previously anticipated . It remains to be investigated whether the low levels of expression we can detect by PCR are functionally significant . We developed the ‘Rabifier’ , a bioinformatics tool to identify and classify Rabs from any set of protein sequences with no need for additional phylogenetic information , which we make available as a web tool for the community . We deployed the Rabifier on 247 proteomes predicted from complete genome sequences , generating the first comprehensive view of the Rab sequence space , which we also make available in form of a browsable database of Rab proteins . We envisage that cell biologists interested in specific organisms may use RabDB and the Rabifier as a first description of the family , at accuracy levels we showed to be very high . In fact , our predictions are well suited to be the first step towards high quality manual annotations . Furthermore , we introduced unified and objective criteria for the annotation of Rabs which is especially important for large-scale comparative studies , which can now be grounded on a coherent body of data . The classification of Rab repertoires in hundreds of genomes gives us the first global view of the Rab family in evolution , revealing that this family followed different routes in each branch of the tree . Massive expansions co-exist with extensive losses . These expansions can vary from taxon to taxon , suggesting that care must be taken when transferring information amongst different branches of the tree of life . In this respect , future work may focus on understanding the detailed evolutionary patterns in eukaryotic taxa other than Metazoa , which we analysed here . It appears that plants are ideal candidates for such a study as multiple genomes have been sequenced covering both unicellular and multicellular organisms . One of the perhaps most surprising observations we made was the extension of RabX's , i . e . Rabs that cannot be assigned to any previously characterised subfamily . Hence , a major bioinformatic and cell biological challenge now is to identify how many Rab subfamilies exist overall , and to establish their conservation or taxon-specificity . Here , we started this classification by proposing new Rab subfamilies derived from clustering of RabX's with respect to their sequence similarity . We hope to stimulate further research which may allow the refinement of our criteria and ultimately the definition of a Rab subfamily . The notion of Rab subfamily is supposed to reflect both evolutionary history and functional information , but has historically been mixed with less clear criteria . In the absence of functional information for all Rabs , phylogenetic analysis becomes particularly important , especially for functional prediction . In this context , it is all the more serious that we found a notorious frailty of Rab trees . Factors such as choice of sequences , outgroups , alignment program , probabilistic model and program implementing it contribute to very different trees ( compare for example [52] , [97] , [98] and Figures S4 A–C ) . We thus need to derive objective criteria that define a Rab subfamily which go beyond the clearly outdated yet still useful sequence identity cutoff [24] . Possibilities are for example to introduce soft thresholds depending on background divergence levels within a given taxon , or to restrain the area considered to measure sequence divergence to the functionally relevant regions . We focused on the evolutionary path from the LECA to mammals in order to gain insight into the mechanism of functional innovation within the Rab family . Based on objective and re-usable criteria we were able to map directionality to duplications clarifying the origin of some human subfamilies . Crossing these relations with data on tissue-expression patterns of Rab genes , we proposed that neo-functionalisation best explains the emergence of new subfamilies . More recent subfamilies are most likely retained for newly evolved tissue-specific functions and coexist with older ones in a subset of tissues . It remains to be determined whether the same happens within a subfamily , i . e . whether a RabXa and a RabXb represent cases of neo- or sub-functionalisation [99] . This is particularly relevant to conceptually tell apart isoforms and distinct subfamilies . As we restricted our analysis to subfamilies present in humans , it is important now to test whether the same neo-functionalisation scenario is observed in other branches of the tree of life . As mentioned before , plants appear to be ideal candidates to extend this analysis . Finally , while we studied the fate of new subfamilies in the context of tissue-specific expression , it will be important to understand the contribution of subcellular re-localisation to neo-functionalisation [100] , [101] . New generations of sequencing methods promise to change that scale at which we perform comparative analysis in cell biology . But for this change to reach the cell biology community , we need the appropriate tools that allow the non-bioinformatician to take advantage of all the emerging data . The Rabifier is one such tool , tailored to enable the cell biologist to analyse protein repertoires in hundreds of genomes .
C57BL/6 mice were bred and housed in the pathogen-free facilities of the Instituto de Gulbenkian de Ciência ( IGC ) . Mouse experimental protocols were approved by the Institutional Ethical Committee and the Portuguese Veterinary General Division . Before we devised a workflow able to identify and classify Rabs , we decided which protein subfamilies we considered being human Rab subfamilies . Since the early genomic analyses of the human Rab repertoire reporting subfamilies 1 to 40 ( with exception of 16 ) [24] , five subfamilies have been newly discovered ( 41 to 45/Rasef ) [102] . Besides those clear cases , the distinction remained less obvious for those which are termed ‘Ran’ and ‘Rab-like’ , each of which we briefly discuss in the following . Rans control nucleocytoplasmic shuttling [103] , and are frequently considered to be members of the Rab family [97] , [102] . This view is supported by our own phylogenetic analysis ( see tree in Figure S3 ) , although without strong bootstrap support . Due to the distinct function and localisation [103] partly within the nucleus we do not further consider Rans in our dataset . However , Rans have recently been linked to ciliary entry of certain kinesins [104] , and they may be included in the future . RabL2 proteins were already mentioned in [24] where it is concluded that they are not Rabs , amongst others due to non-conforming RabF motifs . In [97] , RabL2's are said to cluster together with Rans , which we do not include in our analysis . The tree of human GTPases shown in [98] suggests that RabL2 proteins branch of Rhos at an early stage . Finally , our own tree of human GTPases ( Figure S3 ) positions RabL2s at the periphery of the Rab branch , yet with little bootstrap support . Altogether , we do not see enough evidence for RabL2 proteins to be considered Rabs . The situation is similar for RabL3 and RabL5 . Colicelli clusters them together with Rans [97] , whereas in [98] both reside on a branch with Arfs though classified as belonging to none of the classes Rab , Ras , Arf , Rho or Ran . Our tree of human GTPases suggests that RabL5 and Arfs have a common ancestor , equally so RabL3 and RabL2 , hence we ignored both in our further analysis . Rab7L1 is nearly identical to Rab29 and represents a simple case of naming ambiguity , as has already been pointed out in [24] . The last case is RabL4 , which all [97] , [98] , [102] consider being a Rab . We confirmed that interpretation by detecting and validating four RabF motifs , as well as by our phylogenetic tree , which places RabL4 within Rabs . However , we only group RabL4 together with Rab28 as suggested in [97] , [102] when no GTPase other than the human Rab subfamilies 1 to 45 are included ( see trees in Figure S3 and Figures S4 A–B ) . In mouse , RabL4 is not classified as being monophyletic with Rab28 ( see Figure S4 C ) . We give some technical details about the implementation of the Rabifier which for the sake of brevity have been omitted above . For information on the computation of the confidence scores see Text S1 . In the first phase ( Figure 1A ) , the profile HMM's representing the G-protein family domain are either run manually using Perl scripts ( as of June 2010 ) provided by Superfamily [36] and HMMER 2 . 3 . 2 [35] , or in the case the sequences have been retrieved from the Superfamily database [41] the domain structure is taken directly from Superfamily . Note that Superfamily is a pure protein resource that contains proteomes predicted from genome sequences . It does not provide information about the underlying genes systematically , hence counts of how many Rab genes are present in a specific genome can generally not be derived from Superfamily . BLASTp [37] queries are performed with soft masking ( parameters -F m S ) and considered up to an e-value threshold of 10−10 . Our reference set of sequences not being Rabs is provided as Dataset S1 , whereas the reference database of Rabs are the sequences accessible at www . RabDB . org with redundancy removed using CDHit ( at a 90% sequence identity threshold ) [105] . Our reference data set of Rabs covers more than just the human subfamilies , namely previously published and functionally described subfamilies from Arabidopsis thaliana ( AtRabA1 , AtRabA3–AtRabA6 , AtRabC2 , AtRabD1 , AtRabF1 , AtRabG1 ) [31] , yeast ( yptA , ypt10 , ypt11 ) , Drosophila melanogaster ( DmRabX1–DmRabX6 , DmRab9D , DmRab9F ) and C . elegans ( CeRabY6 ) [52] . Furthermore , as detailed in the main text we proposed a set of hypothetical subfamilies which we integrated into our reference set . The members and phylogenetic distribution of these hypothetical subfamilies can browsed directly on our web site www . RabDB . org . The last stage of the first phase is performed using the Motif Alignment & Search Tool ( MAST ) ( motif finding threshold 0 . 0005 ) [106] from the MEME-suite [107] , with probabilistic representations of the motifs ‘igvdf’ , ‘klqiw’ , ‘rfxxxt’ , ‘yyrga’ , ‘lvydit’ [24] as input generated on our reference database of Rabs beforehand using MEME . In the second phase ( Figure 1B ) , RPS-BLAST queries [38] are performed with standard parameters and an e-value threshold of 10−5 , with position-specific scoring matrices ( PSSM ) previously generated by Ψ-BLAST on all members of each of the Rab subfamilies present in our reference database . The hypothetical subfamilies result from two distinct clustering steps . First , we clustered sequences classified as RabX by the Rabifier and belonging to the same genome at a sequence identity threshold of 70% [24] . In order to resolve the potential conflicts caused by sequences that belong to several clusters at the same time , we applied MCL [108] ( inflation parameter 2 . 0 ) , which resulted in a clean partition , i . e . non-overlapping clustering , of the sequences . In a second step , we merged the resulting clusters across genomes if at least one pair of sequences across clusters shared a sequence identity over 70% . We chose this threshold as it is the lowest which ensures meaningful clusters , that is clusters which in their majority respect taxa boundaries . All phylogenetic trees of Rabs and GTPases presented in this article have been generated with PhyML [109] , which implements a Maximum Likelihood probabilistic model , using standard parameters and 100 bootstraps . Alignments were performed with MAFFT [110] , and manually edited to remove sites with deletions using Jalview [111] . The human trees have been generated using human kRas as an outgroup , the mouse trees using mouse kRas as outgroup , and the mixed tree of human and Monosiga brevicollis Rabs uses both human and M . brevicollis kRas as outgroups . Sequence accessions of all sequences can be taken from Table S2 . Tree visualisations have been generated with Figtree ( http://tree . bio . ed . ac . uk/software/figtree/ ) . The tree of human Rabs not displaying isoforms ( see Figure 5 , Figure 6 ) has been generated by removing isoforms and keeping the longest branch as representative of the corresponding subfamily .
|
Intracellular compartmentalisation via membrane-delimited organelles is a fundamental feature of the eukaryotic cell . Understanding its origins and specialisation into functionally distinct compartments is a major challenge in evolutionary cell biology . We focus on the Rab enzymes , critical organisers of the trafficking pathways that link the endomembrane system . Rabs form a large family of evolutionarily related proteins , regulating distinct steps in vesicle transport . They mark pathways and organelles due to their specific subcellular and tissue localisation . We propose a solution to the problem of identifying and annotating Rabs in hundreds of sequenced genomes . We developed an accurate bioinformatics pipeline that is able to take into account pre-existing and often inconsistent , manual annotations . We made it available to the community in form of a web tool , as well as a database containing thousands of Rabs assigned to sub-families , which yields clear functional predictions . Thousands of Rabs allow for a new level of analysis . We illustrate this by characterising for the first time the global evolutionary dynamics of the Rab family . We dated the emergence of subfamilies and suggest that the Rab family expands by duplicates acquiring new functions .
|
[
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] |
[
"computer",
"applications",
"genome",
"expression",
"analysis",
"cellular",
"structures",
"subcellular",
"organelles",
"genome",
"evolution",
"evolutionary",
"biology",
"molecular",
"genetics",
"information",
"technology",
"sequence",
"analysis",
"comparative",
"genomics",
"biology",
"databases",
"computer",
"science",
"web-based",
"applications",
"cell",
"biology",
"gene",
"identification",
"and",
"analysis",
"genetics",
"gene",
"duplication",
"genomics",
"molecular",
"cell",
"biology",
"genomic",
"evolution",
"computational",
"biology",
"genetics",
"and",
"genomics"
] |
2011
|
Thousands of Rab GTPases for the Cell Biologist
|
High-throughput drug screening has facilitated the discovery of drug combinations in cancer . Many existing studies adopted a full matrix design , aiming for the characterization of drug pair effects for cancer cells . However , the full matrix design may be suboptimal as it requires a drug pair to be combined at multiple concentrations in a full factorial manner . Furthermore , many of the computational tools assess only the synergy but not the sensitivity of drug combinations , which might lead to false positive discoveries . We proposed a novel cross design to enable a more cost-effective and simultaneous testing of drug combination sensitivity and synergy . We developed a drug combination sensitivity score ( CSS ) to determine the sensitivity of a drug pair , and showed that the CSS is highly reproducible between the replicates and thus supported its usage as a robust metric . We further showed that CSS can be predicted using machine learning approaches which determined the top pharmaco-features to cluster cancer cell lines based on their drug combination sensitivity profiles . To assess the degree of drug interactions using the cross design , we developed an S synergy score based on the difference between the drug combination and the single drug dose-response curves . We showed that the S score is able to detect true synergistic and antagonistic drug combinations at an accuracy level comparable to that using the full matrix design . Taken together , we showed that the cross design coupled with the CSS sensitivity and S synergy scoring methods may provide a robust and accurate characterization of both drug combination sensitivity and synergy levels , with minimal experimental materials required . Our experimental-computational approach could be utilized as an efficient pipeline for improving the discovery rate in high-throughput drug combination screening , particularly for primary patient samples which are difficult to obtain .
Despite great advances in the understanding of cancer , there remains a major challenge to develop more effective anti-cancer treatments . Next generation sequencing has revealed the intrinsic heterogeneity in cancer genomes , which partly explains why patients respond differently to the same therapy [1] . To reach durable clinical responses , cancer patients who relapse and become refractory to standard chemotherapy need novel multi-targeted drug combinations which can effectively overcome the emergence of drug resistance [2–4] . Ideally , a potential drug combination should achieve therapeutic efficacy at reduced dosages , and therefore minimize the toxicity and other side effects associated with high doses of single drugs [5–6] . Therefore , two important properties for a drug combination must be evaluated: sensitivity and synergy . Sensitivity of a drug combination is defined as the level of treatment response , usually measured in the unit of percentage inhibition of cell viability or growth . In contrast , synergy of a drug combination is referred to the degree of drug interactions that contributes to the drug combination sensitivity independent of the single drug effects [7] . In order to identify sensitive and synergistic drug combinations , high-throughput drug screening has been applied on a large variety of cancer cell lines and more recently on patient-derived cancer samples [8–9] . Many high-throughput drug combination screens test pairs of drugs at a dose matrix , for which the cell viability or growth inhibition effects are measured [10–11] . The dose-response matrix results from a full factorial design that involves multiple dose combinations of a drug pair , and thus demands a relatively large amount of cancer cells . For patient-derived cancer samples , which are challenging to obtain and restricted in volume , the full matrix design may be infeasible to test even with a minimal number of drug combinations . Furthermore , cancer samples of different genetic profiles are known to respond differently to the same treatment [12] . With the limited amount of drug combination data points , it becomes a daunting task for any machine learning approach to navigate the combinatorial space to pinpoint the most promising drug combinations that are selectively effective for individual cancer samples [13] . Furthermore , many existing computational tools for drug combination analysis focus on the degree of interaction , i . e . drug synergy , but not the sensitivity of drug combinations . For example , Combenefit [14] and SynergyFinder [15] have been developed to provide multiple reference models to score drug synergy , such that drug combinations that produce higher growth inhibition effects compared to the single drugs will be prioritized . There have been multiple methods to score drug synergy [16] , while the choice of the best synergy scoring methods is under debate [7 , 17] . On the other hand , as synergy is a measure of drug interaction while sensitivity is a measure of drug combination efficacy , these two metrics are expected to capture distinct properties of a drug combination . Therefore , neglecting the drug combination sensitivity may lead to a biased prioritization of drug combinations that are unable to kill cancer cells despite strong synergy [18] . However , unlike the sensitivity of single drugs which can be directly derived from monotherapy dose-response curves [19] , the sensitivity of a drug combination remains largely undefined , as the same sensitivity can be achieved using different dose combinations . Furthermore , there is a lack of scoring approaches to fully capture the synergy and sensitivity simultaneously , which should be ideally interpretable using the same scale , e . g . percentage inhibitions [16] . To overcome these challenges , we proposed a cost-effective experimental and computational procedure to facilitate the prioritization of drug combination synergy and sensitivity . We proposed a novel experimental design to allow either drug to span over multiple doses while the concentration of the other drug is fixed at its IC50 concentration . The resulting drug combination dose-response curves were utilized to determine a drug combination sensitivity score ( CSS ) . Using a large-scale drug combination study , referred to as the O’Neil data [20] , we showed that the CSS is highly reproducible , suggesting its robustness to be utilized as a metric for characterizing drug combination responses . Furthermore , we found that the CSS can be predicted at high accuracy using chemical and pharmacological features of the drug combinations . To assess the degree of synergy from the cross design , we developed an S synergy score based on the difference between the observed CSS score and the baseline effect predicted by a reference model . As there is no consensus on the reference model , we evaluated multiple variants of it and found that the S score can detect the true synergistic and antagonistic drug combinations with high accuracy , irrespective of which the reference model is used . Compared to the full matrix design , the cross design requires minimal amount of experimental materials , while it still maintains a robust and accurate characterization of both drug combination sensitivity and synergy levels . We foresee that such a cross experimental design and its CSS and S scoring methods should allow a scale-up of drug combination testing especially for patient-derived cancer cells . The R scripts for calculating and predicting CSS are available at https://github . com/amalyutina/CSS .
We proposed a cross design to test the synergy and sensitivity of a drug pair by first introducing the concepts of background drug and foreground drug: background drug is the drug fixed at its IC50 concentration while foreground drug is added into the background drug with multiple concentrations . We allow either drug in the pair to be the background drug , so that two vectors of dose combinations will be intersected at the IC50 concentrations ( Fig 1A ) . The dose-response curves for these two vectors are usually measured in a unit of inhibition percentages by cell viability or toxicity assays . Note that the cross design requires specifically the combinations at the IC50 concentrations , which need to be determined based on the monotherapy dose response curves . As shown in Fig 1B , as long as a minimal of two concentrations are tested for a drug combination , the cross design will require less experimental materials than a full matrix design . For a combination screen that tests one concentration only , both the cross design and the full matrix design converge to a point estimate of drug combination effects at IC50 . Therefore , we considered that the cross design is in general more cost-effective compared to the full matrix design . With the drug combination dose-response curves determined in the cross design , the CSS summarizes the area under the curve similar to the scoring approaches [21–22] . Namely , a four-parameter log-logistic function is used to fit the dose-response curve for a concentration x of the foreground drug according to: y=ymin+ymax-ymin1+10λ ( log10IC50-log10x ) , ( 1 ) where ymin and ymax are the minimal and maximal percentage inhibition ( the bottom and top asymptotes of the curve , 0 ≤ y , ymin , ymax ≤ 1 ) ; IC50 is the concentration of the foreground drug with which the drug combination reaches 50% of ymax—ymin inhibition of the cell growth; λ is the slope of the dose-response curve . The dose-response curve ( 1 ) is transformed by substituting x with x' = log10 ( x ) as: y=ymin+ymax-ymin1+10λ ( log10IC50-x' ) ( 2 ) The area under the log10-scaled dose-response curve ( AUC ) is determined according to AUC=∫c1c2ymin+ymax-ymin1+10λ ( m-x' ) dx'=ymin ( c2-c1 ) + ( ymax-ymin ) 1λlog10 ( 1+10λ ( c2-m ) 1+10λ ( c1-m ) ) , ( 3 ) where [c1 , c2] is the log10 concentration range for the foreground drug tested in the experiment , and m = log10 ( IC50 ) . The AUC is further normalized as the proportion of its maximal possible inhibition ( i . e . 100% inhibition ) according to: AUC'=AUC–inhmin ( c2-c1 ) ( 1-inhmin ) ( c2-c1 ) , ( 4 ) where inhmin is the minimum inhibition level that is considered as the drug effect ( by default it is fixed at 10% , assuming that the inhibition below 10% is experimental noise , S1 Fig ) . The CSS for the foreground drug is defined as a percentage which varies between 0 and 100: CSS=100AUC' ( 5 ) As there are two drug combination dose-response curves depending on which drug is fixed as the background drug , we refer to the results of Eq ( 5 ) for either scenario as CSS1 and CSS2 . The two variants of CSS are expected to reflect similarly the summarized % inhibition for a given drug combination . We considered them as two samples that are generated from the same random variable , and estimated the CSS as an average of CSS1 and CSS2 , i . e . Dose-response was measured as percentage of cell viability and retrieved from the supplementary material of O’Neil et al . [20] , which includes 22 , 737 drug combinations that involve 38 unique drugs in 39 cancer cell lines , representing 7 tissue types . At the first stage , single-drug screening was done using 8 concentrations to determine the IC50 concentration for each drug with six replicates . At the second stage , a 4 by 4 dose matrix was utilized to cover the span of IC50 concentrations for a drug pair with four replicates . To utilize the cross design , we picked up only the row and the column corresponding to the concentrations closest to the IC50 of the single drugs . These two vectors thus allowed the fitting of drug combination dose-response curves with which the CSS can be calculated . The cell viability percentage was first transformed to inhibition percentage according to: %inhibition=100-%viability ( 7 ) In our analysis , the CSS for a drug combination was determined based on the average % inhibition of the four replicates . The robustness of the CSS scoring was assessed using the Pearson correlation across the four replicates . All the correlation analyses utilized Pearson correlations . With the CSS being determined for each drug combination , we sought to evaluate the prediction accuracy of multiple machine learning methods . We considered a drug combination as a combination of their targets and chemical fingerprints . We collected the known targets that have been experimentally validated for the 38 drugs from Drugbank [23] and ChEMBL [24] . Furthermore , we also utilized the Similarity Ensemble Approach ( SEA ) to predict additional secondary targets based on the chemical structures of the drugs [25] . The targets that were predicted with Z-score higher than 20 , Tanimoto coefficient higher than 0 . 4 and P-value smaller than 0 . 01 were included , following the previously reported filtering strategy [25] . The MACCS fingerprints of the drugs were determined using the SMILES strings with the R package rcdk [26] . The resulting feature set for a single drug included 398 validated and predicted targets and 166 MACCS fingerprints ( S1 and S2 Tables ) . The feature vector for a drug combination was determined as the bitwise OR operation over the features of its single drugs . We compared three state-of-the-art machine learning methods for the CSS prediction: Elastic Net [27] , Random Forests [28] and Support Vector Machines [29] . Elastic Net is a regularization and feature selection method that combines both ridge and lasso regression by including the L1 and L2 penalty terms , which are regulated by hyper parameters α and λ . The α parameter controls the penalty term in the elastic net by giving more power either to the lasso regression ( when α is closer to 1 ) or to the ridge regression ( when α is closer to 0 ) . In our studies , α was selected from the interval [0 . 1 , 1] , which determines the level of compromise between the lasso and ridge regressions . The λ parameter regulates the level of shrinkage and was chosen to minimize the difference between predicted and actual CSS scores . Random Forests is an ensemble learning method that constructs multiple decision trees . In our studies , we set the number of randomly selected predictors that is used at each split of the decision tree equal to the rounded down square root of the number of variables . We utilized Support Vector Machines with Radial Basis Function Kernel , which can be used not only for classification but for regression problems . The tuning parameters are the cost parameter C that sets the penalty for prediction error of a training point and a smoothing parameter σ , based on the loss function in cross-validation . We focused on the model performance for predicting new drug combinations within the same cell line , as the set of drug combinations in the training data did not overlap with that in the test data . For each of the 39 cell lines , the number of drug combinations ranges from 290 to 688 , with an average of 338 . We randomly sampled 70% of the drug combinations to train multiple machine learning models using 10-fold cross-validation , which splits the training data randomly into 10 equally folds , 9 of which were used to fit the model and the remaining one was used to evaluate the prediction accuracy . The model with the lowest RMSE out of the 10-fold cross validation was then used for predicting the CSS values for the remaining 30% of the novel drug combinations as the testing data . As the sampling was done randomly for each cell line , the training and the testing data were therefore balanced . Four metrics including coefficient of determination ( R2 ) , root mean square error ( RMSE ) , mean absolute error ( MAE ) and Pearson correlation ( COR ) were utilized for evaluating the prediction performance on the testing data . The whole procedure above was repeated 20 times for each cell line on the predefined seeds , and the final model performance was obtained as the mean values of these iterations . To benchmark the performance of the machine learning methods , we utilized one randomly selected technical replicate as the best possible prediction to obtain the upper limit of the performance . All the methods were implemented and evaluated using the R package caret [30] . The advantage of CSS is that it allows a direct comparison of the sensitivity between a drug combination and its single drugs , and hence facilitates the quantification of drug synergy . The degree of synergy is often calculated as the deviation of the observed drug combination effect from the reference , which is defined as the expectation effect if the drugs are not interacting . However , how much the expected effect should be is a matter of mathematical modelling with certain assumptions . As the choice of the ‘best’ synergy model is rather heuristic , we proposed three variants of CSS-based synergy scores ( termed as S scores ) by assuming the reference model as the sum , the maximal and the mean of the AUCs for the monotherapy drug responses: Ssum=CSS-sum ( AUC1 , AUC2 ) , ( 8 ) Smax=CSS-max ( AUC1 , AUC2 ) , ( 9 ) Smean=CSS-mean ( AUC1 , AUC2 ) , ( 10 ) The AUC for a monotherapy drug response was defined according to [22]: AUC=a ( x-c+log10 ( 1+10b ( c-x ) ) b ) log10a , ( 11 ) where [a , b , c] were the parameters to fit a logistic function on the single drug response at concentration x: y=a1+10b ( c-x ) ( 12 ) To evaluate the prediction accuracy of the S synergy scores , we defined a set of true synergistic and antagonistic drug combinations as the gold standard , which were determined using the full dose-response matrix data including the combination and monotherapy responses . We utilized the R package synergyfinder [31] to calculate multiple versions of synergy scores including the HSA ( Highest Single Agency [32] ) , the Bliss [33] , the Loewe [34] and the ZIP synergy scores [35] . The principles of these four models are summarized below: Consider that drug 1 at concentration x1 and drug 2 at concentration x2 were combined to produce the inhibition effect of yc , while their respective single drug effects were y1 ( x1 ) and y2 ( x2 ) . The synergy score was calculated as the difference between yc and the expected effect ye if there is no synergy . Each synergy scoring took a different model for ye: For each of the four models , the synergy scores were determined first for a given dose combination and then were averaged over the full dose-response matrix . With the four synergy scores determined for each drug combination , the true synergistic and antagonistic drug combinations are those with all four synergy scores consistently higher than 5 and lower than 5 , respectively . The aim was then to use the S synergy score which was determined by the cross design data to predict the ground truth determined by the full matrix design . The areas under the ROC curve and the precision-recall curve were used for evaluating how well the S synergy scores can predict the consensus drug combinations determined using the full dose-response matrix data .
We applied the CSS scoring on the O’Neil drug combination data , which consists of 22 , 737 drug combinations for 39 cancer cells [20] . We found that the CSS1 and CSS2 values calculated using either drug fixed at its IC50 concentration were highly correlated ( Pearson correlation = 0 . 82 , p-value = 2×10−16; Fig 2A ) . Both CSS1 and CSS2 values ranged from 0 to 50 , with a marginal absolute difference of 5 . 62 ( Fig 2B ) . As a CSS score can be directly interpreted as a normalized average % inhibition of the drug combination response ( Eqs 1–6 ) , such a result implies that about 5% inhibition difference is expected between CSS1 and CSS2 . The high level of consistency holds true for all the 39 cancer cell lines and the majority of the 38 unique drugs , suggesting the robustness of the CSS scoring method ( Fig 2C and 2D , S2 Fig ) . To evaluate the robustness of the CSS values further , we permuted the O’Neil data randomly and recalculated the correlations between CSS1 and CSS2 . The correlations deteriorated quickly to near zero ( Pearson correlation = 0 . 075 , S3 Fig ) , suggesting that the high correlation can be attributed to the robustness of CSS on the actual drug combination data . We also found high correlation between the CSS value and those derived from individual replicates ( minimal Pearson correlation = 0 . 97 , S3 Table ) . In order to check whether the CSS values are within the range of the CSS replicates for each drug combination , we calculated the minimal and maximal values over the CSS replicates for each drug combination and plotted them together with CSS values over the standard deviation of the CSS replicates . For a better visualization , we applied a generalized additive model to smoothen the CSS lines and obtain 95% pointwise confidence interval around the mean ( S4 Fig ) . Only 4% of the drug combinations have the CSS values being out of the CSS replicate-based limits , however this can be explained by the higher variance over the replicates . Notably , we found that drug combinations that involved bortezomib showed much lower correlation ( 0 . 26 ) between the CSS1 and CSS2 values compared to other drug combinations . Since the O’Neil data contains the replicates for single drug screening , we analyzed the coefficient of variation ( CV ) of the cell viability readout for each drug in the replicates . As expected , we found that bortezomib has the highest CV ( 0 . 26 ) , suggesting a relative low quality of the drug combination sensitivity data involving this drug ( S5 Fig ) . When filtering out the drug combinations with a decreasing threshold of difference between CSS1 and CSS2 , the remaining drug combinations showed an increasing CSS1 and CSS2 correlation ( S6 Fig ) . We found that the threshold of 10 is close to the middle point that reached the correlation of 0 . 91 ( i . e . the average of 1 and 0 . 82 ) . We therefore applied an empirical threshold of 10 to filter out the drug combinations that were in poor quality , as CSS1 and CSS2 in these combinations showed bigger than 10% inhibition difference . In total , there were 18 , 905 drug combinations after the filtering , constituting the 83 . 1% of the original data . As expected , the correlation between CSS1 and CSS2 was further improved ( Pearson correlation = 0 . 93 , p-value = 2×10−16 ) . Furthermore , the mean absolute difference between CSS1 and CSS2 was 3 . 83 , which was comparable to the variability determined from the technical replicates of CSS1 and CSS2 ( 2 . 92 and 3 . 06 respectively ) , suggesting that the difference between CSS1 and CSS2 is similar to what is expected when repeating the experiment . Taken together , CSS1 and CSS2 values are highly consistent and therefore supported their averaging as a summary for the drug combination sensitivity score . Given that the CSS is highly reproducible as a summary of the overall sensitivity of a drug combination , we explored whether CSS can be predicted using pharmacological and chemical information of the drugs . We considered a drug combination as a combination of its drugs’ target profiles as well as their chemical fingerprints , with which the machine learning approaches illustrated in the previous section can be optimized by exploring the feature space using the training data . We examined three major machine learning methods for predictions: Elastic Net , Random Forests and Support Vector Machines . We found that all of these machine learning approaches worked reasonably well , where Elastic Net consistently achieved the best performance , with a mean MAE of 4 . 01 which is comparable to that ( 2 . 07 ) of a technical replicate ( Table 1 ) . Note that in our cross-validation setting the drug combinations in the test data were not present in the training data , however , the machine learning methods were still able to predict the CSS values for new drug combinations by exploring the feature similarity in the drug targets and chemical fingerprints . The prediction performance thus validated our hypothesis that a drug combination can be considered as a combination of their drug target profiles and chemical-structural properties , with which the CSS score can be predicted with high confidence using state-of-the-art machine learning approaches . Since both the drug target profiles and chemical fingerprints were considered as the drug combination features , we next evaluated their prediction performances separately using the Elastic Net method . For drug-target profiles we collected known targets that were experimentally validated as well as the additional secondary targets that were predicted with high confidence using the SEA method . For chemical fingerprints we used the MACCS fingerprint which contains 166 structural features [36] . As expected , when combining all the features the model achieved the best performance ( Table 2 ) . We found that in general drug target profiles were predictive of CSS , especially when including the experimentally validated targets . The predicted targets using the SEA method did not improve the prediction accuracy significantly , indicating that even though secondary drug target interactions may occur , most likely they have minor functional impact that may not lead to changes in cancer cell viability and thus does not contribute to the prediction of CSS . On the other hand , we found that chemical fingerprints were less predictive of CSS compared to the drug-target profiles , suggesting that the use of MACCS might be suboptimal to capture the relevant structural information for predicting the drug combination sensitivity . However , as the focus of this study was to show the validity of using machine learning methods to predict the CSS score , we decided to explore other chemical fingerprint features as a future step . We considered the regression coefficients that were determined in the Elastic Net model as an indication of their importance to contribute to the CSS prediction . We found that certain drug target features were present with high coefficients across all the cell lines ( Fig 3 ) . For example , DNA topoisomerases including TOP1MT , TOP2A and TOP2B and TOP1 were selected , with average coefficients of 8 . 2 , 2 . 7 , 2 . 6 and 1 . 0 , respectively . Despite the difference in the level of variable importance , all the DNA topoisomerases showed positive coefficients in 38 of 39 cell lines , suggesting that targeting DNA topoisomerases were associated with a higher CSS . DNA topoisomerases are known proteins which are essential for cell replication and metabolism [37] . Including a topoisomerase inhibitor can thus enhance the drug combination sensitivity in many cancer cell lines . On the other hand , the only cell line that showed negative coefficients for TOP1MT was LNCAP ( prostate cancer ) , which turned out to be the cell line that has the smallest average CSS scores for drug combinations involving the TOP1MT inhibitor ( topotecan ) ( S7 Fig ) . If the CSS profiles for two cell lines are similar , then their feature importance vectors are expected to be similar . We focused on the most important features that have their absolute coefficients greater than 3 for at least one cell line , resulting in 67 top features . We then utilized these feature importance scores to cluster the cancer cell lines , using unsupervised hierarchical clustering with the Euclidean metric . We found that cell lines of the same tissue type did not necessarily cluster together , indicating their distinctive drug combination response profiles . For example , we found that breast cancer cell lines did not form a single cluster due to the outlier MDAMB436 . Indeed , MDAMB436 is the only triple negative breast cancer ( TNBC ) subtype , while the other cell lines are either ER positive ( KPL1 , ZR751 and T47D ) , or HER2 positive ( EFM192B and OCUBM ) . It has been known that TNBC respond to anticancer drugs differently from ER and HER2 positive breast cancers due to distinct disease mechanisms [38] . The top drug combination features separated these two distinctive breast cancer subtypes , suggesting the validity of using CSS-predictive features to cluster cancers of different subtypes . Furthermore , we found that AKT targets ( AKT1/2/3 ) were among the top features that showed higher importance in the non-TNBC group . A combination of an AKT inhibitor and a TOP1MT inhibitor therefore can be proposed to treat non-TNBC , but not necessarily for TNBC breast cancers . On the other hand , we found that CHEK1 and PARP3/4 targets were selected for the TNBC cell line MDAMB436 but not for the non-TNBC group , suggesting that a combination of a CHEK inhibitor and PARP inhibitor might be effective for TNBC . The mechanisms of actions for the proposed drug combination may prove interesting for experimental validations . Taken together , the pharmaco-features that were determined from the CSS prediction may help pinpoint the underlying target combinations , which are of pivotal importance to explain the drug combination responses . Next , we defined the degree of drug synergy as the differences between the dose-response curves of a drug combination and its single drugs . We derived three variants of the S synergy score ( Ssum , Smax , Smean ) and compared them with the HSA , Bliss , Loewe and ZIP synergy scores that were determined using the full-dose response matrix . Although being determined using only one row and one column from the dose-response matrix , all the S synergy scores managed to obtain a good correlation with the synergy scores based on the full matrix design ( Table 3 ) . We found that the S synergy scores correlated relatively well with the HSA and Bliss scores , while the correlation started to decrease when comparing to the Loewe and ZIP scores . Since all the synergy scoring models utilized different assumptions for the reference of no synergy , we therefore did not expect a perfect correlation in such pairwise comparisons . For example , the Bliss model assumes that two non-interactive drugs act independently while the Loewe model assumes two non-interactive drugs act as one drug . Their differences in mathematical models have been discussed in our previous publications , such as [16] and [35] . Of all the three variations of S synergy score , we found that Ssum showed the best correlation with those determined using the full dose-response matrix . As Ssum considers the additive effect of single drug sensitivities as the expectation of no synergy , it thus can be considered a more conservative scoring method compared to Smax and Smean , where the maximal and average effect of single drugs were considered as reference separately . To control the false discovery rate of detecting synergistic combinations , we therefore proposed Ssum as a more appropriate scoring method for the cross drug combination design . Next , we evaluated the predicting accuracy of the S synergy scores for true synergistic and antagonistic drug combinations . To be able to formulate a binary classification problem , we first selected the true positive and true negative drug pairs by applying stringent criteria to determine the ground truth from the full dose response matrix data . The rationale of the ground truth was based on the assumption that full dose-response matrix can capture the truly synergistic and truly antagonistic drug combinations , as it allows the full factorial testing of all the possible doses for a given drug combination . However , as there exist different models for synergy scoring , we decided to apply the most stringent criteria to determine the ground truth , such that a truly synergistic or truly antagonistic drug combination has to fulfill the criteria of all the four existing synergy scoring models ( HSA , Bliss , Loewe and ZIP ) . Namely , the drug combinations with all the four synergy scores ( HSA , Bliss , Loewe and ZIP ) higher than 5 , or lower than -5 , were classified as true synergistic or antagonistic drug combinations , respectively . The threshold of [–5 , 5] was determined by the empirical distribution of the synergy scores in the O’Neil data , assuming that most of the drug combinations are non-interactive ( S8 Fig ) . Furthermore , we considered that a synergy score of [–5 , 5] range corresponds to a [–5 , 5] % inhibition which might be due to experimental variation , as indicated by the analysis of replicates from O’Neil dose-response matrix data ( mean standard deviation of % inhibition is 6 . 6% ) . Therefore , any synergy score between [–5 , 5] may be simply experimental noises . From the O’Neil data , we showed that 78 . 1% of the drug combinations were within the [–5 , 5] range by at least one synergy scoring , suggesting that the majority of the drug combinations indeed should not be reliably considered as true synergistic nor true antagonistic ( S8 Fig ) . As a result , we identified 3 , 716 true synergistic and 218 true antagonistic drug combinations . All the other drug pairs were considered as non-interactive and thus excluded from the analysis . Once the ground truth had been determined using the full dose-response matrix data , we then asked the question: Can the S synergy scores that were determined using the cross design correctly identify the most significant synergistic and antagonistic drug combination hits that were confirmed using the full matrix design ? We showed that the S synergy scores achieved the area under the ROC curves of 0 . 997 ( Ssum ) , 0 . 996 ( Smax ) and 0 . 992 ( Smean ) to detect the true synergistic and antagonistic combinations ( Fig 4A ) . The area under the precision-recall curves were 0 . 9997 ( Ssum ) , 0 . 9995 ( Smean ) and 0 . 9998 ( Ssum ) , suggesting that the S scores retrieved a majority of synergistic combinations with minimal false positive and false negative rates ( S9 Fig ) . The S synergy score was derived from the cross design , where only two vectors of the drug combination responses are needed . Still , the S synergy scores can predict the most synergistic and antagonistic drug combinations with high accuracy . These results showed that the cross design can be reliably utilized as a cost-effective strategy in a primary screen to detect the most significant synergistic or antagonistic drug combinations . On the other hand , the S synergy score and the CSS drug combination sensitivity score utilize the same unit as the percentage of inhibition of cell viability . Therefore , the synergy score can be interpreted as the extra percentage inhibition effect beyond the expectation . We summarized both CSS and S scores for all the drug combinations as an S ( sensitivity ) -S ( synergy ) plot ( Fig 4B; S4 Table ) . By applying a threshold of the 3rd quantiles for CSS and S , we can clearly identify the most promising drug combinations that fulfill both the sensitivity and synergy criteria , while avoiding the false positive drug combinations that might be synergistic but do not achieve a sufficient high level of sensitivity ( S10 Fig ) . Taken together , the combined use of CSS drug combination sensitivity score and the S synergy score allows a simultaneous evaluation of the sensitivity and synergy for a drug combination , which will facilitate a more systematic analysis of high-throughput drug combination data with much less experimental materials .
Drug combinations may potentially lead to more durable clinical responses by overcoming intra-tumoral heterogeneity and drug resistance to monotherapies . Identifying drug combinations that are tailored for personalized medicine is a challenge , as the number of possible combinations may easily grow exponentially [39] . High-throughput drug combination screening has been increasingly utilized for early detection of true synergistic and effective drug combinations . However , systematic identification of drug combinations is difficult , as the concepts of synergistic versus effective drug combinations are often intertwined and sometimes interchanged without sufficient clarification . Furthermore , there is a lack of consensus on what the definition of synergy and sensitivity are , which might contribute to the poor reproducibility of many drug combination studies . The uncertainty about the endpoint measurement in drug combination screens brings additional complexity for any machine learning approach to tackle the prediction problem . We developed a novel scoring approach called CSS for drug combinations that can be efficiently determined using the cross design . We found that the CSS is highly reproducible and therefore can be considered as a robust metric to characterize drug combination sensitivity . To understand the drug combination sensitivity , we implemented a systematic evaluation of the prediction accuracy of three machine learning methods . We showed that machine learning in general worked well for the prediction of CSS , where Elastic Net showed the best performance according to our cross-validation setting . We found that the drug target information for the compounds as well as their chemical fingerprints are highly predictive of the CSS values , with an accuracy comparable to the level of experimental replicates . Therefore , the rationale of considering a drug combination as a function of their target and fingerprint profiles can be justified . This would also allow the augmentation of single-drug screening and drug combination screening data together to train a machine learning model , as many drugs are multi-targeted which are equivalent to a drug combination with the same target profile . In our study , we utilized the SEA method to predict new targets of compounds , by applying a combination of thresholds of Z-score , Tanimoto coefficient and p-value suggested by the authors of SEA [25] . In addition to the predicted targets , we also included the known primary and secondary targets of the compounds , so that the risk of false negative is minimal . However , we could not find significant improvement on the prediction accuracy when including the SEA-predicted secondary targets ( Table 2 ) , suggesting that the unknown targets for a compound may contribute minimally to the drug combination sensitivity . In this study we focused on drug combination prediction within the same cell line . In the future , we could include the molecular features of the cell lines to improve the prediction accuracy , aiming to identify drug combination specific biomarkers across different cell lines . On the other hand , as the focus of the study is to propose the new experimental design and to justify its associated drug combination scoring methods , we tested the predictability of CSS using only conventional machine learning methods with standard cross-validation schemes . A more comprehensive evaluation of machine learning approaches may be developed by including multiple cross-validation schemes and data pre-filtering techniques . More advanced machine learning methods such as Deep Learning [13] or network-based methods [40] may further improve the prediction accuracy as well as help the biological understanding of the mechanisms of action . A truly promising drug combinations shall reach sufficient therapeutic efficacy via a strong sensitivity and synergy . Therefore , both these aspects should be evaluated for the prioritization of most potential drug combination hits . While there have been multiple synergy scoring methods that can be applied to the full matrix design , they do not always produce consistent results . The truly synergistic and antagonistic drug combinations may therefore be determined by achieving the consensus across the different scoring methods [16] . Based on the CSS drug combination sensitivity scoring , we developed an S synergy score to quantify the degree of interactions in a drug pair , and showed that it can identify truly synergistic and antagonistic drug combinations accurately . Tailored for the cross design , the S synergy score can be used for the prioritization of a primary drug combination screen , after which only the most significant drug combinations should warrant a confirmation screen using the full matrix design . Furthermore , we proposed a novel S-S plot ( Fig 4B ) to visualize drug combination sensitivity and synergy using the same scale , which enables an unbiased way to explore high-throughput drug combination data more efficiently with minimal bias . Notably , the CSS is defined at the IC50 concentrations of the background drugs . Therefore , a synergistic drug combination determined by the S score should be more therapeutically relevant than a drug combination where the synergy is detected at higher concentrations , which are often associated with unwanted off-target effects and side-effects . The proposed cross design aimed for alleviating the limitation of the conventional dose-response full matrix design which usually requires a large amount of cancer cells that may not be amenable especially from patient-derived samples . Empowering the cross design with the CSS and S scoring methods , we are foreseeing a lower technical barrier to carry out large-scale drug combination studies with minimal cellular materials . Although we showed the proof-of-concept using the drug combination data involving cancer cell lines , the cross design coupled with the CSS and S scoring methods can be readily applied for ex_vivo drug screening , where the amount of patient-derived materials can be extremely limited and technically difficult to obtain due to culture constraints [41] . While the majority of drug combination screens are limited to cytotoxic and molecularly targeted molecules , the cross design should be also favored for the testing of combinations that involve immunotherapies and antibodies . Furthermore , the cross design is not only applicable for cancer but also for other diseases , as long as cellular phenotypes of interests can be measured at multiple dose levels . With the help of cross design and its data analysis tools , drug combination discovery can be more quickly advanced and may eventually lead to the validation of personalized drug combinations in clinical trials .
|
Cancer is one of the main causes of death worldwide . Although new treatment strategies have been achieved , they still have limited efficacy as cancer cells can easily develop drug resistance . To achieve more sustainable therapies to treat cancer , we need multi-targeted drug combinations that can inhibit cancer cells more effectively and synergistically . However , the increasing number of possible drug combinations makes a full matrix design unfeasible , even with automated drug screening instruments . Therefore , we proposed a novel cross design to access drug combinations more efficiently . We further developed a drug combination sensitivity score ( CSS ) that is tailored for the cross design to quantify the efficacy of a drug combination . Using public datasets , we showed that the CSS is a robust metric and highly predictive with an accuracy comparable to the experimental replicates . We also developed a CSS-based synergy score to assess the degree of drug interaction and showed its capability to correctly identify synergistic and antagonistic drug combinations . Taken together , we showed that the cross design and its scoring methods allow a more systematic and cost-effective evaluation of drug combinations . The proposed experimental and computational techniques are expected to be widely applicable in the field of drug combination discovery .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"oncology",
"computer",
"and",
"information",
"sciences",
"medicine",
"and",
"health",
"sciences",
"machine",
"learning",
"drug",
"research",
"and",
"development",
"pharmaceutics",
"artificial",
"intelligence",
"cancer",
"drug",
"discovery",
"dose",
"prediction",
"methods",
"cancer",
"treatment",
"pharmacology",
"drug",
"discovery",
"basic",
"cancer",
"research",
"drug",
"interactions",
"drug",
"screening"
] |
2019
|
Drug combination sensitivity scoring facilitates the discovery of synergistic and efficacious drug combinations in cancer
|
Patterns of spontaneous activity in the developing retina , LGN , and cortex are necessary for the proper development of visual cortex . With these patterns intact , the primary visual cortices of many newborn animals develop properties similar to those of the adult cortex but without the training benefit of visual experience . Previous models have demonstrated how V1 responses can be initialized through mechanisms specific to development and prior to visual experience , such as using axonal guidance cues or relying on simple , pairwise correlations on spontaneous activity with additional developmental constraints . We argue that these spontaneous patterns may be better understood as part of an “innate learning” strategy , which learns similarly on activity both before and during visual experience . With an abstraction of spontaneous activity models , we show how the visual system may be able to bootstrap an efficient code for its natural environment prior to external visual experience , and we continue the same refinement strategy upon natural experience . The patterns are generated through simple , local interactions and contain the same relevant statistical properties of retinal waves and hypothesized waves in the LGN and V1 . An efficient encoding of these patterns resembles a sparse coding of natural images by producing neurons with localized , oriented , bandpass structure—the same code found in early visual cortical cells . We address the relevance of higher-order statistical properties of spontaneous activity , how this relates to a system that may adapt similarly on activity prior to and during natural experience , and how these concepts ultimately relate to an efficient coding of our natural world .
The classic debates of nature vs . nurture , or innate vs . learned , are pervasive in the literature of early visual development . A variety of studies have shown that the visual system requires external experience to mature ( e . g . , [1]–[3] ) . On the other hand , many animals are able to see at birth , and have a functioning primary visual cortex even before eye opening ( e . g . , [4] , [5] ) . It might seem straightforward to assign the properties found at birth to be innate and the properties dependent on visual experience to be learned . However , a strict dichotomy may unnecessarily limit our integrated understanding of visual development . In particular , we wish to focus on the issue of a form of learning that occurs before birth on patterns of activity that are generated internally . It is well known that spontaneous endogenous activity is necessary , or permissive , for the proper development of the visual system ( see [6] for review ) . The point of this paper is to discuss the statistical aspects of this activity that may be sufficient , or instructive , to guide development in much the same way that visual experience refines the mature visual system . Essentially we propose an “innate learning” approach which prepares the system for later experienced-based refinement – a diplomatic balance between nature and nurture . Several Studies have shown that in the early stages of rat visual development , retinal neurons are spontaneously active and correlated in their bursting patterns of activity [7] , [8] . Later , these retinal wave patterns were recorded from many animals by calcium imaging in the developing retina [6] , [9] , with one example shown in figure 1A . Experiments since then have manipulated these waves by abolishing them , over-stimulating them , or otherwise altering their properties and have shown how they are necessary for proper development [10]–[15] . Several models have been proposed for the production of these waves [16]–[18] . For the two most recent models , cholinergic amacrine cells mediate this activity with general agreement about the mechanism . Neurons begin bursting spontaneously , while neighboring cells can be recruited if enough cells in the local area are also bursting . With such rules for wave formation and propagation , biologically plausible models of retinal wave formation have been able to create complex images , such as those in figure 1B . Although retinal spontaneous activity has been well studied , many areas beyond the retina exhibit patterned , spontaneous neural activity . In the visual system , both the LGN [19] and V1 [20]–[22] have patterned , spontaneous activity during development . The effects on LGN and V1 connectivity have been analyzed functionally by layer segregation and orientation column formation [23] , [24] . Patterned , spontaneous activity is also known to occur in the developing auditory system and is necessary for proper development [25] , [26] . Similar developmental mechanisms are also found in hippocampus [27]–[29] and spinal cord [30] , [31] . From a biophysical perspective it has been shown that spontaneous neural activity is necessary to mediate many mechanistic effects such as axon branching [32] , dendritic patterning [33] , and synaptic pruning [23] , [34] . With the ubiquity of spontaneous activity in development and its ability to affect various aspects of neural connectivity , understanding the general role of spontaneous activity in early visual development is likely to have implications beyond vision . In adult primary visual cortex , it has been known for nearly half a century that V1 cells respond strongly to bars and edges [35] with later experiments demonstrating that simple cells in V1 have a characteristic filter description much like a 2D gabor function [36] , [37] as shown in figure 2A . The V1 cell has specific elongated subregions of visual space where relatively bright or dark parts in the visual image will stimulate the cell . Note that this characterization is purely descriptive as a stimulus-response paradigm by answering “what” the neuron responds to instead of “why” the filters have that appearance . According to the efficient coding hypothesis , the role of the early visual system is to remove statistical redundancy in the visual code [38] , [39] . From this hypothesis , one way to understand the visual system is to develop and analyze a visual encoding scheme to remove the redundancy in images of natural scenes . This was done using sparse coding [40] and independent components analysis ( ICA ) [41] on a set of natural images – pictures of rocks , trees , forest scenes , etc . . . The derived filters resemble the 2D gabor filters found in V1 simple cells – see figure 2B-C . One can conclude from these results that V1 is developed and tuned to efficiently encode the visual world . In this paper , we make the claim that there is a parsimonious computational reason for the existence of spontaneous patterns - a functional strategy that the early visual system can employ to guide this development both prior to and throughout experience . In addition to molecular guidance cues we believe the visual code is refined from training on patterns of spontaneous activity during development in a similar manner to how the juvenile animals refine the visual code on statistical patterns found in natural images . Many statistical structure models rely on the pairwise correlations between neighboring units ( also known as second-order statistics ) – an implicit assumption in other functional descriptions of spontaneous activity [42]–[47] . However , many efficient coding models applied to natural images , such as sparse coding and ICA , rely on statistics beyond pairwise correlations . In fact , often as a first step these correlations are removed in a process known as decorrelation or “whitening” ( e . g . , [40] , [41] , [48] ) ; a process that at least in part is considered a function of retinal ganglion cells [49] ( see [50] for a discussion ) . Although the developmental activity patterns are known to have relevant pairwise correlations , we argue receptive field refinement may also rely on higher-order statistics – thus bridging the gap between models of sparse , efficient coding and spontaneous activity . We will demonstrate that simple patterns of activity can be used as training images for refining the visual code . The patterns we use resemble the only 2D imaged spontaneous activity available – retinal waves; this is demonstrated in figure 1 , with specific examples of our generated patterns in figure 1D . Beyond a visual resemblance , our pattern generation technique also abstracts from the general properties and parameters of the current retinal wave models . We strongly note , however , that this is strictly not a retinal wave model but an abstraction of what we believe are the essential features of the relevant endogenous activity . We are more concerned in this paper with the statistical nature of the produced activity than its precise localization – including whether the activity originates in one particular area or is part of a larger , dynamical system . For example , in comparison to retinal waves , LGN/V1 spontaneous activity has a more direct influence on cortical receptive field formation . In ferrets the LGN remains spontaneously active at the beginning of V1 activation , while V1 activity and retinal wave production do not significantly overlap in time [6] . LGN and V1 activity have been experimentally characterized [19] , [22] , but are far less understood than retinal activity , thus prompting our analogies to retinal waves in this paper . Our patterns are generated using a variant of traditional site percolation models [51] - the analogy to retinal wave propagation and its relation to physiological models is detailed in the discussion section . Models common to the study of critical phenomena in physics , such as percolation models or the Ising model , have been used in artificial neural networks and understanding adult retinal neurons and can be equally useful in understanding models of development . Ising models , for example , have been adapted as artificial neural networks since Hopfield's network [52] . Recent work has also shown that Ising models are apt analogies for the maximal entropy and high-predictability neural firing in the retina upon natural stimulation [53] . Although the pattern generation technique we use is quite abstract , similar networks have been shown to be relevant biologically and demonstrate desirable statistical properties . The main goal of this paper is to show how the same adaptive , efficient algorithm can be applied for both natural inputs as well as spontaneous activity . We show that certain wavefront-containing patterns possess the relevant statistics and a percolation network provides a useful abstraction for demonstrating this property . These patterns , independent of how they were generated , can simply be used as an existence proof for the possible training role of spontaneous activity . First , we will show our generated patterns qualitatively resemble known patterns of spontaneous activity . We will then compare various methods of learning V1 receptive fields –showing how both natural images and spontaneous activity patterns can be used to produce V1-like gabor filters . We will also demonstrate how significant variations of receptive field properties can occur even at the threshold for scale invariance – showing flexibility of learning even for this simplified model . Finally , one of the main points of this paper , as expressed in the final figure , is that the relevant statistics for sensory coding go beyond simple correlations . There are higher-order statistics which are still present after decorrelation . Sparse and independent efficient coding algorithms rely on these statistics , which are found in natural scenes and are also present in the particular amorphous , wavefront-containing structure of spontaneous activity patterns . We will present how this fact points to the conclusion that the same adaptive coding strategy may then be present both before and during visual experience .
We believe the relevant statistical properties for an efficient “innate learning” strategy are present in a wide class of amorphous , wavefront patterns in which current models of spontaneous activity belong . We hope to demonstrate this generality , and avoid the pitfalls of selecting a particular physiological model , by using an abstract technique for pattern generation . This technique , described in detail in the methods section , can be summarized as a simple , three parameter model – a threshold , site percolation network model . Despite its abstract nature , this technique is analogous to known spontaneous activity patterns in generation and final pattern statistics , as mentioned in the discussion section . We began by exploring the parameter space for a suitable training pattern by varying the proportion of nodes which are able to spread activity , p , and the threshold of active neighboring nodes needed to initiate activity , t . For a fixed t , there is a clear phase transition , the critical percolation threshold , pc . For p>pc activity would spread over the whole image , and in the extreme case only a few small areas would remain inactive . At p<pc , active clusters would be finite in size , and in the extreme would be exceptionally small clusters – approaching random noise . Although not strictly a property of physiological spontaneous activity patterns , we were interested in scale-invariant patterns in this model . For this reason , sampling was done at p = pc ( along the phase transition boundary ) as shown in figure 3A . Approximate scale invariance is a property shared with natural images [54] . In this case , it also allows neurons with limited dendritic fields to produce consistent , large-scale statistical effects . We also chose this sampling as a mathematical convenience so results would not require a defined scale of analysis . Note that known spontaneous patterns - such as retinal waves - are clearly not on this self-similar boundary , but may be considered close , with many species having limited wave sizes , and others - such as chick retina - covering large areas of retina and often terminating at the edges [55] . The next step was to find if these patterns could be used to train an efficient coding system for natural vision . Sparse coding and ICA have been used to find approximately independent codes for natural images with resulting filters resembling those found in primary visual cortex , as shown in figure 2A-C . Figure 2C shows the filters derived from natural images given the parameters of image patch collection and coding as detailed in the methods section . Following the main thesis of this paper , one might ask if an efficient coding of activity from more physiologically precise models is capable of producing similar V1-like filters . To show this , we efficiently encoded thresholded , time-lapsed retinal wave images as in figure 2 of Godfrey and Swindale [18] . These moderately resemble images of experimentally determined retinal wave extent as shown in figure 1C ( from figure 1 of [56] ) . The resulting V1-like filters from this data are shown in figure 2D . Although an efficient coding of this model qualitatively produces physiological filters , we would like to demonstrate that these images are embedded in a larger class of amorphous , wavefront-containing patterns capable of producing relevant filter properties . We believe the question of whether or not the activity comes from a particular model - or even from the retina vs . the LGN/V1 - is important , but we would like to stress the necessary statistical properties independent of the particular source . To demonstrate this we generated a set of images from our abstract pattern generation technique with the resulting filters shown in figure 2E for comparison . Note how the general statistical structure of natural scenes , our abstract patterns , and more physiological models of spontaneous activity all produce filters resembling those found in V1 . To further demonstrate the ability of these amorphous , wavefront patterns to generate physiological filters , we generate sets of images along the phase-transition boundary . Filters derived from a representative sample are shown in figure 3C . A qualitative difference between the gabor filters is visible , and we analyzed at least one aspect of these filters – the orientation bandwidth . We chose orientation bandwidth because it is a well-defined , physiologically measured parameter . We fit 7 parameter gabor filters to the 16×16 pixel derived filters . After this fit , we used the parameters of the gabor fits to find the orientation bandwidth , with histograms of these fits shown in figure 3E , along with the primate physiological median [57] . We also coarsely explored the area below the phase transition boundary for this parameter; the transparent color contour in figure 3A indicates how the median orientation bandwidth changes in this region . Note for p<pc a manipulation of ‘p’ is more effective at changing the orientation bandwidth than ‘t’ – one indication of how models such as this one could lead to testable predictions through pharmacological manipulations . However , we do not intend to stress a direct comparisons to physiological filters; we know that even within neurophysiological literature , orientation and spatial frequency bandwidth decreases as newborn macaques age [58] complicating direct comparison . We show that even with this simple generation technique and imposed self-similarity constraint , a significant variation of filters can be produced . This variation demonstrates one way a method like this may adjust local parameters to affect global pattern statistics and more closely resemble properties of adult physiological filters and natural scene efficient coding filters . Current models proposed to explain pre-experience cortical receptive field development rely primarily on hebbian mechanisms and pairwise correlations . These approaches do not address the relevant statistical structure for receptive field formation related to efficient coding . Although hebbian models are capable of achieving arbitrary levels of complexity - and can even implement sparse coding strategies in specific configurations [59] - we note that the fundamental computational insight of hebbian models relies on pairwise correlations . In figure 4 , we address the fundamental differences between these second order and higher order correlations with respect to relevant statistical structure and receptive fields . Note that uncorrelated noise ( “white” noise ) has no second order or higher order statistics , so techniques that rely on pairwise correlations , as in PCA , or higher-order statistics , as in ICA , do not produce filters with discernable structure . In patterns with only second order correlations , as in the random 1/f patterns ( “pink” noise ) , PCA can produce relevant filters . However , in these 1/f noise patterns the sparse structure on which ICA relies is not present , and structured filters do not form . For the natural and our patterned images in this figure , we have partially removed the second order correlations by a procedure to flatten a 1/f slope in the Fourier amplitude spectrum . This removes the correlations in images that have an approximately 1/f slope , as was shown to be the case for natural images [54] . When we whiten the images by removing the pairwise correlations , PCA bases resembling receptive fields are , by definition , unable to form , and we see that natural images as well as the wavefront patterns still retain important image structure . This whitened structure , reminiscent of line drawings , is efficiently encoded using ICA . Also note that for these image sets the ICA filters are inherently localized within the filter patches . However , encoding using PCA will not produce localized filters without the use of additional constraints . Whichever encoding scheme is used , it should be noted that the generated wavefront patterns have both pairwise correlation structure as well as sparse , edge-like structure used by ICA . If only correlations were necessary to prepare the visual system , there are a number of even simpler ways to create these correlations without the additional wavefront , edge-like structure . This additional , higher-order structure can be exploited by the visual system to guide receptive field formation and maintenance . The fact that it exists in both spontaneous activity patterns and natural scenes suggests that both endogenous and external activity may use the same method of receptive field adaptation .
The point of this paper is to show how seemingly random patterns of activity can be used as training patterns for the visual system before eye opening . We believe that real spontaneous activity patterns are part of a class of amorphous , wavefront-containing patterns with the relevant efficient coding statistics . The patterns we create are also part of this class but abstract out the minimal , essential features while still retaining some biological plausibility . This pattern generation technique is of interest for the following reasons: 1 ) conceptual and analytical simplicity , 2 ) statistical properties – both self-similar/correlation-based and sparse coding/edge-like structure , and 3 ) biological plausibility . First , the technique is a simply stated three parameter model , collapsing to a one parameter model if you fix the neighborhood radius ( r = 3 , here ) and require fractal self-similarity ( p = pc ) . Also , this technique is not only conceptually simple , but simple to implement given a biological substrate of dendritic fields , local activity pooling , and activation thresholds . Second , the statistical properties have been discussed in detail – this pattern generation technique is capable of extremes from complete noise to clusters of activity to full activation; self-similar fractal patterns with similar statistics at all spatial scales; and edge statistics which vary the fractal dimension of the edges and consequently the sparse-coding structure of the resultant filters . Third , this technique can be considered an abstraction of more biologically plausible models . The retinal wave model of Butts and Feller [17] showed that wave propagation speed and termination were primarily determined by a 2-D map of one summary variable , f – the local fraction of recruitable amacrine cells – similar to our variable ‘p’ . Their random variation of this parameter came from variations in cell refractory period , temporal dynamics from multiple waves , and influence of non-propagating spontaneous activity . Although the more recent Godfrey and Swindale model [18] does not offer an equivalent summary variable , we believe a similar abstraction of local network excitability is equally possible . In the Butts model a neuron would only fire if a threshold of neighboring cells fired , similar to our ‘t’ , while in the Godfrey model this threshold varied over time . Both models also had a fixed dendritic field size , analogous to our ‘r’ . Their parameters were chosen to match known physiological parameters such as wave size , speed , and frequency given neurophysiological constraints . Our parameters choice , however , was more constrained by theoretical and computational concerns . It may be useful to compare these models; for example , pharmacological manipulations of amacrine cell recruitability or neural firing threshold could move pattern generation along our p-t phase plane vertically or horizontally respectively , leading to potentially testable predictions . We however consider this particular pattern generation technique better suited as a conceptual model to address a developmental paradigm and limitation of current statistical techniques , rather than a guide for directly verifiable experiments . As stated in the introduction , we believe that the use of highly theoretical models such as percolation networks and Ising models have been of sufficient use in understanding neural phenomena [52] , [53] to warrant application in this domain . This method provides an alternative approach to understanding the relation between spontaneous activity and V1 development by stressing the relation to image statistics and efficient coding in individual receptive fields . There are a number of models that stress other physiological dimensions , such as cortical column map formation , which can provide more insights to development . Linsker [42] demonstrated orientation column ( OR ) formation in a multi-layer model beginning with uncorrelated noise . Grabska-Barwinska and von der Malsburg demonstrate orientation column formation using recent experimental evidence of patchy , spatially periodic cortical spontaneous activity [60] . Miller [45] developed ocular dominance ( OD ) as well as orientation column ( OR ) formation . More current models have become even more ambitious in the development of map features . Bednar and Miikkulainen demonstrated direction selectivity ( DR ) to create a combined map ( OR/DR ) [61] . A later model combined these features ( OC/OR/DR ) using translated natural images [44] . Carreira-Perpinan et . al . [46] using the elastic net model [62] included a spatial-frequency map ( OC/OR/DR/SF ) , although the relation of their input to either natural stimulation or spontaneous activity is not clear . In each of these models the goal was to synthesize a cortical map and receptive fields which mimic known neurophysiology . Our use of functional , efficient coding methods precludes any relation to a particular topography , but with this we generate individual receptive fields with properties more relevant to physiological filters – more spatially bandpassed and localized . Our technique also directly addresses how the resulting code reflects its function during natural vision - by similarly efficiently encoding natural inputs using the same adaptive algorithm . Although our model clearly lacks a columnar organization , it does uniquely address the relation of spontaneous activity to current statistical methods of efficient coding . Although this paper stresses the effects and theoretical justifications of spontaneous activity , there are clearly limitations to this method for preparing V1 . Crowley and Katz [24] stated that ocular dominance columns initially form through molecular guidance mechanisms , and subsequent activity was needed for maintenance and plasticity during the critical period . Also , Ringach's connectivity model [63] , [64] shows how V1 receptive fields and functional topography could form based on the quasi-regularity of the ON/OFF center retinal ganglion cells in the retinal mosaic; with closely located ON and OFF-center cells forming simple receptive fields . Certainly a number of molecular-guidance mechanisms are necessary for proper development , and even if rudimentary receptive fields can form through simple axon guidance mechanisms , we still believe the simplicity and functional benefits of endogenous activity suggest a plausible role in development . The visual system will eventually refine based on the statistical structure in the experienced natural signals , and the pre-experience receptive fields can refine using the same mechanism on simple patterns . This conceptual model is able to address general properties of this process; however , it is more difficult to address the precise nature of the receptive fields between molecular guidance cues and the onset of natural experience . In addition to physiological details , such as optical and retinal maturity , the goals of this handoff between development processes need to be specified for a given animal . Some precocial animals may require a well functioning visual system from the onset , implying a goal of immediate efficient coding at the expense of later adaptability . On the other hand , altricial animals , such as monkeys and humans , may trade off immediate optimality for a certain amount of environmental adaptability; this may be one functional justification for the large spatial frequency and orientation bandwidths in neonatal monkeys . Although a more detailed , species-specific analysis may require additional assumptions , the general strategy may be universally beneficial . The functional benefits are an increased refinement beyond rough molecular cues using techniques which are relatively simple given the existence of a separate , adaptive learning system . In summary , our pattern generation technique resembles known patterns of spontaneous activity in both appearance and how they are generated . We have demonstrated that simply-generated , sparse , wavefront-containing patterns have the statistics to produce a sparse , efficient code with filters resembling those found in primary visual cortex and those produced by an efficient coding of natural scenes . Also , this work demonstrates the critical importance of statistics beyond simple pairwise correlations ( figure 4 ) which exist in wavefront-containing patterns . Efficient coding models relating natural scene statistics to V1 activity have relied on higher-order statistics for over a decade . Previous spontaneous activity models that try to explain V1 formation rely only on lower-order statistics that may not be as relevant to early visual processing from a functional perspective . The combination of a simplified abstraction of physiological methods of spontaneous activity and the demonstration that it provides a richer theoretical and computational understanding of why these patterns emerge is clearly attractive as it gives us a better , deeper understanding of the nature of spontaneous activity in development . With spontaneous activity present in sensory systems , the hippocampus , and motor systems [65] , any additional methods of understanding this activity may lead to insights of value in many other areas of brain development . We believe it is useful to add a computational perspective to the mechanistic interpretation of this activity - in addition to the role of spontaneous activity in axon branching , dendritic patterning , and synaptic pruning . Clearly these implementation-level goals are necessary for function , but do not address the general , functional purpose for this connectivity . A statistical , computational perspective is more likely to address the universal and ubiquitous nature of these patterns during development . In this paper , we have given a parsimonious explanation of both why this activity has particular sparse , edge-like statistics beyond simple correlations and how this allows the same adaptive learning system to use both endogenous spontaneous activity and natural inputs to refine the visual code . But more generally , we believe that by examining spontaneous activity in this way , we bring about a conceptual shift in the way people interpret developmental strategies . In the context of the visual system , it appears that the system both learns from patterns extrinsic to its functionality , but strictly internal to the animal; a bridging point between both learned and innate interpretations of mental phenomena .
The pattern generation is a variation of site percolation [51] . We use a simple , three parameter ( p , r , t ) model involving initiation and complete propagation of wave activity – thus the patterns have no temporal component . On a square array of points , mark a random fraction ‘p’ of the points on the grid as potentially active . To initiate an active cluster , we randomly select a location and activate all available points in a neighborhood radius ‘r’ . Neighboring potentially active points near the wave can only become active if there are at least ‘t’ active points within a distance ‘r’ . The wave is allowed to propagate until no more cells can become active . This completely explains the method of pattern generation , but not the interesting aspects of the behavior . Introductory percolation theory involves networks with t = 1 , often with r = 1 as a typical example . When ‘p’ approaches a value known as the percolation threshold , ‘pc’ , the pattern of activity is known to be fractal – the image statistics appear similar at all scales . For example , when p<pc , the activation terminates forming small clusters , when p>pc the activation spreads without bound leaving small holes without activity , but when p = pc , both the activity and holes are nearly infinite in extent – leading to a fractal interpretation . Images at increasingly larger scales become indistinguishable . For examples , the following ( p , r , t ) triplets are known empirically , and in some cases analytically , to produce fractal images – ( ∼0 . 592 , 1 , 1 ) , ( ∼0 . 407 , 1 . 8 , 1 ) , ( ∼0 . 288 , 2 , 1 ) [66] . For our measurements , for a given ‘r’ and ‘t’ pair , we found pc by finding the maximum derivative in the function of cluster size to ‘p’ value . To obtain enough edge statistics on these waves , we randomly chose points to begin wave propagation until more than 20% of the available points were activated – stopping when the last wave was allowed to fully propagate . Only the spatial statistics of these final patterns were explored . All encoding was done by downsampling the image by 2 to minimize any local edge effects due to aliasing . Unless otherwise noted , we set r = 3 for simplicity . The method used to analyze these patterns is demonstrated in figure 3 . 1 ) Generate a series of patterns from a given set of parameters , 2 ) extract image patches from that set , 3 ) preprocess ( “whiten” ) the data , and find the optimal code for the data using independent component analysis ( details below ) , 4 ) fit 2D gabor functions to the resulting filters , and 5 ) analyze properties of the resulting gabors . We show how they vary with a change in the underlying pattern generation parameters and compare to experimentally measured filters . In this case , we chose orientation bandwidth to demonstrate how filters that qualitatively appear similar can vary in a dimension useful to efficient encoding in the adult . For both natural images and activity patterns we randomly sampled image patches – excluding patches within a patch width of the border . We also excluded patches without a significant variation for learning – specifically , patches with a pixel variance less than 0 . 16 . This is equivalent to a requirement that between 20%–80% of the original downsampled sites were occupied . For each 256×256 image , downsampled to 128×128 , up to 100 viable 16×16 pixel patches were selected . This was done until 10 000 patches were collected , which were then encoded using the fastICA algorithm [48] using the ‘tanh’ contrast function . PCA dimensionality was reduced to 100 .
|
Before many animals first open their eyes , neurons in the retina , thalamus , and visual cortex fire spontaneously in highly structured , patterned ways . Experimental manipulations have demonstrated that this activity is necessary for proper function , but it is difficult to answer certain fundamental questions about the role of this activity by using experimental manipulations alone . We know that the early visual system can adapt to better encode statistical regularities in the environment . Can the same learning system that adapts to natural input be applied to this patterned activity to learn the visual code before birth ? What qualities would we want in an instructional pattern of activity in the developing visual system ? We answer these questions by presenting an abstract model of spontaneous activity in the early visual system—with direct relations to more physiological models . We demonstrate that instructive statistical properties can exist in spontaneously generated patterns based on very simple , local interactions . Also , we demonstrate that these patterns not only have the necessary pairwise correlations , which previous models have relied upon , but also additional sparse , edge-like structure . This higher-order statistical structure is universal to natural visual scenes and is necessary to understand neural responses as an efficient coding of our natural world . Most importantly , this additional structure would allow the visual system to use the same adaptive efficient coding strategy in two cases previously treated as separate—learning from natural visual experience as well as through innately generated patterns before visual experience .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"neuroscience/theoretical",
"neuroscience",
"neuroscience/sensory",
"systems"
] |
2008
|
Innate Visual Learning through Spontaneous Activity Patterns
|
Unexplained cardiac arrest ( UCA ) with documented ventricular fibrillation ( VF ) is a major cause of sudden cardiac death . Abnormal sympathetic innervations have been shown to be a trigger of ventricular fibrillation . Further , adequate expression of SEMA3A was reported to be critical for normal patterning of cardiac sympathetic innervation . We investigated the relevance of the semaphorin 3A ( SEMA3A ) gene located at chromosome 5 in the etiology of UCA . Eighty-three Japanese patients diagnosed with UCA and 2 , 958 healthy controls from two different geographic regions in Japan were enrolled . A nonsynonymous polymorphism ( I334V , rs138694505A>G ) in exon 10 of the SEMA3A gene identified through resequencing was significantly associated with UCA ( combined P = 0 . 0004 , OR 3 . 08 , 95%CI 1 . 67–5 . 7 ) . Overall , 15 . 7% of UCA patients carried the risk genotype G , whereas only 5 . 6% did in controls . In patients with SEMA3AI334V , VF predominantly occurred at rest during the night . They showed sinus bradycardia , and their RR intervals on the 12-lead electrocardiography tended to be longer than those in patients without SEMA3AI334V ( 1031±111 ms versus 932±182 ms , P = 0 . 039 ) . Immunofluorescence staining of cardiac biopsy specimens revealed that sympathetic nerves , which are absent in the subendocardial layer in normal hearts , extended to the subendocardial layer only in patients with SEMA3AI334V . Functional analyses revealed that the axon-repelling and axon-collapsing activities of mutant SEMA3AI334V genes were significantly weaker than those of wild-type SEMA3A genes . A high incidence of SEMA3AI334V in UCA patients and inappropriate innervation patterning in their hearts implicate involvement of the SEMA3A gene in the pathogenesis of UCA .
Unexpected sudden death in healthy individuals remains a daunting problem . Unexplained cardiac arrest with documented ventricular fibrillation ( UCA ) including idiopathic ventricular fibrillation ( IVF ) is defined as spontaneous VF that is not associated with a known structural or electrical heart disease . IVF is diagnosed in up to 10% of survivors of out-of-hospital cardiac arrest [1] . Many reports have documented the role of abnormal sympathetic innervations as a trigger of VF [2]–[6] . Sympathetic innervation of the heart is determined during development by chemoattractive and chemorepulsive factors . Semaphorins , members of a conserved family of both secreted and integral membrane proteins , are typical chemorepulsive factors acting on the growth cone as guidance cues to control the establishment of neural connections [7] , [8] . Recently , SEMA3A was shown to form an epicardial-to-endocardial transmural sympathetic innervation pattern in the heart . In addition , disruption of innervation patterning in both SEMA3A -deficient and SEMA3A-overexpressing mice resulted in sudden death or lethal arrhythmias [9] , [10] . Identification of the genes responsible for UCA may further increase our understanding of the pathophysiology of UCA and facilitate the diagnosis and prophylactic treatment , especially in asymptomatic , disease-carrying relatives of the proband . In the current study , we investigated the significance of the SEMA3A gene polymorphisms in the etiology of UCA .
The subjects were divided into two geographic regions based on their birthplace information , as shown in Figure S1 . The characteristics of the two regional groups of UCA patients enrolled in this study are listed in Table 1 . There was no significant difference in the clinical characteristics among the two UCA groups . As for control groups , the gender distribution was similar in the two groups , but individuals were older in Eastern Japan as compared to Western Japan ( 70±9 years vs . 47±16 years ) . Only one nonsynonymous polymorphism was identified in exon 10 of the SEMA3A gene through resequencing of the coding region . This polymorphism causes an amino acid substitution from isoleucine to valine ( I334V , SEMA3AI334V ) and is identical with the SNP that was recently submitted to dbSNP ( rs138694505 ) . There was a significant difference in genotype frequencies between UCA cases and controls in the western Japan ( dominant model P = 0 . 007 ) . This association was replicated in the Eastern Japan ( P = 0 . 008 ) . The Breslow-Day test showed no heterogeneity among the groups , and the overall degree of association by the Mantel-Haenszel test was P = 0 . 0004 ( OR 3 . 08 , 95%CI 1 . 67–5 . 70 ) ( Table 2 ) . Collectively , 13 of the 83 UCA patients ( 15 . 7% ) carried the risk genotype G , whereas only 5 . 6% did in the controls . The SEMA3AI334V carrier frequency appeared to be relatively stable throughout the age classes ( Data not shown ) . According to the 1000 Genomes Project , regional differences in the SEMA3AI334V ( rs138694505 ) frequency are evident among populations . For example , the frequency of the G allele is 2 . 1% in East Asians ( 3 . 93% in Japanese ) , 1 . 35% in West Africans , 1 . 86% in Americans , and 0% in Europeans ( Table 3 ) . Two UCA cases were severe ( Figure 1 , patients 1 and 2 ) . They suffered from VF at a young age and had a family history of sudden cardiac death . VF attacks recurred on several occasions in these patients . In one patient ( Patient 1 ) , VF recurred twice after discharge and was terminated by an implanted cardioverter defibrillator ( ICD ) shock ( Figure 2 upper panel ) . According to the ICD records , a preceding transient bradycardia was followed by short coupled ectopic ventricular beats , finally leading to VF . Another patient ( Patient 2 ) went into an electrical storm at midnight one day after hospitalization ( Figure 2 lower ) . VF occurred suddenly during sinus bradycardia . She had been suffering repeated epileptic seizures with loss of consciousness from the age of 15 . Most patients with SEMA3AI334V were found to have sinus bradycardia and sinus node dysfunction by an electrophysiological study . Figure 1 shows the ECGs before the ICD implantation in patients 1 , 2 and 3 . Because of the sinus bradycardia and in order to prevent a VF recurrence , the ICD was set to the AAI+ mode at 60–75 bpm . The number of tests that was performed in the UCA patients is shown in Table S2 . The phenotype characterization in each UCA patient with SEMA3AI334V is shown in Table S3 . Patient 3 had persistent AF and patient 13 had chronic AF . Patients 1 , 2 , 5 , 7 and 9 had 1st degree atrioventricular block . In Table 4 we present the clinical , electrocardiographic , and echocardiographic findings between the UCA patients with and without SEMA3AI334V . VF occurred predominantly at rest and during the night in the patients with SEMA3AI334V . In contrast , it occurred during exercise and during the day in most patients without SEMA3AI334V ( VF occurred during the night 69 . 2% vs . 37 . 1% , P = 0 . 032 , VF occurred at rest 69 . 2% vs . 34 . 3% , P = 0 . 015 ) . Some of the patients with SEMA3AI334V had sinus bradycardia , and their RR intervals on the 12-lead ECG tended to be longer than those without ( 1031±111 ms vs . 932±182 ms , P = 0 . 039 ) . None of the UCA subjects regularly took β-blockers during their ECG recordings . One patient without Sema3aI334V took 100 mg/day of oral amiodarone when recording the ECG . The other cases did not have any anti-arrhythmic agents . Early repolarization ( ER ) was evident in only two SEMA3AI334V cases ( 15 . 4% ) , whereas 34 patients ( 48 . 6% ) without SEMA3AI334V demonstrated ER ( P = 0 . 02 ) . The other 12-lead ECG parameters , signal-averaged ECG , and echocardiographic findings , were similar in the patients with and without SEMA3AI334V . To screen the entire SEMA3A gene , 47 tag SNPs were additionally genotyped in the UCA patients from Hiroshima/Nagasaki University and the healthy controls from Hiroshima University ( Table S1 ) . All SNPs were successfully genotyped in >98% of the samples . Among them , one SNP , rs1533996 , was not polymorphic . The other SNPs were within the Hardy-Weinberg equilibrium ( P>0 . 01 ) in the controls except for rs13437857 ( P = 0 . 0031 ) and rs10280701 ( P = 0 . 000086 ) . The p value of the I334V in the population ( p = 4 . 53E-08 ) was still significant even if a Bonferroni correction for the tag-SNP approach was applied ( p = 2 . 12E-06 ) . None of the 47 tag SNPs were significantly associated with UCA after the Bonferroni correction . The I334V variant showed a moderate linkage disequilibrium only with rs740948 ( r2 = 0 . 43 ) . A haplotype analysis revealed that no haplotype had a stronger association with UCA than the single marker analysis ( data not shown ) . Representative immunofluorescence images for vinculin ( a cell surface marker ) and anti-tyrosine hydroxylase ( TH ) in the sympathetic nerves in the subendocardial layer of patients with and without SEMA3AI334V are shown in Figure 3 . Under normal conditions , the TH nerves were reported to exist in the subepicardial layer of cardiomyocytes , not in the subendocardial layer ( 9 ) . In patients without SEMA3AI334V , no TH nerves were observed in the subendocardial layer , consistent with earlier findings in normal subjects . In patients with SEMA3AI334V , in contrast , TH nerves were distributed in the subendocardial layer ( right panel , the arrowheads indicate TH positive nerves ) . This finding was consistently observed in patients with SEMA3AI334V ( N = 4 ) but not without SEMA3AI334V ( N = 8 ) , suggesting abnormal sympathetic innervation in the heart of UCA patients with SEMA3AI334V . On the other hand , NGF , a neural attractant factor , was similarly expressed in the subendocardial layer in patients with and without SEMA3AI334V ( Figure 4 ) . As a result of a DRG repulsion assay , SEMA3AWT-expressing cells repelled the DRG axons on the proximal side of the ganglia ( Figure 5 , left ) . In contrast , DRG explants were less responsive to SEMA3AI334V ( Figure 5 , middle ) . Figure 6 shows the percentage of collapsed growth cones in the E8 chick embryos incubated with media containing SEMA3AWT , SEMA3AI334V and vector only ( negative control ) at 0 . 3 , 0 . 1 , and 0 . 03 dilutions of a concentrated media , respectively . At all dilutions , SEMA3AWT , and SEMA3AI334V were similarly expressed and secreted ( Figure 6 ) . The secreted proteins for both SEMA3AWT and SEMA3AI334V were similar in size ( approximately 65 kDa ) . The growth cone collapse by SEMA3AI334V was less frequent than that of SEMA3AWT at all concentrations . ( SEMA3AWT vs . SEMA3AI334V: 84 . 8±1 . 5% vs . 75 . 8±1 . 8% at a dilution of 0 . 3 , P = 0 . 009 , and 70 . 2±1 . 1% vs . 57 . 2±2 . 4% at a dilution of 0 . 1 , P = 0 . 009; Figure 6 , lower ) .
To the best of our knowledge , this is the first report demonstrating that UCA patients have a high incidence of I334V SNP ( rs138694505 ) in the SEMA3A located at chromosome 5 . Furthermore , new experimental data presented here indicates that SEMA3AI334V disrupts the SEMA3A function of inhibiting neural growth and impaired appropriate innervation patterning in the heart . Finally , this study suggested that SEMA3AI334V is a risk factor for human UCA and contributes to the pathogenesis of UCA . Many studies have reported the relationship between abnormal autonomic nerve activity and lethal ventricular arrhythmias , and in most of them I123-MIBG imaging was used to aid in the detection of sympathetic innervation abnormalities [3]–[5] , [11] . However , the molecular mechanisms determining these innervation densities in patients with lethal arrhythmia have not been fully clarified . Elucidation of underlying genetic defects will provide further insight into the pathogenesis of UCA , but identification of the genes involved in UCA is very difficult because of its high mortality rate and subsequent diagnostic difficulties . Unlike other monogenic arrhythmia syndromes ( e . g . , long QT syndrome , catecholaminergic polymorphic ventricular tachycardia and Brugada syndrome ) , the diagnosis of UCA cannot be made on the basis of ECG abnormalities prior to the occurrence of VF . In addition , UCA is only diagnosed by excluding any identifiable structural or functional cardiac diseases among the few survivors of VF . One case report indicated that a missense variant of the KCNJ8 gene , a subunit of the KATP channel , conferred a predisposition to dramatic depolarization changes and ventricular vulnerability [12] . In another report , Alders et al . demonstrated that a haplotype on chromosome 7 , which includes the DPP6 gene ( associated with potassium channel Ito subunits ) , was the causal gene of IVF [13] , [14] . Sympathetic innervation of the heart is sculpted during development by chemoattractive factors such as NGF and chemorepulsive factors such as SEMA3A . NGF acts through the Trk A and p75 neurotrophin receptors in sympathetic neurons . Lorenz et al . reported heterogeneous ventricular sympathetic innervation , altered β-adrenergic receptor expression , and rhythm instability in mice lacking the p75 neurotrophin receptor within the heart [15] . Ieda et al . [9] , [10] reported that cardiac innervation patterning is disrupted in SEMA3A-deficient and SEMA3A-overexpressing mice , leading to lethal arrhythmias and sudden death . On the basis of this background information , we focused on SEMA3A , which plays a crucial role in cardiac innervation patterning [7]–[10] , [16] , as abnormal sympathetic innervations have been demonstrated in patients with UCA . We observed that a polymorphism in exon 10 of the SEMA3A gene ( i . e . , SEMA3AI334V ) , located in the semaphorin domain , which plays an essential role in SEMA3A [17] , was highly prevalent in patients with UCA and strongly associated with UCA pathophysiology . To our knowledge , this is the first report that investigates the relevance of functional mutations or polymorphisms in SEMA3A with respect to human diseases . We divided the case and control subjects into two geographical groups based on their birthplace in Japan . Significant results observed in Western Japan were replicated in the Eastern Japan group , and the combined P value and odds ratio calculated by the Mantel-Haenszel test were 0 . 0004 and 3 . 08 , respectively . According to publicly available data from the 1000 Genomes Project , the frequency of this risk allele of SEMA3A is similar among populations other than Europeans , suggesting that this variant may be relevant to the etiology of UCA across these populations . In our study , the G allele frequency was 2 . 8% in the controls , which was consistent with that reported in Japanese ( 3 . 9% ) and East Asian populations ( 2 . 1% ) in the 1000 Genomes Project . Haïssaguerre et al . reported an increased prevalence of ER characterized by J-point elevation among patients with a history of UCA [18] . Antzelevitch et al . classified ER patterns for risk stratification of VF [19] . The genetic basis for ER is slowly coming into better focus . Burashnikow et al . identified loss of function mutations in the α1 , β2 , and α2δ subunits of the cardiac L-type calcium channels ( CACNA1C , CACNB2 , and CACNA2D1 ) in patients with ER syndrome [20] . Abe et al . reported that ER may be closely associated with depolarization abnormalities and autonomic modulation [21] . In this study , only two UCA cases with SEMA3AI334V demonstrated ER . Instead , the characteristics of the cases with SEMA3AI334V suffered VF attacks in a relaxed state and presented with sinus bradycardia/sinus node dysfunction . These findings are consistent with the report by Ieda et al . [9] , [10] that SEMA3A−/− mice lacked a cardiac sympathetic innervation gradient and exhibited satellite ganglia malformations , which led to marked sinus bradycardia due to sympathetic dysfunction . Some of the UCA cases in our study may have a mild degree of depolarization or repolarization abnormalities , although we could not detect any obvious organic diseases such as cardiomiopathy by diagnostic imaging or manifest conduction disturbances . The other patients did not have any depolarization or repolarization abnormalities . The patients with SEMA3AI334V do not have a homogeneous phenotype and we have to follow up the clinical course of the UCA patients with SEMA3AI334V for a long period . The frequency of AF was 21 . 6% and rather high in the UCA subjects of our study for unknown reasons and the frequency was similar in the patients with and without SEMA3AI334V . One possible reason was that the episodes of AF after resuscitation were included in the past history of AF . In our study , immunofluorescence staining of the RV revealed that sympathetic nerves were distributed in the subendocardial layer only in patients with SEMA3AI334V . If SEMA3A exists in adequate quantities in the endocardial layer and functions normally , sympathetic nerves extending to the endocardial layer are suppressed . We assumed that in UCA patients with SEMA3AI334V , the epicardial-to-endocardial transmural sympathetic innervation patterning had deteriorated . An SEMA3AWT- and SEMA3AI334V-concentrated media did not grossly affect the expression , stability , or secretion of the ligand . As for the molecular weight of SEMA3A , when it was expressed in HEK293 , the full semaphorin domain ( 65 kDa ) was cleaved and detected in a conditioned media [22] . The sizes of the secreted proteins in both SEMA3AWT and SEMA3AI334V were equal and coincident with the semaphorin domain including a dimerization interface and Neurolipin-1 ( Nrp-1 ) -binding residue , and the biological activity was sufficient for the acquisition of a high repulsive activity [22] . The function of repelling the DRG axons was weaker and growth cone collapse was less frequent in SEMA3AI334V than in SEMA3AWT . Therefore , one allele of SEMA3A leads to a disruption of the sympathetic innervation of the heart under relevant conditions . These findings were consistent with immunofluorescence observations strongly suggesting that SEMA3AI334V can disrupt the ability of SEMA3A to repel or collapse DRG axons and sensory neuron growth cones under equal conditions of the neural attractant NGF . Merte et al . reported that a forward genetic screen in mice identified a novel loss of function SEMA3AK108N mutation , which bound to Nrp-1 but failed to repel or collapse DRG axons in vitro [23] . SEMA3AI334V exists in blade 5 of the 7-bladed propeller structure of the semaphorin domain and performs a crucial function in SEMA3A . Residues 333–335 in 5S of SEMA3A constitute the dimerization interface . The SEMA3A-65K dimerization interface overlaps with sites responsible for the initial high-affinity binding to the domain of Nrp-1 . Binding of SEMA3A to Nrp-1 leads to a conformational change in Plexin-A1 , which is transmitted to the cytosolic domain [17] . In the association analysis , SEMA3AI334V was highly prevalent in patients with UCA and associated with the UCA pathophysiology . On the other hand , none of the control subjects with SEMA3AI334V had any signs of disease at the time of the study , indicating incomplete penetrance or additional environmental or genetic factors . Our study had several limitations . First , it was very difficult to congregate many UCA cases and therefore the size of our study population was too small to obtain any robust findings . Secondly , we were not able to study the segregation data in the UCA patients with SEMA3AI334 because their families refused screening . A future prospective study with a larger cohort will be required to obtain these data . A further functional study would also be desirable to determine whether any abnormal innervation can be observed in healthy carriers by using autopsy specimens In conclusion , a polymorphism of SEMA3AI334V diminishes the cardiac sympathetic innervation gradient and partially contributes to the etiology of UCA . This finding is important in elucidating the pathogenesis of UCA .
We recruited a total of 83 UCA patients ( 64 male and 19 female , mean age 43±16 years ) from Hiroshima University Hospital , Nagasaki University Hospital , Shiga University of Medical Science , and the National Cerebral and Cardiovascular Center . We recruited 2958 controls ( 1540 male and 1452 female , mean age 54±18 years ) from Hiroshima University Hospital , Osaka-Midosuji Rotary Club ( Osaka , Japan ) , Shiga University of Medical Science , and Niigata University Graduate School of Medical and Dental Sciences . All patients and controls in this paper were unrelated Japanese individuals . Case and control subjects were collected from various regions of Japan . Although the Japanese population has rather low genetic diversity , it has been shown that population structures may lead to spurious associations [24] . Therefore , to eliminate the possibility of a population stratification , we divided case and control subjects into two groups geographically based on their birthplace information ( i . e . , Western Japan and Eastern Japan ) ( Figure S1 ) . The Institutional Ethics Committee of the Graduate School of Biomedical Science at Hiroshima University approved all procedures involving human tissue usage . Written informed consent was obtained from all subjects prior to participation . Twelve subjects enrolled in the study were diagnosed and treated at the Hiroshima University Hospital; the other subjects were diagnosed and treated at other affiliated hospitals and their information was provided to us . We defined UCA as that without structural heart disease and in the absence of signs of an arrhythmia syndrome such as Brugada syndrome , catecholaminergic polymorphic ventricular tachycardia and long QT syndrome . All patients with cardiac arrest underwent a physical examination , 12 lead ECG [25] , echocardiography and coronary angiography to rule out any underlying heart disease . Those who met the inclusion criteria were enrolled and underwent additional testing ( signal averaged ECG , T wave alternance , cardiac magnetic resonance imaging , computer tomography , provocation tests , cardiac biopsy or an electrophysiological study ) , if possible . The numbers of further noninvasive or invasive tests against UCA patients varied from institute to institute . Patients with exonic mutations in SCN5A and a positive pilsicainide challenge test were excluded from the sample . Early repolarization ( ER ) was defined as a QRS slurring or notching of ≥0 . 1 mV in more than two consecutive leads of the 12-lead ECG . Peripheral blood was obtained from all the subjects . Genomic DNA was extracted from leukocytes using a QIAamp DNA Blood Mini Kit ( QIAGEN , Hilden , Germany ) according to the standard protocol . Using Go Taq ( Promega , Madison , WI , USA ) , all coding regions of the SEMA3A located at chromosome 5 were amplified by PCR from 2 . 5-ng genomic DNA using our original primers in 17 UCA patients and 15 healthy controls entered from Hiroshima University . These amplified coding regions were then resequenced using an ABI PRISM 310 Genetic Analyzer ( Applied Biosystems , Foster City , CA , USA ) to identify mutations and polymorphisms . Subsequently , SNP genotypes were genotyped in All of the UCA subjects and healthy control subjects using the Invader assay or the TaqMan assay , as described previously [26] , [27] . The 47 tag SNPs were genotyped only in the UCA patients and the healthy controls entered from Hiroshima University and Nagasaki University . Using the HapMap database ( public release #27 , hapmap . ncbi . nlm . nih . gov ) and the Haploview program ( www . broad . mit . edu/mpg/haploview ) and based on selection criteria of r2>0 . 8 and a minor allele frequency of >0 . 01 for the Japanese population , tagging-SNPs were selected from the SEMA3A region spanning approximately 247 kb , from approximately 5 kb upstream of the transcription start site to 5 kb downstream of the 3′ untranslated region . The complete coding region of human SEMA3A was amplified from cDNA with forward ( tgttagtgttgccatgaggtct ) and reverse ( gcattcacctgtgttctctgttag ) primers . To generate Flag- SEMA3A , the coding sequence DYKDDDD was introduced between the codons for G25 and K26 ( NM_006080 . 2 ) . The I334V mutation was introduced by site-directed mutagenesis using the QuickChange ( Stratagene , La Jolla , CA , USA ) . Full-length human wild-type ( SEMA3AWT ) or mutant SEMA3A ( SEMA3AI334V ) cDNA was cloned into pcDNA3 . 1 ( + ) ( Invitrogen , Carlsbad , CA , USA ) . Transverse sections of a septal site of the RV outflow tract were obtained by biopsy from 12 UCA subjects ( 4 patients with SEMA3AI334V and 8 patients without SEMA3AI334V ) . These sections were embedded in an OCT compound ( Sakura , Torrance , CA , USA ) and frozen with liquid nitrogen . Immunofluorescence staining was performed using the frozen sections with rabbit anti-TH ( AB152 , Millipore , Billerica , MA , USA ) antibodies and mouse anti-vinculin ( Sigma-Aldrich , St . Louis , MO , USA ) antibodies diluted at concentrations of 1∶100 and 1∶200 , respectively , in 1% BSA/PBS . Alexa 488-conjugated goat anti-rabbit and Alexa 568-conjugated goat anti-mouse antibodies ( Invitrogen ) were used as secondary antibodies . As for NGF , sheep polyclonal to NGF ( ab49205 , Abcam , Cambridge , MA , USA ) and rabbit anti-TH ( ab152 , Millipore ) were used as primary antibodies at concentrations of 1∶100 in 1% BSA/PBS . Alexa568 donkey anti-sheep ( A21099 ) and Alexa488 donkey anti-rabbit ( A21206 ) antibodies were used as secondary antibodies . Nuclei were stained with 10 µM of Hoechst 33342 ( Molecular probes ) . Samples were examined using a confocal microscope and captured with a 20× and 40× objective lens on a Zeiss LSM 510 laser scanning microscopy system ( Carl Zeiss , Thornwood , NY , USA ) . The DRG were dissected from E8 chick embryos . HEK293T cells were transfected with Flag- SEMA3AWT or SEMA3AI334V expression vector or equal amounts of empty vector ( control ) using Gene Juice Transfection Reagent ( Novagen , Madison , WI , USA ) . The DRG and SEMA3A -expressing HEK293T cell aggregates were embedded as described previously [28] . Samples were incubated at 37°C in a 5% CO2 humidified incubator for 48 h and examined using an inverted microscope . For DRG repulsion assays , 10–15 DRG cells were examined , each with Sema3AWT , Sema3AI334V , or a control . For the purpose of a growth cone collapse assay , the conditioned medium of the SEMA3A -expressing HEK293T cells was concentrated [22] . A Western blot analysis was performed using both dilutions of the SEMA3AWT and SEMA3AI334V concentrated media with anti-FlagM2 ( Sigma ) . Growth cone collapse assays were performed as previously described using chick E8 DRG explants grown on laminin ( Invitrogen ) - and poly-L-lysine ( Sigma ) -coated 48-well plates ( BD Falcon/353078 ) . The dilution series of the SEMA3AWT , SEMA3AI334V and vector only concentrates were added to each well and incubated at 37°C in a 5% CO2 humidified incubator for 30 min . The explants were fixed with 4% paraformaldehyde in 10% sucrose PBS ( pH 7 . 4 ) , and the samples were examined using an inverted microscope [29] . In each dilution series , 5 or 6 growth cone collapse assays were investigated . Each in vitro assay was performed in triplicate . For quantification , we counted at least 50 growth cones to score on each explant . We assigned each growth cone as either collapsed or not collapsed , and the results were expressed as the percentage of collapsed to all counted growth cones . We compared the percentage of those collapsed between the SEMA3AWT and SEMA3AI334V . Normally distributed continuous variables are presented as the mean ± SD . Continuous data between the two groups were analyzed using the nonparametric Mann–Whitney U test . For testing the genetic associations in the case–control studies , the chi-square test and Cochran–Armitage trend test were used . Tests for the Hardy–Weinberg equilibrium among the cases and controls were conducted for observed and expected genotype frequencies using an ordinary chi-square test , where a P-value of <0 . 05 was considered statistically significant . For a meta-analysis of 3 individual cases and controls , we used the Mantel-Haenszel test .
|
Unexplained cardiac arrest with documented ventricular fibrillation ( UCA ) is defined as spontaneous ventricular fibrillation ( VF ) that is not associated with known structural or electrical heart diseases and is one of the major causes of sudden cardiac death . Identification of the genes responsible for UCA may further increase our understanding of mechanisms of UCA and facilitate more accurate diagnosis and preventive treatment , especially in asymptomatic disease-carrying relatives of the patient . However , molecular mechanisms of UCA have not been fully clarified due to the high mortality rate and difficulty of diagnosis . In this study , UCA patients are shown to have a high incidence of a polymorphism in the Semaphorin 3A gene ( rs138694505 , SEMA3AI334V ) . The result confirms previous reports that the abnormal sympathetic innervation is a trigger of UCA because SEMA3A is crucial for the establishment of normal innervation patterns in the heart . Furthermore , experimental data presented here indicate that SEMA3AI334V disrupts the SEMA3A function and impairs appropriate innervation patterning . Finally , the study suggests that SEMA3AI334V is a risk factor for human UCA and contributes to the etiology of UCA .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"arrhythmias",
"neurology",
"genetics",
"biology",
"human",
"genetics",
"genetics",
"of",
"disease",
"autonomic",
"nervous",
"system",
"cardiovascular"
] |
2013
|
A Nonsynonymous Polymorphism in Semaphorin 3A as a Risk Factor for Human Unexplained Cardiac Arrest with Documented Ventricular Fibrillation
|
Leptospirosis is a worldwide zoonotic bacterial disease caused by infection with leptospires . Leptospirosis in humans and livestock is an endemic and epidemic disease in Thailand . Livestock may act as reservoirs for leptospires and source for human infection . Data on leptospirosis infection in humans and livestock ( Buffaloes , Cattle , and Pigs ) species during 2010 to 2015 were analyzed . Serum samples were examined using Microscopic Agglutination Test ( MAT ) to identify antibodies against Leptospira serovars using a cut-off titer ≥ 1:100 . The seroprevalence was 23 . 7% in humans , 24 . 8% in buffaloes , 28 . 1% in cattle , and 11 . 3% in pigs . Region specific prevalence among humans and livestock was found in a wide range . The most predominant serovars were Shermani , followed by Bratislava , Panama , and Sejroe in human , Shermani , Ranarum , and Tarassovi in buffaloes , and Shermani and Ranarum in cattle and pigs . Equally highest MAT titers against multiple serovars per one sample were found mainly in buffaloes and cattle showing equally titers against Ranarum and Shermani . The correlations of distribution of serovars across Thailand’s regions were found to be similar in pattern for cattle but not for buffaloes . In humans , the serovar distribution in the south differed from other regions . By logistic regression , the results indicated that livestock is more susceptible to infection by serovar Shermani when compared to humans . This study gives a detailed picture of the predominance of Leptospira serovars in relation to region , humans and typical livestock . The broad spatial distribution of seroprevalence was analyzed across and within species as well as regions in Thailand . Our finding may guide public health policy makers to implement appropriate control measures and help to reduce the impact of leptospirosis in Thailand .
Leptospirosis is a worldwide zoonotic bacterial disease particularly in tropical and subtropical countries [1] . Pathogenic Leptospira species are causative agents of the disease , specifically Leptospira interrogans sensu lato . There are approximately more than 250 recognized pathogenic serovars and 24 antigenically related serogroups [2 , 3] . Serovars , that are antigenically related , can be grouped into serogroups . The serogroups of L . interrogans can have some common serovars [4] . The infection in humans is caused by direct contact with products of infected animals , mainly urine , and also by indirect contact as the organisms can be transmitted to humans through cut skin or mucous membranes via a contaminated environment [5] . The continuing epidemic of human leptospirosis in Thailand produces an annual incidence rate of about 5 . 9 cases per 100 , 000 population each year during the last ten years [6] . From March 2003 to November 2004 , most confirmed cases occurred in the north and northeast regions of the country [7] . The number of reported cases was highest during the rainy seasons . Farmers and other agricultural workers make up the main occupational risk groups , which are likely to be exposed with contaminated wet soil and water during their daily activities [8–10] , for example during rice cultivation , fish capture and animal farming . In the environment , leptospires may survive from a few weeks to almost a year in wet soil on dry days or in surface waters on rainy days [11] . The animal hosts ( e . g . , cattle , buffaloes , pigs , dogs , and wildlife ) are considered as common reservoirs of leptospires and may act as a source for human infection [1] . Domesticated livestock in Thailand usually dwell in close proximity to their owners in rural areas , which poses a certain risk of interspecies transmission . In a previous study conducted in Thailand during January to August 2001 , it was found that the most commonly detected Leptospira serovars in cattle were Leptospira serovars Ranarum , Sejroe , and Mini , whereas Mini , Sejroe , and Bratislava were mostly detected in buffaloes , Ranarum , Pomona , and Bratislava in pigs , and Mini , Shermani , and Ranarum in sheep and goats , respectively [12] . The sera of patients in Bangkok Thailand were found to have the highest reactions to Leptospira serovars Shermani and Bratislava [13] . Undoubtedly , animals hosting and shedding leptospires pose a certain risk to public health as even vaccinated livestock are reported to shed Leptospira into urine [14] . The host reservoirs may be infected asymptomatically while infected humans with exactly the same serovars may develop serious illnesses [15 , 16] . Preventive measures by public health sections were typically focused on the increase of public awareness on the exposure risks when dealing with animals [17] . However , understanding of the diversity of Leptospira serovars in animals and humans should also be taken into account in order to identify associations of animal reservoirs with human infection . We established a large-scale dataset gathered from passive surveillance over a six year period , which may represent the largest and longest study of leptospirosis among buffaloes , cattle , pigs and humans in Thailand so far . This basic knowledge of serovars and their maintenance hosts is critical to provide more understanding of the epidemiology of leptospirosis . The aims of this study were i ) –to determine predominant Leptospira serovars circulating between humans and livestock in the epidemiological context of Thailand , ii ) –to identify spatial distributions of Leptospira serovar seroprevalence in humans and livestock in each of the 5 regions Thailand , iii ) –to assess the similarity between the distribution of predominant serovars across and within both species and regions in Thailand using cross-correlation analysis , and iv ) to provide a detailed investigation of Leptospira serovar seropositivity according to species and/or regions based on statistical analysis . This study also gives an update on leptospirosis in Thailand and information about host and human serovar association to support public health management .
A total of 7 , 218 livestock serum samples derived from 432 buffaloes , 3 , 648 cattle , and 3 , 138 pigs were submitted to National Institute of Animal Health , the Department of Livestock Development , Ministry of Agriculture , Thailand from January 2010 to December 2015 under the passive surveillance program of leptospiral seropositivity , which is a part of the passive surveillance program . Most of samples in this study were collected from rural areas . This study was not basically designed for research proposal; an animal ethic protocol was not required . The samples were sent by different reasons , e . g . , routine diagnosis , health check , and leptospirosis investigation . Most animals were not vaccinated against leptospirosis ( 6 , 934 , of 7 , 218 samples , 96 . 07% ) and were collected from rural areas . The 1 , 990 human serum samples under suspicion of leptospirosis , which have some clinical symptoms of leptospirosis such as high fever , headache , muscle aches , Jaundice , and diarrhea , were sent to National Institute of Health , Department of Medical Sciences , Ministry of Public Health , Thailand during January 2010 to December 2015 . Data collection was performed as a part of routine clinical examination procedures for which results were previously transmitted to patient and consent was thus not required by the Ethics Committee . Data contained in the patient’s records , without any patient information , except location and time , was de-identified prior to an anonymous analysis . All serum samples were examined for the presence of Leptospira serovar antibodies by the microscopic agglutination test ( MAT ) . The MAT is the most widely used method in identifying leptospiral positive samples [3] . Serological tests and leptospira culture protocol in this study were based on the standard methodology [18 , 19] using a panel of 23 reference serovars [12] . The panel of antigens included L . interrogans serovars Bratislava ( BRA ) ( serogroup Australis ) , Autumnalis ( AUT ) , Ballum ( BAL ) , Bataviae ( BAT ) , Canicola ( CAN ) , Celledoni ( CEL ) , Cynopteri ( CYN ) , Djasiman ( DJA ) , Grippotyphosa ( GRI ) , Hebdomadis ( HEB ) , Icterohaemorrhagiae ( ICT ) , Javanica ( JAV ) , Louisiana ( LOU ) , Manhao ( MAN ) , Mini ( MIN ) , Panama ( PAN ) , Pomona ( POM ) , Pyrogenes ( PYR ) , Ranarum ( RAN ) , Sarmin ( SAR ) , Sejroe ( SEJ ) , Shermani ( SHE ) , and Tarassovi ( TAR ) . Based on practical approaches , the MAT titer ≥ 1:100 is recommended cut-off and was used to determine seropositivity [15 , 20] . However , previously , the MAT showed high sensitivity and specificity of 95% and 89% at a cut-off titer 1:50 , respectively [21] . A dilution of 1:100 sera was screened as positive sera [22] . This cut-off titer was also used in previous studies [23 , 24] . This study , the occurrence of humans and livestock leptospirosis was determined by serological test . The cut-off titer was chosen to ≥ 1:100 to increase the previously specificity . Cross-reactivity between different serogroups may occur in MAT due to the detection of both IgM and IgG antibodies [5] . In this case , the highest MAT titer criteria were used to identify the predominant serovar ( s ) . Infection by multiple serovars was assumed in case of equally highest MAT titers against two or more serogroups . The distribution of leptospiral serovar seroprevalence was measured within and across both species and regions in Thailand . The seroprevalence was calculated from the proportion of leptospiral culture positive hosts and all tested samples . The spearman correlation test was employed to test the correlation coefficient between pairwise comparisons of seropositive frequency distribution within and across both species and regions . According to Thai Meteorological Department , Thailand can be divided into 5 regions , i . e . , northern , northeastern , central , eastern , and southern regions by using climate pattern and meteorology [25] . In northern region , most area is hilly and mountainous , which have lowest average temperature . Northeastern region is naturally a high level plain with slightly higher average temperature than northern region , while central region is a large lower plain with high average temperature . Eastern region is the part of adjacent of the Gulf of Thailand , which have high average temperature as well as in central region . The topography of southern region is the peninsula in Andaman Sea , which has the highest average annual rainy day comparing to the other 4 regions . The season of southern region is divided into two major seasons which are the rainy ( June-February ) and the summer ( March-May ) seasons , while the other 4 regions have 3 seasons , i . e . , a rainy season ( mid-May to mid-October ) , a winter season ( mid-October to mid-February ) , and a summer season ( mid-February to mid-May ) . All calculations were performed using R software [26] . The Bonferroni correction was used to adjust the P-value and control type I error rates in multiple comparisons . All pairwise comparisons with adjusted a P-value < 0 . 05 were considered significant . To investigate the relation of serovars according to species and/or regions , a logistic regression model ( Generalized Linear Model ( GLM ) with binomial function ) [27] was performed using R software [26] . A total of 9 , 208 serum samples were analyzed . The presence or absence ( Yes/No ) of antibodies against serovars was analyzed . Five serovars were selected based on the highest seroprevalence , i . e . , Bratislava , Ranarum , Sejroe , Shermani , and Tarassovi ( S1 Table ) . The best univariable or multivariable model was selected using a stepwise forward approach based on the Akaike Information Criterion ( AIC ) . The models were compared in regard to deviance and degree of freedom at a significance level of P-value < 0 . 05 under chi-squared distribution . The interaction between all effects was also investigated . The regression diagnostic model was performed by Cook's distance and leverage methods .
The Leptospira seroprevalence in 4 species ( buffaloes , cattle , pigs , and humans ) and 5 regions ( northern , northeastern , central , eastern , and southern ) as determined by MAT using a 23 serovar panel during the period 2010–2015 is summarized in Table 1 . A total of 7 , 218 animals were tested with 1 , 489 ( 20 . 6% , 95% CI [19 . 7–21 . 6%] ) animals found positive by MAT . The seroprevalence was 24 . 8% [20 . 8–29 . 1%] in buffaloes , 28 . 1% [26 . 7–29 . 6%] in cattle , and 11 . 3% [10 . 3–12 . 5%] in pigs . For humans , a total of 1 , 990 samples were tested and 471 ( 23 . 7% [21 . 8–25 . 6%] ) were found seropositive . Seropositive samples reacting with two or more serovars was highest in cattle measuring 43 . 3% [40 . 2–46 . 4%] and lowest in pigs counting 16 . 9% [13 . 1–21 . 1%] . In humans , seropositivity against multiple serovars occurred in 15 . 3% [12 . 1–18 . 9%] of tested samples . To assess the spatial seroprevalence , Thailand was divided into 5 regions: northern , northeastern , central , eastern , and southern according to climate variations [25] . Animal locations were based on the owner’s address . The highest overall prevalence ( 35 . 4% [31 . 6–39 . 4%] ) ( pooled data of 4 species ) was found in the southern region and the lowest overall prevalence was identified in the north ( 13 . 1% [11 . 0–15 . 5%] ) . The highest prevalence reacting with multiple serovars by region was found in the eastern region ( 47 . 3% [40 . 2–54 . 5%] ) , while the lowest was found in the central region ( 28 . 9% [26 . 0–32 . 0%] ) . Region specific prevalence in buffaloes ranged from 16 . 4% [11 . 7–22 . 2%] in the northeastern region to 41 . 2% [27 . 6–55 . 8%] in the central region ( one positive sample ( 3 . 33% ) in the south ) ( Fig 1 ) . In cattle , findings ranged from 6 . 1% [4 . 0–8 . 8%] in the northern region to 52 . 0% [44 . 7–60 . 3%] in southern region . The spatial seroprevalence in pigs ranged from 1 . 9% [0 . 7–4 . 1%] in the eastern region to 23 . 2% [31 . 1–50 . 2%] in the southern region with no positive sample in the north . In humans , the value ranged from 12 . 8% [10 . 5–15 . 5%] in the central region to 39 . 0% [32 . 7–45 . 7%] in the southern region . The predominance of serovars in humans and livestock was presented in Fig 1 . High titer seropositivity against single serovars was observed for 7 serovars and against multiple serovars for 6 serovars . Less prevalent serovars were grouped into “Other” . In buffaloes and in the northern and central regions , multiple serovar association was identified showing equally titers against serovars Ranarum and Shermani ( 57 . 1% and 28 . 6% , respectively ) , whereas in the northeastern and eastern regions the predominant serovar was serovar Tarassovi ( 38 . 2% and 33 . 3% , respectively ) . In cattle , the most common MAT reaction was Ranarum/Shermani for samples from all regions ( 38 . 6% in northern , 34 . 5% in northeastern , 29 . 6% in central , 42 . 6% in eastern and 42 . 9% in southern regions respectively ) . The most common serovar in all regions ( except northern , where no positive sample was observed ) in pigs was serovar Shermani ( 63 . 2% in the northeast , 68 . 6% in central , 83 . 3% in the east , and 38 . 6% in the southern regions , respectively ) , which coincides with the human samples ( 53 . 6% in northern , 60 . 3% in northeastern , 51 . 6% in central , 47 . 4% in eastern , and 70 . 8% in southern regions , respectively ) . By considering each serovar separately , regardless of species and regions , seropositive MAT titers of the first 5 predominant serovars were Shermani ( 67 . 9% ) , Ranarum ( 38 . 9% ) , Sejroe ( 6 . 4% ) , Bratislava ( 5 . 9% ) , and Tarassovi ( 4 . 5% ) , while titers against Canicola and Celledoni were not observed ( S1 Table ) . Seropositivity against three serovars ( Ballum , Cynopteri , and Panama ) was observed only in humans , whereas seropositive samples against serovars Manhao and Pyrogenes were observed only in livestock . In general , the MAT titer was higher in humans than in livestock . The most common serovars in all species was serovar Shermani , while serovar Ranarum was found mainly in livestock . We extracted the information of the predominant serovar that is specific to species and region from S1 Table and illustrated the results in Fig 2 . In buffaloes , the MAT titers against the primary serovar were different in each region; high titers were found against serovars Shermani and Ranarum in northern and central regions , against Tarassovi , Shermani , and Sejroe in the northeast , against Ranarum , Tarassovi , and Shermani in the east , and against Louisiana in the south ( with only one positive sample ) . In cattle , the most common serovar titers were against Ranarum and Shermani from all regions , whereas the primary infecting serovar in pigs and humans was serovar Shermani . The distribution of the Leptospira serovar seroprevalence in buffaloes differed in composition across regions whereas similar prevalence was found in cattle for all regions ( Fig 2 ) . In pigs , the serovar distribution in the central region differed to other regions showing highest serovar diversity ( present 11 serovars ) . It shall be noted that the central region had the highest sample size and seroprevalence in pigs . In the southern region , the serovar distribution in humans differed in comparison to other regions by higher proportions of serovars Shermani and Tarassovi . The pairwise correlation in distribution of serovar prevalence ( data in Fig 2 ) across and within both species and regions was determined using the Spearman method with Bonferroni adjustment of P-value ( Fig 3 ) . Strongly positive correlations across regions within species were found in buffaloes between northeastern and eastern regions ( Spearman’s correlation ( cor . ) 0 . 75 , 95% confidence interval [0 . 40–1 . 00] ) , in cattle between all regions ranging from cor . 0 . 70 [0 . 28–0 . 90] to cor . 0 . 82 [0 . 52–0 . 97] ( except between northern and central , and between northern and southern ) , in pigs between northeastern and eastern ( cor . 0 . 99 [0 . 99–1 . 00] ) , northeastern and southern ( cor . 0 . 74 [0 . 49–1 . 00] ) , and eastern and southern ( cor . 0 . 74 [05 . 0–1 . 00] ) , and in humans between northeastern and central ( cor . 0 . 70 , CI 0 . 31–0 . 90 ) , and northeastern and eastern ( cor . 0 . 70 [0 . 29–0 . 92] ) ( all P-values <0 . 05 ) . Strongly positive correlations across both regions and species were observed only between buffaloes and cattle and measured 0 . 71 [0 . 38–0 . 99] between both species in the northern region , 0 . 75 [0 . 42–0 . 94] in the central region for buffaloes and in the southern region for cattle and 0 . 80 [0 . 51–0 . 95] for buffaloes in the eastern region and for cattle in southern region . For correlation between livestock and humans , there was a medium positive correlation between buffaloes in the northeastern region and humans in the southern ( cor . 0 . 65 [0 . 33–0 . 86] ) , and buffaloes in the central region and humans in the northern ( cor . 0 . 64 [0 . 32–0 . 85] ) ( all P-values <0 . 05 ) . A negative correlation was observed between humans and livestock , however being statistically significant . We fitted data from all serum samples using the logistic regression model to investigate seropositivity with regard to species and regions . The selection of the best model used a forward stepwise approach based on the Akaike Information Criterion . The best logistic regression model was the univariate model for the serovars Bratislava , Ranarum , and Sejroe ( Table 2 ) and the multivariate model for the serovars Tarassovi and Shermani , and MAT appearance ( Table 3 ) . The effect of the region was analyzed in reference to the central region and the effect of the species in reference to humans . The central region was chosen as it is the region for which has the highest number of sample . Thus it is the most powerful disposition to study contrast between regions in reference to the central . For species effect , we would like to highlight the species with higher seroprevalence compared to human . The seroprevalence of serovar Bratislava alone in buffaloes ( odds ratio ( OR ) 0 . 1 ) , cattle ( OR 0 . 05 ) , and pigs ( OR 0 . 17 ) was significantly ( P-value <0 . 001 ) lower than in humans . When investigating the seroprevalence according to regions , a higher chance of seropositivity was associated with the northern ( OR 2 . 22 , P-value <0 . 05 ) , northeastern ( OR 3 . 26 , P-value <0 . 001 ) and southern regions ( OR 2 . 86 , P-value <0 . 01 ) when compared to the central region . There was no satisfying univariable model for seropositivity against serovars Ranarum and Sejroe associated with species . Seropositivity against serovar Ranarum was more frequent in the northeastern region ( OR 1 . 22 , P-value <0 . 05 ) , eastern ( OR 2 . 24 , P-value <0 . 001 ) , and southern regions ( OR 1 . 63 , P-value <0 . 001 ) when compared to the northern region ( OR 0 . 5 , P-value <0 . 001 ) . Seropositive samples against serovar Sejroe were significantly ( P-value <0 . 001 ) more frequent in the northeast ( OR 2 . 47 ) and the east ( OR 2 . 61 ) . The results of the multivariable logistic model suggest a significant influence of the type of species on the seroprevalence of serovar Tarassovi . The risk of infection was significantly higher ( P-value <0 . 001 ) in buffaloes ( OR 10 . 76 ) and cattle ( OR 2 . 26 ) compared to humans . Analysis according to region showed a higher risk of infection ( OR 3 . 57 , P-value <0 . 001 ) in the southern region and a lower risk in the north ( OR 0 . 23 , P-value <0 . 05 ) . Multivariable logistic model with interaction effects were found to be significant for serovar Shermani and MAT positivity . In general , seropositive samples would be identified for serovar Shermani that are significantly ( P-value <0 . 001 ) associated with buffaloes ( OR 4 . 72 ) and cattle ( OR 2 . 95 ) and also with the northeastern ( OR 4 . 42 ) and southern regions ( OR 5 . 02 ) . However , a significant interaction effect must be considered . For example , to compare the risk in buffalo in northeastern to the reference ( in humans in central ) , three terms must be multiplied , i . e . , OR for buffalo ( 4 . 72 ) , OR in northeastern ( 4 . 42 ) , and OR for buffalo in northeastern ( 0 . 03 ) . So the result of interaction OR equaled 0 . 64 . The overall MAT appearance indicated higher seroprevalence in buffaloes ( OR 4 . 75 ) and cattle ( OR 2 . 88 ) as the species effect , and in higher seroprevalence in the northern ( OR 1 . 51 ) , northeastern ( OR 4 . 07 ) and southern region ( OR 4 . 35 ) as the region effect . Moreover , the interaction terms must also be considered .
This study provides basic knowledge on serological examination based on a large data set that was gathered by passive surveillance during 2010–2015 among livestock ( buffaloes , cattle , and pigs ) and humans in Thailand . Anti-Leptospira antibodies are prevalent in all 5 regions as well as in livestock and humans . This study is the first to investigate the link between livestock and humans on Leptospira serovars endemic in Thailand . This study constitutes an important epidemiological approach and the results may increase comprehension of leptospiral serovar distribution at the regional level . Previous studies suggested that livestock could play an important role as source of human leptospirosis infection [12 , 28] . High seroprevalence in buffaloes and cattle was also observed , which may increase the exposure level and thus a high risk of infection in humans [12 , 29] . In this study , we used the MAT to determine seropositivity against a panel of 23 reference serovars including the local ones . As there is no absolute congruency about MAT cut-off titers , we used the recommended cut-off titer ≥ 1:100 [15 , 20] . MAT is the decade-long gold standard and most commonly used serological test in routine leptospirosis laboratories [30] . However , the test may yield false-negative results and be flawed by cross-reactions [4] . It is also a reported poor predictor of the infecting serovar [31] . However , the MAT provides information about Leptospira serogroups circulating in respective species and the immune response of the host . In livestock , the highest seroprevalence was found in cattle , whereas lowest prevalence occurred in pigs regardless the region . This finding could be explained by the living conditions of pigs leading to a low exposure level to leptospires in the environment due to being caged and/or fenced . In contrast , most cattle are free to move in their environment . Another factor could be the feeding of antibiotics to pigs to prevent leptospirosis infection . In general , pigs also have a shorter life span when compared to cattle and buffaloes leading to shorter exposure time . Regardless of species , the highest seroprevalence was found in southern Thailand and the least was found in the north . This may be attributed to higher rainfall in the south leading to higher exposure to contaminated water and soil [32] even though there are more agricultural areas cultivating rice in northern and northeastern regions . This study indicates a diversity of Leptospira serovars occurring in humans as well as in livestock in all 5 regions of Thailand . In humans , serovar Shermani has the highest seroprevalence , followed by serovars Bratislava , Panama , and Sejroe . Shermani prevalence was also highest in all regions . The high prevalence of serovar Shermani corresponds well with a previous report detecting mainly serovar Shermani in humans [13] . However , the previous study was limited to sample collection in Bangkok . Previously , the patients with a clinical diagnosis of leptospirosis from March 2003 to November 2004 in Thailand were found to positive cultured Leptospira , i . e . , L . interrogans serovar Autumnalis ( 7 ) , L . interrogans serovar Bataviae ( 2 ) , L . interrogans serovar Pyrogenes ( 2 ) , L . borgpetersenii serovar Javanica ( 1 ) , L . interrogans serovar Hebdomadis ( 1 ) , L . interrogans serovar Grippotyphosa ( 1 ) , and an unidentified serovar ( 1 ) [7] . However , those serovars were not consistence with this study . This may be attributed to the time lag , land use , and environmental factors [10] . The infection by serovar Shermani in humans could be attributed to a high contact rate with livestock or leptospira contaminated environmental sources . Livestock is known to be a host reservoir for leptospires and is potential source of infection for humans [33 , 34] . Our statistical analysis indicated that livestock were more frequently seropositive for the infecting serovar Shermani than were humans . We observed low prevalence of antibodies against serovars Bratislava and Sejroe in human samples for all regions . Historically , Bratislava [35–38] and Sejroe [38] were the main infecting serovars in humans across different areas in Thailand . The logistic modeling results with respect to serovar Bratislava prevalence indicated a lower occurrence in livestock when compared to humans . The species effect for serovar Sejroe was found to be insignificant . Sejroe infection could also be attributed to the interaction between rodents , as they are the natural carriers of this serovar in Thailand [39] . Seropositive samples against serovar Panama were detected in humans , particularly in the central region , but not in livestock . High MAT titers against this serovar were also frequently found in the province of Khammouane , Lao PDR [40] , in the province of Tien Giang ( Mekong delta ) , Viet Nam [41] . However , associations may need further investigation . In buffaloes , the most common infecting serovars were Shermani , Ranarum , and Tarassovi . The identification of serovar Ranarum infections corresponds well with a previous study [12] , and the seropositivity against serovars Shermani and Tarassovi corresponds to previous positive identifications made in Sakon Nakhon Province , Thailand [42] . The present study suggests that buffaloes could be the maintenance host of the serovars Shermani , Ranarum , and Tarassovi . It shall be noted that a single positive sample was identified in the south . A smaller data set was available for that region . Serum samples from cattle showed highest MAT titers against serovars Shermani and Ranarum in all regions . Similar to buffaloes , cattle might be a host for those two serovars . The transmission of these serovars between cattle and buffaloes is likely due to shared pastures and water sources . A similar observation of seropositive samples with the other serovars amongst cattle and buffaloes was made in Katavi-Rukwa , Tanzania [33] and in Turkey [43] . Seropositivity against multiple serovars was most common for serovars Ranarum and Shermani in samples from cattle and buffaloes . As a consequence of singlet tests , we were not able to distinguish between cross-reaction and co-infection ( sensu stricto or sensu lato ) . In the case of co-infection sensu stricto , animals may have been infected by different serogroups during the same period [14] . In the case of co-infection sensu lato , animals may have been exposed to a previous serogroup and subsequently exposed to another serogroup resulting in samples positive against multiple serogroups [44] . The later infection may then elicit an immune-response of cross-reactive polyclonal anti-leptospiral antibodies [13 , 33] . It shall be noted that this is the first study to report seropositive samples against RAN/SHE in Thailand . In pigs , the predominant infecting serovar was serovar Shermani followed by serovar Ranarum . This study also suggests that pigs can be maintenance hosts of serovar Shermani beside buffaloes and cattle . A great difference in prevalence between serovars Shermani and Ranarum was observed in pigs but not in buffaloes and cattle . We explain this observation with different feeding and living behavior as described above . Serovar Shermani may have a broader distribution with respect to region , while serovar Ranarum occurrence may be limited to grazing grounds and wet lands , which are not the natural habitat of pigs . Our results highlight that the most abundant infecting serovar is Shermani across humans , buffaloes , cattle , pigs and regions . This suggests a possible transmission pathway between humans and livestock . However , a previous study based on active surveillance [12] found less frequent seropositive samples against serovar Shermani when compared with this study , in particular for livestock . When we compared ratio of serovar Shermani relative to serovar Ranarum between previous study and current one , we obtained an increase of 3 . 33-fold , 3 . 62-fold , and 153-fold in buffaloes , cattle , and pigs respectively . Possible reasons are the time lag between the two studies , landscape ecology variations , differences in land use and environmental factors , such as humidity , climate , and animal behavior [10] . Differences in data from livestock between the studies may arise from the individual decision making of farmers to send samples as well as recognition of the pathological conditions of the animals . The correlation of distributions of Leptospira serovar proportions across regions showed a similar pattern for cattle but differed for buffaloes . Cattle have greater commercial relevancy for bush meat production than buffaloes . As a consequence , greater movement of cattle between regions may occurs , which then may lead to similar serovar patterns and higher serovar diversity than in buffaloes . In humans , our results suggested that the infecting serovar distribution in the southern region ( high prevalence with 39% ) was different to all other regions showing a high proportion of serovars Shermani and Tarassovi . The correlations of titers against the 23 serovars across regions in humans compared to livestock also differed , measuring greater serovar diversity in humans . Some seropositive samples would only be found in humans ( serovar Ballum , Cynopteri , and Panama ) and some were absent ( serovars Manhao and Pyrogenes ) . The exclusivity of some serovars to human samples could be explained by several other sources of infection , for example the transmission path via rodents . Indeed , the present study was based on data derived from passive surveillance system . One must interpret the results with caution . Thus , an active surveillance with well sample collection design is also suggested to confirm our passive surveillance to perform to better understand the disease transmission in the field . The size of samples in each species and region could not be controlled and depended on farmers decision making and public health campaigning . The samples in this study were not exclusively collected for leptospiral surveillance . However , this sample size was large enough to illustrate the whole picture of leptospiral infection in livestock in Thailand . Additionally , the samples were collected from different scales of livestock operations . The risk of leptospiral infection was , therefore , different . However , the present study focused on holistic picture of leptospirosis occurrence in livestock at national level regardless types and sizes of the farms . A further study on comparison of leptospirosis in different livestock settings is suggested to elaborate this point . Logically , animal density should affect epidemiology of leptospirosis . The regions with higher density may pose a higher risk of leptospiral infection . However , this aspect was not focused in our study as the reliable data on animal census in each region of Thailand , have not yet existed . The animal identification system has still been developing . Once the system is fully set up , the study on animal density and risk of leptospiral infection in different regions of Thailand is strongly recommended . Animal movement is also an important factor that may contribute to the exchange of leptospiral serovars among different regions in the country . Nonetheless , animal movement data in Thailand is not publicly accessible . The data is officially hosted by the Department of Livestock Development . The joint research with this institute to visualize the dynamic network of livestock movement is suggested . Subsequently , the study on animal movement and leptospiral distribution along the movement network should be initiated . Other animal such as wild animal and rodents were not included in this study because we focused on livestock . Wild animals may results in seroprevalence in humans and livestock . Wild animal can be identified as carries of leptospires , which can transfer to humans and livestock [45] . Rodent has been found as reservoirs for human leptospirosis in Thailand especially for L . interrogans and L . borgpetersenii species [46] . Rats can be an environmental risk factor by infestation in slums residents in Brazil [47] , whereas other mammals can be major reservoir for human leptospirosis , that highlights an importance of leptospiral surveillance beyond rodent species [48] . Other factors such as environmental effects and climate change were also not included . The temporal aspects were not included , this can influence control measures as leptospirosis pattern peak in rainy season and flood events [28] . In fact , the results were analysis based on the interpretation of seropositivity , which indicates a previous exposure to leptospires in the past . The interpretation of occurred infection should be concerned . The age of animals was not recorded in this study . Older animals have more opportunity for leptospires exposure , resulting in an increase of opportunities for seropositivity against multiple serovars [44] . However , leptospiral serovar diversity was observed in humans and livestock . We identified the current most abundant Thailand-endemic leptospiral serovars , foremost serovar Shermani , using MAT to test human and livestock serum samples . Serovar Shermani could be considered a potential public health risk as an emerging serovar occurring at high frequency in humans as well as livestock . The risk of human infection via livestock may be caused directly by contact with an infected animal or indirectly via animal products , mostly contaminated urine [34] . The infection in animal may continue for several months to a year [14] . The public health sector should increase awareness of high-risk groups , in particular abattoir workers , livestock keepers , farmers , and other such individuals with close contact to host and carrier animals . The finding of same serovar distributions may support public health officials in setting up effective intervention and control measures . However , further studies employing molecular typing or real-time PCR are recommended to identify leptospires and confirm the interspecies transmission . A cross-sectional survey should be conducted , especially in abattoirs and animal farms .
|
Leptospirosis is an important worldwide zoonotic disease , particularly in tropical and subtropical countries . The infection in humans is caused by either direct contact with products of infected animals , mainly urine , or by indirect contact via a contaminated environment . The animal hosts are thus considered reservoirs for human infection as livestock in Thailand usually live in close contact with householders in rural areas . However , the links of Leptospira serovars in humans and livestock in Thailand are poorly understood . Therefore , we illustrate the circulation of Leptospira serovars in humans and livestock during the past six years . The cross-correlations of the seroprevalence distribution were investigated to assess similarity between serovars across and within both , species and regions . The results suggest that livestock could be a potential source for human infection as sample analysis revealed a predominance of the same serovars . This information will increase public health awareness and may benefit especially high risk groups such as abattoir workers , livestock owners , farmers and other animal handling personal .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"livestock",
"medicine",
"and",
"health",
"sciences",
"leptospira",
"pathology",
"and",
"laboratory",
"medicine",
"ruminants",
"pathogens",
"geographical",
"locations",
"microbiology",
"vertebrates",
"tropical",
"diseases",
"buffaloes",
"animals",
"mammals",
"bacterial",
"diseases",
"neglected",
"tropical",
"diseases",
"bacteria",
"bacterial",
"pathogens",
"infectious",
"diseases",
"swine",
"zoonoses",
"medical",
"microbiology",
"microbial",
"pathogens",
"leptospirosis",
"agriculture",
"people",
"and",
"places",
"asia",
"biology",
"and",
"life",
"sciences",
"leptospira",
"interrogans",
"cattle",
"amniotes",
"bovines",
"organisms",
"thailand"
] |
2017
|
Investigation on predominant Leptospira serovars and its distribution in humans and livestock in Thailand, 2010-2015
|
In August 2014 , the National Institute for Communicable Diseases ( NICD ) in South Africa established a modular high-biosafety field Ebola diagnostic laboratory ( SA FEDL ) near Freetown , Sierra Leone in response to the rapidly increasing number of Ebola virus disease ( EVD ) cases . The SA FEDL operated in the Western Area of Sierra Leone , which remained a “hotspot” of the EVD epidemic for months . The FEDL was the only diagnostic capacity available to respond to the overwhelming demand for rapid EVD laboratory diagnosis for several weeks in the initial stages of the EVD crisis in the capital of Sierra Leone . Furthermore , the NICD set out to establish local capacity amongst Sierra Leonean nationals in all aspects of the FEDL functions from the outset . This led to the successful hand-over of the FEDL to the Sierra Leone Ministry of Health and Sanitation in March 2015 . Between 25 August 2014 and 22 June 2016 , the laboratory tested 11 , 250 specimens mostly from the Western Urban and Western Rural regions of Sierra Leone , of which 2 , 379 ( 21 . 14% ) tested positive for Ebola virus RNA . The bio-safety standards and the portability of the SA FEDL , offered a cost-effective and practical alternative for the rapid deployment of a field-operated high biocontainment facility . The SA FEDL teams demonstrated that it is highly beneficial to train the national staff in the course of formidable disease outbreak and accomplished their full integration into all operational and diagnostic aspects of the laboratory . This initiative contributed to the international efforts in bringing the EVD outbreak under control in Sierra Leone , as well as capacitating local African scientists and technologists to respond to diagnostic needs that might be required in future outbreaks of highly contagious pathogens .
Ebola virus ( EBOV ) was previously only identified to cause outbreaks of Ebola virus disease ( EVD ) in central equatorial Africa [1] , but emerged in West Africa in late 2013 in Guinea [2] , and spread into Liberia in March , Sierra Leone in May , and Nigeria in late July 2014 [3] . The EVD outbreak was declared by the World Health Organization ( WHO ) as a public health emergency of international concern ( PHEIC ) on 8 August 2014 [4] . The United Nations ( UN ) Security Council recognised the outbreak as a threat to peace and security on 19th September 2014 [5] , and consequently the UN Mission for Ebola Emergency Response ( UNMEER ) was deployed to combat the Ebola crisis in West Africa [6] . The index case is thought to have become infected late in December 2013 in a village located in the forest region of south-east Guinea [2 , 7] . EVD spread unrecognised from its unexpected source in Guinea to the neighbouring countries [8] to become the largest outbreak in the recorded history of the disease [9 , 10] . The PHEIC related to EVD in West Africa was formally lifted on 29 March 2016 [11] . A total of 28 , 616 confirmed , probable and suspected cases were reported in Guinea , Liberia and Sierra Leone , with 11 , 310 deaths [12] . Confirmed EVD cases among health care workers ( HCWs ) amounted to 881 of which 513 ( 58 . 2% ) were fatal [13] . The emergence of EVD in West Africa had a severe impact on public health , social life , national education programmes , agriculture , transportation , domestic and international markets [14] . The high EVD death toll amongst HCWs had impaired the already weak health system in Guinea , Sierra Leone , and Liberia , and might have a long term and severe effect on maternal , infant and under-5 year mortality [15] . A scarcity of resources in testing for EBOV infection left patients stranded in isolation wards for days with little to no management . The establishment of rapid and more widely accessible diagnostic capacities in the West African countries affected by the EVD epidemic was identified as one of the most urgent priorities to prevent the further escalation of the EBOV-related health and humanitarian crisis [10 , 16] . The first EVD cases in Sierra Leone were reported in May 2014 in the Kailahun District that shares its borders with Guinea and Liberia . The disease subsequently spread to the neighbouring Kenema District in June 2014 . Between 23 May and 30 August 2014 , the two districts had the highest number of EVD confirmed cases , with the incidence rates during this time of 3 . 8 and 7 . 0 per 100 000 population per week , respectively [17] . Laboratory diagnosis of the first EVD cases in Sierra Leone was made on 25 May by the Lassa diagnostic laboratory at the Kenema Government Hospital ( KGH ) [18] . KGH was the only facility in the country that could provide care to EVD patients at the onset of the outbreak [19] . The first cases were likely exposed to EBOV when attending the funeral of a traditional healer who had been treating Ebola victims in neighbouring Guinea [3 , 18] . Due to severe staff losses , increasing numbers of EVD patients , shortages in the medical supplies , compounded by hostile situations and striking of the remaining nurses and laboratory technicians , operations of both the KGH and the Lassa diagnostic laboratory were difficult to continue and needed urgent assistance . Aid was eventually provided by Médecins Sans Frontières/Doctors Without Borders ( MSF ) , WHO and other international partners [18 , 19] . The cluster of EVD cases involving HCWs in the Kenema District was one of the largest ever reported in the history of the disease [19] . Of the eight laboratory technicians , only one survived the EVD outbreak [20] . By late July 2014 , EVD cases were detected in the capital of Sierra Leone , Freetown [3 , 10] . The fear that the deadly virus would spread to the congested , impoverished , post-conflict West African urban areas had become reality . Since the introduction of EBOV into Sierra Leone , new EVD cases were reported daily in increasing numbers in nearly all provinces , but particularly in the Western Urban and Western Rural regions of Sierra Leone . During the period May 23 to October 31 , 2014 , there were 3 , 854 laboratory-confirmed cases of EVD reported in Sierra Leone , including 199 cases in HCWs . The confirmed EVD incidence was 103-fold higher in HCWs than that in the general population in Sierra Leone [21] . The weak health system , lack of experience with management of EVD , poor community awareness , and highly deficient laboratory diagnostic capacity further contributed to the spread of the disease [10 , 16 , 22 , 23] . In response to the public health emergency caused by the EVD epidemic , the National Institute for Communicable Diseases ( NICD ) established a modular high biosafety field Ebola diagnostic laboratory ( FEDL ) within the compound of the Lakka Tuberculosis ( TB ) Hospital located on the outskirts of Freetown in the second half of August 2014 . This paper describes the deployment , organisation , diagnostic and biosafety procedures utilised by the South African ( SA ) FEDL in Sierra Leone , the training of Sierra Leonean national staff during the EVD outbreak , laboratory results , and the hand-over of the SA FEDL to the Ministry of Health and Sanitation ( SL MoHS ) as a part of the SA EVD diagnostic capacity building initiative in West Africa .
We did not seek institutional review board approval for data collection in this study because data were collected as part of routine case management under an emergency response mandate from the government of Sierra Leone . As a part of routine practise , patients orally agreed to be tested for Ebola virus infection . Permission to carry out post-outbreak studies on samples transported from Sierra Leone to BSL4 facility in South Africa was given by the Government of Sierra Leone Office of the Sierra Leone Ethics and Scientific Review Committee , Directorate of Training and Research , Ministry of Health and Sanitation . To support the SA FEDL initiative in Sierra Leone , South African scientists and technologists throughout the NICD and the University of Pretoria were voluntarily mobilised . They undertook intensive in-house pre-deployment training in the safety and diagnostic procedures of the FEDL and a further 2–4 days on-site training by the preceding team upon arrival in Sierra Leone . The SA FEDL staff members completed the UN basic and advanced field safety and security programme training and were all deployed through the WHO Global Outbreak Alert and Response Network recruitment system . Travel restrictions leading to the discontinuation of commercial flights to West African countries affected by the EBOV epidemic presented difficulties in the rotation of SA FEDL teams and shipping of laboratory supplies to Freetown . The UN World Food Programme established the Humanitarian Air Service ( UNHAS ) to augment the capacity of humanitarian efforts under UNMEER . With the assistance of the SA National Department of Health , most of SA FEDL teams were transported using commercial flights from Johannesburg via Dakar ( Senegal ) or Accra ( Ghana ) combined with UNHAS flights to Freetown . The infrastructural and operational capacity of the SA FEDL was designed to primarily conduct molecular EVD and Lassa fever laboratory testing and to ensure adequate safety procedures for working with BSL-4 pathogens under field conditions in a poorly resourced country . Originally , the diagnostic capacity of the laboratory was designed to process and test a maximum of 58 specimens per day . The first SA team consisted of four NICD staff members experienced in laboratory diagnostics of viral haemorrhagic fevers ( VHFs ) , and the construction and technical management of biocontainment facilities . This team was tasked with the selection of an appropriate site for the deployment of the SA FEDL and establishing its full operation . The team arrived in Freetown , on 17 August 2014 and the following day visited the WHO Country Office where it was introduced to the National Coordination Team of the Ebola Epidemic Response . The same day , the team visited the headquarters of the UN Development Programme in Freetown to obtain country security briefing by the UN Department of Safety and Security . On 18 August the team visited the Lakka Tuberculosis ( TB ) Hospital in Freetown where potential operating sites for the SA FEDL were inspected . The next day , the decision was made to deploy the SA FEDL to the campus of the Lakka TB Hospital , in a vacant building planned to be used in the future as a TB reference laboratory . On 19 August , the team inspected the interior structure and the layout of the building and started preparations for the establishment of the SA FEDL . On 20 August , the first shipment containing the negative pressure biological containment system ( IsoArk® , Beth-el Zikhron Yaaqov Industries Ltd . , Zikhron Yaaqov , Israel ) arrived from South Africa and the same day it was assembled in a designated room for “hot” processing of specimens . The second shipment containing laboratory equipment , a negative pressurised glovebox , rapid containment kits , generator , uninterrupted power supplies ( UPS ) , reagents , laboratory consumables , personal protective equipment , and other SA FEDL operation supplies arrived on 22 August . On 24 August , mock EVD testing , using synthetic RNA controls , was conducted to evaluate the work flow and the performance of the EBOV reverse transcription polymerase chain reaction ( RT-PCR ) assay under field conditions and to confirm the integrity of reagents after shipping . On 25 August the SA—FEDL was declared fully operational and the testing of the first blood specimens from EVD suspected cases commenced . The FEDL was located on the premises of the Lakka TB Hospital compound , Peninsula Road , Lakka near Freetown , in a building consisting of brick structure , steel roofing , lockable windows and three lockable entrances . All windows and doors were fitted with burglar bars . Air conditioner units were located in most of the rooms of the building . Two emergency electricity generators of different output capacities were used . The main back-up 60 kVA diesel-powered generator was routed to the main distribution board to supply electricity to the entire building . The second 5 . 5 kVA petrol generator was fed to the building by means of electrical extension cords and only supplied power to essential laboratory and storage equipment such as biocontainment devices , real time PCR instruments , refrigerators and freezers . In addition , essential laboratory equipment was connected to UPS to prevent power loss during operation and to serve as voltage stabilisers . The lay-out of the SA FEDL is given in Fig 1 . Procedures for centrifugation , aliquoting and inactivation of specimens as well as the addition of positive control RNA and the RT-PCR amplification of RNA templates were carried out in designated rooms that were separated from the PCR master mix preparation room , RNA extraction room and storage area by a laboratory airlock area ( Fig 1 ) . Chlorine disinfectant solution ( Medisure ) at a concentration of 0 . 5% ( 5000 ppm ) was used for the decontamination of pipette tips that came into contact with clinical specimens , primary specimen containers , the exterior of sample tubes containing plasma/lysis buffer in ziplock bags for their transfer out of biocontainment , and for the decontamination of the glovebox working area during and after “hot” processing . A lower concentration of 0 . 05% chlorine ( 500 ppm ) was used for all other decontamination procedures , including but not limited to decontamination of operators PPE when exiting the IsoArk , decontamination of specimen submission forms , secondary containers , work benches , and floors . Work benches in the RNA extraction and PCR master mix rooms , rapid containment kits and class II biosafety cabinet work surfaces were decontaminated daily with 0 . 05% chlorine disinfectant solution followed by wipe down with 70% ethanol . At the end of each day , the ultra-violet ( UV ) germicidal lamp fitted to the biosafety cabinet was operated for 15 minutes , while portable UV germicidal lamps were operated for 20 minutes in the RNA extraction and PCR master mix rooms . Staff receiving specimens wore PPE that included , a surgical gown , gloves , a N95 mask and eye protection . The outer boxes/specimen containers , as well as case investigation forms were first placed into plastic biohazard bags and then taken to the airlock of the IsoArk for decontamination by mist spray with chlorine disinfectant . Specimens were then either moved to the temporary storage fridge for later processing or taken into IsoArk biocontainment chamber for immediate processing . Specimens were cross-matched with the information provided in the case investigation forms . A unique numerical laboratory identification number was assigned to each specimen . Information on specimens accepted for testing was recorded in the specimen register and included a laboratory submission number , a patient name and surname , age , sex , reference number and name of referral facility . More detailed information was collected and electronically recorded from case investigation forms by the WHO employed data capture clerk , using the reporting template provided by the SL MoHS . Reception of specimens was refused or subjected to storage until specimen submission forms could be obtained or if the required patient history could be provided by other confirmable means . Patient data captured electronically were copied into the Ebola master database with the corresponding FEDL reference laboratory submission number and the sample RT-PCR cycle threshold ( Ct ) value . The leader of each FEDL rotation team was responsible for ensuring that the master database was safely maintained and that the data and confidentiality of results was safeguarded . Results were mailed electronically to authorised email addresses using the reporting template supplied by the SL MoHS and WHO . Urgent results were communicated directly to attending healthcare workers by telephone , while reporting of results to family or community members by FEDL staff was strictly forbidden . Protection of staff from direct or indirect contact with infectious substances present in clinical samples or in droplets that could be formed during handling and processing of diagnostic specimens was achieved using PPE ( Fig 2 ) , including a Powered Air Purifying Respirator ( PAPR ) coupled to a full-face hood . The full set of PPE for “hot” work in the IsoArk for one operator included: washable socks; single use overshoes; a single use scrub set ( pants and shirt ) ; a single use Tyvek coverall; reusable gumboots; two pairs of single use surgical examination gloves; one pair of single use long-cuff examination gloves; a single use fluid resistant , rear fastening folio gown and the reusable parts of the PAPR ( filter-blower unit , air-hose , double shrouded full-face hood and rechargeable battery ) . The IsoArk ( Figs 2 and 3 ) features a high air change rate , low noise level , and adjustable airflow rate up to 2 , 200 m3/hr . The main components of IsoArk included: an airlock , a main chamber , and an air filtration system . When entering the airlock , an electro-optical eye sensor automatically switched the filtration system to a high flush mode that increased the airflow through the airlock . This reduced the waiting time for a complete air change in the airlock and ensured that the negative pressure was maintained even when entering or exiting the main chamber . Electrical cables were passed into the isolated area through specialised utility sleeves . The integrated air Filtration System FA 2000 combined a highly efficient HEPA-filter with a UV-radiation source . The glovebox used for primary containment of specimens ( Fig 3 ) features a negatively pressurised vinyl box that allows safe processing of biohazardous material and a solid perspex pass-box that enables safe passage of materials into and out of the containment area . PPE worn by staff exiting the containment area was decontaminated by two rounds of mist-spray using a pressurised container with spray nozzle containing 0 . 05% chlorine disinfectant . The first mist spray was performed within the IsoArk biocontainment . After the removal and disposal of overshoe coverings and the first pair of gloves into a biological waste bag , a second mist-pray was performed inside the airlock . After allowing 3 minutes of contact time with the disinfectant solution , operators were permitted to leave the airlock and complete PPE doffing . Procedures in the “hot” room followed by exit from IsoArk are additionally illustrated in supplementary video material ( S1 Video ) . Manual RNA extraction was carried out using the QIAmp viral RNA kit ( Qiagen , Hilden , Germany ) according to the manufacturer’s instructions . Clotted or whole blood in vacutainer tubes were centrifuged , using safety buckets , inside the IsoArk for 3 minutes at 1 , 500 x g to separate serum/plasma from cellular content . Following centrifugation , the sealed buckets containing the blood tubes and swabs were placed inside the glovebox . Serum/plasma or swab medium was transferred to the pre-labelled tube containing AVL-carrier RNA buffer . Remaining plasma was transferred into an empty cryotube for long-term storage . The cryotubes containing processed specimens and biohazard bags containing the original blood tubes were mist-sprayed before transfer from the glovebox into the main chamber of the IsoArk . To ensure effective and rapid inactivation of EBOV in specimens before removal from biocontainment , a volume of 560 μl molecular grade absolute ethanol was added to each sample/AVL tube [24 , 25] . Tubes containing inactivated specimens were then placed into an airtight ziplock bag filled with chlorine disinfectant . The ziplock bag exterior was mist-sprayed and moved to the airlock of the IsoArk , and after 10 minutes contact time for inactivation was transferred to the RNA extraction room to finalize extraction . Extracted RNA specimens were transferred to the 4°C fridge until later addition of the PCR mastermix . Used blood tubes or tubes containing buccal swabs were filled with chlorine disinfectant within the glovebox , closed , and transferred from the glovebox to the IsoArk biocontainment area , where they were placed in a biological waste bag for incineration . Automated RNA extraction was carried out according to the manufacturer’s instructions ( MagMax96 , ThermoFisher Scientific , Waltham MA , USA ) using MagMax Express-96 Deep Well magnetic particle extractor ( Life Technologies , New York , USA ) . Clotted blood was centrifuged as described for processing of blood specimens for manual extraction . Vacutainer tubes containing whole blood samples ( EDTA or heparin blood ) were not centrifuged . Serum , whole blood or swab medium was transferred into the pre-labelled cryotube containing MagMax lysis buffer and carrier RNA . The remaining volume of processed serum , or buccal swab medium was transferred to a cryotube for long-term storage . Vacutainer tubes containing the remaining whole blood were not centrifuged but were temporarily stored at 4°C in case of the need for retesting . If not required for retesting , the next day the remaining blood was centrifuged and plasma aliquoted for long-term storage . Extracted RNA was aliquoted into pre-labelled tubes and stored at 4°C until addition to the PCR master mix . To evaluate the potential impact of automated versus manual RNA extraction on the diagnostic performance of EBOV L gene TaqMan assay , EBOV RNA was extracted from 72 sera of EVD suspected cases using both extraction methods , and subjected to comparative analysis . The real time RT-PCR protocol for the detection of EBOV RNA targeting the polymerase ( L ) gene was carried out as previously described [26] . The assay was run using real-time PCR instrument ( SmartCycler , Cepheid , Sunnyvale , CA , USA ) . The manufacturer’s protocol was followed when preparing the RT-PCR master mixes ( Qiagen One-Step RT-PCR kit , Qiagen , Hilden , Germany ) . The monitoring of quality performance of the EBOV L gene TaqMan assay involved testing of internal RNA controls having pre-determined Ct values . The L gene TaqMan assay run was considered valid when the following internal quality control ( IQC ) acceptance criteria were met: ( 1 ) the positive RNA control had a Ct value < 35 and fell within the predetermined upper ( UCL ) and lower ( LCL ) IQC limits ( 2 ) the positive extraction control had a Ct value < 35 and fell within the predetermined UCL and LCL IQC limits , ( 3 ) no amplification was detected in the negative control , and ( 4 ) no amplification was detected in the negative extraction control . When the IQC criteria for the PCR run were met , the results in the test sample were interpreted as follows: ( 1 ) positive result when Ct value was ≤40 , and ( 2 ) negative results when no amplification was detected or the Ct value was > 40 . To exclude possible low level of cross-contaminations , specimens yielding high Ct values , from 35 to 40 , were retested using re-extracted RNA . Clinicians were advised that patients who tested negative and with Ct values of 35 to 40 should be re-bleed and tested again . The SA FEDL took part in two external quality assessment runs conducted by WHO Country Office Sierra Leone , the Centers for Disease Control and Prevention ( CDC ) Atlanta , WHO and the European Network for Diagnostics of Imported Viral Diseases [27] . Laboratory generated bio-waste was burned in 210 litre capacity oil drums re-designed for field incineration using diesel fuel as ignition source .
Initially , the operation of the SA FEDL was planned for 2–3 teams and for a maximum of 3 months duration . This plan had to be revised following the ground experience of the first SA FEDL team . The lack of effective control of the EVD outbreak in the second half of 2014 prompted the need for drastic upscaling of the EVD diagnostic capacity in Sierra Leone Preparations were then made for the deployment of up to ten SA teams . However , this was eventually not required due to the success of outbreak response measures that resulted in controlling the EVD crisis by the end of January 2015 . In addition , the training of national staff enabled the handover of the SA FEDL to SL MoHS in March 2015 . Eight SA teams each consisting of 2–5 members , were deployed from 17 August 2014 to 25 March 2015 . Each team was deployed for 5–7 weeks , with 2–4 days overlap with the outgoing team . Each team worked 12–16 hours a day , 7 days a week , and usually no days off . On most days , RNA extractions and RT-PCR were performed both in the morning and afternoon . After two weeks of operation in Freetown , FEDL staff undertook intensive training of five Sierra Leonean staff in all aspects of the FEDL logistic and diagnostic operations . The Sierra Leonean nationals were gradually incorporated into all routine functions of the FEDL and became invaluable in its long-term and sustainable operation . Of the five Sierra Leonean nationals trained , two had BSc degree and two had MSc degree in biology , one had Diploma in Paramedic Sciences . Specimens from suspected EVD cases were submitted to the FEDL throughout the day , sometimes until late at night . During the first weeks of operation , specimens were submitted from several regions of Sierra Leone . Specimens received from October 2014 onwards mostly originated from Freetown health care facilities , including the Emergency Ebola Treatment Centre , Connaught Hospital , 34 Military Hospital , Macauley Street Clinic , Hastings Ebola Treatment Centre , Princess Christian Hospital , Lumley Governmental Hospital , and Ola During Children’s Hospital . The EBOV L gene TaqMan assay results were usually issued twice a day . Submission of specimens rapidly increased during the first weeks of operation and exceeded the maximum testing capacity of 58 specimens per day on several occasions during the last two weeks of September 2014 ( Fig 4 ) . The week of 14–20 September 2014 included a 3-day National House-to-House Campaign to enhance active EVD surveillance [17] and resulted in increased demand for laboratory confirmation of EVD suspected cases . The increasing demand for EVD diagnosis necessitated improvement of the technical laboratory capacity . This was achieved by the use of the MagMAX™ 96 Express automated nucleic acid extractor . In a comparison study , the mean Ct value of 72 sera subjected to manual and automated extraction was 31 . 9 ±7 . 9 standard deviations ( SD ) ( 95% confidence interval [CI] , 30 . 0–33 . 7 ) and 31 . 6±8 . 3 SD ( 95% CI , 29 . 7–33 . 6 ) , respectively . The Wilcoxon matched-pairs signed-ranks test gave the two-tailed P value of 0 . 2194 , indicating no significant differences between the two extraction methods used . When using the cut-off of ≤40 Ct value for the EBOV L gene TaqMan assay , of 67 positive specimens after automated extraction , 65 ( 97% ) were also positive and two were negative after manual extraction . Of 5 negative specimens after automated extraction , 4 ( 80% ) were also negative , and one was positive after manual extraction . The observed discrepant results were noted only in samples with high Ct values ranging from 38 . 42 to 40 ( Fig 5 ) . Between 25 August 2014 and 22 June 2016 , the laboratory tested 11 , 256 specimens mostly obtained from the Western Urban and Western Rural regions of Sierra Leone , of which 2 , 379 ( 21 . 1% ) tested positive for Ebola virus RNA . These also included post-outbreak specimens tested during the WHO-recommended increased EVD surveillance period after the WHO declared Sierra Leone free from EVD on 07 November 2015 . From 25 August to 30 September 2014 ( 37 days ) when the SA FEDL was the sole operational Ebola diagnostic facility in Freetown , the laboratory tested 1 , 219 blood samples ( Table 1 ) and 260 buccal swabs ( Table 2 ) of which 62 . 2% and 38 . 8% were positive by RT-PCR , respectively . From 25 August 2014 to 31 March 2015 , a total of 7 , 267 specimens were tested , of which 4 , 299 were blood specimens ( Table 1 ) and 2 , 968 were buccal swabs ( Table 2 ) . Over this period the percentage of positive blood samples and buccal swabs drastically decreased from 62 . 2% to 13 . 2% ( Table 1 ) and from 38 . 8% to 1 . 4% ( Table 2 ) , respectively . From April 2015 to 22 June 2016 ( after the handover of the FEDL to the SL MoHS ) a total of 3 , 170 blood specimens ( Table 1 ) and 813 ( Table 2 ) buccal swabs were tested of which 8 ( 0 . 25% ) and 1 ( 0 . 12% ) were positive , respectively . The last two blood specimens that tested positive were sampled at the end of January 2016 ( Table 1 ) , while the last buccal swab that tested positive was sampled in May 2015 ( Table 2 ) . Irrespective of the month of submission , the mean EBOV L gene TaqMan assay Ct values for positive blood specimens were similar . In the total number of 1 , 758 positive blood specimens , the Ct values ranged from 12 . 37 to 39 . 99 with a mean value of 26 . 5 ± 5 . 38 ( Table 1 ) . The mean Ct values for positive buccal swabs were also comparable across the months tested . In the total number of 621 positive buccal swabs , the Ct values ranged from 15 . 82 to 40 . 0 with mean value of 26 . 8 ± 4 . 68 ( Table 2 ) . Of the total of 1 , 758 RT-PCR positive blood specimens , 1 , 552 ( 88 . 3% ) were collected from potentially viraemic EBOV patients ( Table 1 ) . Categorisation of EBOV L gene TaqMan assay positive results as indicating infectious versus non-infectious specimens was based on recently published correlation between Ct value of the L gene TaqMan assay and virus isolation results of EBOV from blood specimens [28] . In 1015 patients tested RT-PCR positive for the presence of EBOV in their blood and for whom day of onset of clinical symptoms was known , there was a strong linear correlation ( R2 0 . 9504 ) between Ct values and the days after onset . The lowest Ct values were recorded on days 0–3 ( mean 25 . 9 ± 4 . 7 SD ) , and the highest on days 20–27 ( mean 35 . 7 ± 1 . 7 SD ) after onset . The percentage of viremic patients ranged from 96 on days 4–7 ( n = 432 ) to 5 . 9 on days 21–27 ( n = 17 , one viremic patient with CT value of 33 . 07 on day 21 ) after onset ( Fig 6 ) . The mean time from symptoms onset to diagnosis ranged from 7 . 5 ± 3 . 8 SD days in October 2014 to 2 ± 2 . 7 SD in March 2015 . The mean time from collection of specimens to diagnosis ranged from 2 . 2 ± 1 . 2 SD days in October 2014 to 0 . 7 ± 0 . 8 SD in March 2015 . Through August 2014 to March 2015 , on average 53% , 34 . 6% , 10% , and 2 . 4% results were reported <6 hours ( h ) , <12 h , <24 h , and >24h after receiving of specimens , respectively . Owing to the excessive workloads , assays other than EBOV RT-PCR were not performed until early March 2015 . Since then malaria rapid test ( SD Bioline , Standard Diagnostics , Korea ) was conducted on request of the attending physician . Of 26 Ebola positive patients , one was also malaria positive ( 3 . 8% ) and of a total of 568 Ebola negative patients , 117 were malaria positive ( 20 . 6% ) . In two independent laboratory quality control trials the SA FEDL scored 100% in both trials [27] . One of the major challenges encountered during the first weeks of operation was a poorly organised system for the delivery of specimens , including: delivery of specimens very late at night , unsafe packaging resulting in specimens leaking , inappropriate primary containers ( e . g . glass blood collecting tubes , syringes loaded with blood and uncapped needles wrapped in plastic bag ) , the delivery of specimens without patient clinical history , discrepant patient information written on primary containers versus specimen submission form , and unlabelled primary containers . Another challenge was the lack of a consistent electricity and water supply . During the first three weeks of operation , the biocontainment negative pressure chamber , refrigerators , and laboratory equipment , including PCR instruments were mostly run using the 5 . 5 kVA petrol generator shipped from South Africa . During the third week of operation this generator broke down due to 24 hours of continuous operation for six consecutive days , resulting in suspension of FEDL services for half a day on the 4th of September 2014 . Extraordinary and commendable efforts were made by local authorities to provide an emergency generator which permitted the FEDL to resume operation later that same evening . The building where the FEDL was housed was eventually connected to the national electricity power grid as well as to an emergency diesel generator provided by the Sierra Leone government . However , the provision of uninterrupted electrical power and the required maintenance for the technical infrastructure of the building often remained an unresolved issue and resulted in damage to laboratory equipment . Three PCR instruments broke down , most likely due to operation in high ambient temperature and humidity as well as problems with the power supply . Additionally the UPS unit supporting the operation of strategic laboratory equipment eventually broke down . Dysfunctional air-conditioning units made work in the biocontainment chamber , glove box and other lab areas difficult due to high temperatures and humidity and posed not only a high level of discomfort but also potentially affected human judgement resulting in errors and safety risks . Technical problems forced the laboratory to be closed for a few days in the last week of February 2015 . The laboratory has since then continued to provide the Ebola diagnostic services and arrangements were made to address repairs and maintenance needs on a more regular basis . Communication via internet , including reporting of results was very time consuming and often impossible due to highly inefficient and occasionally completely dysfunctional internet and/or 3G cellular network connectivity . The maximum daily diagnostic capacity was exceeded shortly after the establishment of the FEDL , however , the demand for EVD laboratory testing with turnaround times of less than 24 hours after receipt of specimens continued to increase . By allocating extra working time this situation was partially addressed , but such efforts were not sustainable for a longer period of time for the small SA FEDL teams . The increasing demand for Ebola infection diagnosis and the high expectation for shorter turn-around times necessitated up-scaling of the laboratory capacity and the deployment of more staff . This was eventually achieved by the provision of additional local human resources , the use of additional PCR instruments as well as technical improvements to the laboratory . These included the redesign of the PCR master mix room and the use of UV lamps in the PCR master mix room and RNA extraction room , the replacement of the rapid containment kit by a biosafety cabinet , which allowed quicker and safer addition of RNA templates to PCR master mixes , and the use of the MagMAX™ 96 Express automated nucleic acid extraction system which increased the speed and capacity for extraction of RNA from clinical specimens . The use of a lockable -80°C freezer improved the capacity for the secure and long-term storage of processed and tested clinical materials before shipping them to a maximum biocontainment facility ( BSL-4 ) in Johannesburg , South Africa for further laboratory analysis and research , including the evaluation of the prototype Cepheid GeneXpert Ebola diagnostic assay [28] . On the 17th of March 2015 , a meeting took place at the SL MoHS , Freetown , to discuss the protocol for the official handover of the FEDL . The meeting was attended by the Minister of Health and Sanitation , Chief Medical Officer , the Director of Hospital and Laboratory Services , and NICD staff . During this meeting the handover protocol was accepted and agreed to be fully implemented . Consequently , on the 21st of March 2015 an on-site meeting was held at FEDL Freetown-Lakka to finalise the protocol and procedures for handover . On 23 March 2015 the NICD delegation visited the President of Sierra Leone , Dr Ernest Bai Koroma to report on the South African work in combating the EVD in Sierra Leone and specifically to discuss the handover of the SA FEDL to the SL MoHS . The official handover and donation of all laboratory , biosafety , and office equipment of the FEDL as well as diagnostic reagents allowing testing of at least 20 , 000 specimens took place on 24 March 2015 . The handover meeting was attended by several representatives of the Sierra Leone MoHS , including the Deputy Minister of Health , Deputy Chief Medical Officer , the Director Hospitals and Laboratory Services , the National Manager Laboratory Services , the Deputy Manager of the National Laboratory Services , the Sierra Leonean staff trained by NICD in the operation of the laboratory , and the staff of the NICD . Documents that were reviewed , signed and issued during the handover meeting included: the Safety and Operational Manual of the FEDL , training documents and the certification of local staff in diagnostic and operational competence , equipment inventory list , reagents and consumables inventory list , laboratory specimen register , EVD investigation forms , and the agreement on FEDL capacitation by the NICD after handover to MoHS which would be on-going until the end of the epidemic . The technical support included further guidance and ad hoc consultation via telecommunication , the supply of reagents , and consumables and financial assistance with repairs of air conditioners and water supply systems . The FEDL was fully decontaminated and its function as EVD diagnostic laboratory terminated on 12 November 2016 . Sierra Leonean staff taught by NICD trained ten additional nationals to enhance local rapid response team capacity for epidemic-prone diseases .
The Western Urban and Western Rural Area where the SA FEDL operated remained a “hotspot” of the EVD epidemic in Sierra Leone for many months; from 23 May 2014 to 31 January 2015 districts in these areas , which include Freetown , had the highest number of confirmed cases ( n = 3158 ) , 39 . 2% of all cases nationwide [17] . At the time the first SA FEDL team arrived in Freetown delays in testing samples for EVD kept patients stranded for days in isolation wards and contributed to fears raised against seeking treatment . Severely sick patients who eventually recovered after admission to a hospital had watched Ebola virus killing others in the same room and it took days to test specimens from suspected EVD cases [16] . The nearest EVD diagnostic capacity available was at the KGH located 300 km east from Freetown ( about five hours drive ) . At that time the laboratory was overloaded with blood samples from around the country . The delay in the laboratory confirmation of EVD suspected cases meant that patients were dying before they could be transported to a treatment centre . Within 6 days upon the arrival of the first SA FEDL team , diagnostic results were being sent to isolation wards twice a day and allowed some patients to leave the hospital on the same day that they were admitted [16] . Most results ( 87 . 6% ) were issued by SA FEDL within less than 6–12 h after receiving of specimens . The test procedures only , from RNA extraction to reading RT-PCR results for each RT-PCR run ( max 29 specimens per run including controls run on two Smartcyclers ) took less than 4 h . However , diagnostic process concerns not only conducting an assay . Due to the high number of samples received daily and irregular delivery times specimens were batched to conduct at least two diagnostic assay runs per day: morning and afternoon runs . Much time had to be spent to safely receive and handle specimens submissions often delivered unsafely packed and/or collected in unsafe primary containers , including blood-filled syringes with needles attached , all warped in plastic shopping bags . Also much time had to be spent on qualifying samples for testing due to discrepancies of information provided in submissions ( investigation ) forms versus those written on primary containers . For these reasons even the most experienced laboratory could on average process only about 71% of samples on the day they arrived at the laboratory [29] . Intensification and improved coordination of outbreak control measures , including timely case finding and contact tracing , and collection of specimens resulted in significant shortening of time between onset of clinical symptoms and diagnosis , and between collection and diagnosis . Improvements of laboratory capacity enabled more rapid testing and reporting of results . In the future , field laboratories will likely utilise simpler on-site detection techniques , including point-of-care testing [28] to reduce time from sample collection to laboratory diagnosis . Rapid and accurate laboratory confirmation of EVD cases was paramount for case management and contact tracing , and the subsequent control of the outbreak . Some of the major contributions of the SA FEDL included: ( 1 ) augmenting the local diagnostic capacity , ( 2 ) alleviating the problem of logistics that may have led to delayed testing when specimens had to be shipped to regional or international reference laboratories for testing , ( 3 ) aiding in patient management that involved resource intensive barrier nursing and isolation of suspected cases , ( 4 ) rapid molecular testing of buccal swabs from deceased persons that was essential to manage secure burials . Training of local staff in all operations of the SA FEDL was instrumental in the long-term provision of EVD diagnostic services by this facility . It was highly beneficial for both partners , including the cost-effectiveness of laboratory operation when demand for testing decreased towards the end of the outbreak . Initially the proposal of incorporating Sierra Leonean staff in laboratory operations during the acute phase of the EVD outbreak was met with some resistance and there were doubts as to whether this could be achieved . The SA FEDL was the only facility which undertook training of national staff during the escalation stage of the EVD outbreak in the Western Urban and Western Rural Area of Sierra Leone . Most other international laboratories planned training of Sierra Leonean nationals towards the end or after the outbreak . The successful training and incorporation of national staff into all SA FEDL operations was possible because of the trust of national staff in the safety procedures and biocontainment devices used , including the high level of both primary and secondary biocontainment barriers and the use of the best available portable biocontainment devices . Having finally contained the EVD situation in March 2016 [13] , Sierra Leone has maintained heightened surveillance with testing of all reported deaths and prompt investigation and testing of all suspected cases . The SA FEDL played an important role in the WHO-recommended enhanced surveillance for EVD post-outbreak until it was decommissioned in November 2016 . By the end of September 2014 there were only three field-established laboratories operating in Sierra Leone . , An additional eight laboratories were established between early December and mid-January 2015 The combined effort and collaboration of these laboratories , of which six operated in the most affected Western Area of Sierra Leone , eventually not only helped to address the high demand for laboratory EVD diagnosis but also increased the total EVD diagnostic capacity in Sierra Leone . All of the field-established laboratories in Sierra Leone used real-time RT-PCR diagnostic assays [30] targeting either the viral polymerase ( L ) [26] or nucleocapsid ( NP ) and VP40 genes of EBOV [29] . However , these assays were not extensively validated in the field , mostly due to limited availability of clinical specimens in the past . Our results confirm earlier findings that buccal swabs are a useful alternative specimen source to whole blood/serum [31] , and that this non-invasive specimen collection should be considered in testing unexpected community deaths . Testing of post-mortem throat swabs was also shown to be a reliable and sensitive method for EVD diagnosis and surveillance [32] . The levels of EBOV RNA detection and duration of viremia in our study were similar to other published results [33 , 34] . EVD patients who have a Ct of <17 have a casa fatality rate ( CFR ) of 95% , and those with a Ct of >26 have a CFR of 15% [32] . It was also shown that most blood samples having Ct value > 33 . 7 tested negative for replicating EBOV [28] . Thus the RT-PCR Ct value could be used in discrimination between infectious and non-infectious specimens and has strong prognostic utility for aiding EVD patient management and care . It was not always possible to retest EVD suspect patients . In this context , one had to emphasise , that a single PCR result needed to be interpreted with great caution with regards to a definitive confirmation of EVD diagnosis . Inhibitors present in the sample may have caused low or no amplification resulting in false negative results . Common inhibitors derived from red blood cells include haemoglobin and lactoferrin that may be released in haemolysed specimens . In addition molecules such as bile , CaCl2 , EDTA , FeCl3 and heparin may also inhibit Taq polymerase activity [35] . The presence of inhibitory factors in a biological specimen is difficult to anticipate . To address some of these problems , the human betaglobin gene was used as a control for sample integrity in some of the RT-PCR protocols [29] . It is recommended that a PCR result be interpreted along with other test results where available ( e . g . serology ) whilst also considering the clinical history of the case . When possible , results should be confirmed on subsequent or repeat specimens . Typically , very low stringency PCR conditions are used and erroneous amplification may occur . Furthermore nucleic acid amplification methods are prone to contamination and this should always remain a concern . One of the concerns related to potentially uncertain results was associated with handling and testing of large numbers of specimens over a long time and under pressing circumstances which might contribute to contaminations . Each RT-PCR run usually consisted of specimens with very high to low concentration or no EBOV presence , thus specimens yielding high Ct values could potentially represent cross-contaminations . Consequently , these specimens were always retested using re-extracted RNA . The presence of RT-PCR inhibitors is a potential problem , but likely not very common . Most field laboratories operating in West Africa were using RT-PCR assay for the detection of EBOV L gene . There was a concern that the unprecedented number of EVD cases in West Africa ( exceptional high number of EBOV person-to-person transmissions ) might induce mutations which could impact the diagnostics performance of the RT-PCR . The L gene represents a much more conserved part of the EBOV genome than glycoprotein ( GP ) and nucleocapsid ( NP ) genes . One then can argue that mutations ( evolution of the virus during the outbreak ) would rather hamper diagnostic performance of RT-PCR targeting GP and NP than RT-PCR targeting L gene . In our recent work we demonstrated 100% correlation between RT-PCR targeting EBOV L gene and virus isolation . In addition we demonstrated that the performance of the L-gene based Taqman assay was highly comparable to an assay targeting two different genome targets [28] . Although nucleic acid sequencing may be performed to determine the molecular identity of an amplicon for confirmation , this capacity is usually not available in field-operated mobile laboratories . Due to overwhelming demand for rapid diagnosis of EVD suspected cases , and technical and staff limitations of the SA FEDL , serological testing could not be conducted . The same applied to most if not all mobile laboratories deployed in West Africa . As for most viral haemorrhagic fevers ( VHFs ) , the non-specific presentation , especially in early stages of infection , makes it difficult to diagnose clinically . Therefore , the differential diagnosis concerns a broad array of conditions ( e . g . malaria , rickettsial infections , Q fever , typhoid fever , dysentery , plague , brucellosis , leptospirosis , meningitis , other sepsis from bacterial infections , viral hepatitis , different VHFs , and non-infectious causes of disseminated intravascular coagulopathy ) especially during a yet unrecognised outbreak/causative agent . Emergency laboratory response cannot provide brood spectrum differential diagnosis for suspected VHFs . However , once the outbreak is recognised , the laboratory response is mostly focused on confirmation of cases . As for all VHFs diagnostic process has to consider all available laboratory results in the context of clinical , pathological and epidemiological data . The value of differential diagnosis was clearly demonstrated when we started testing EVD suspected cases for malaria infection towards the end of the outbreak; most of the cases turned out to be malaria positive and negative for EBOV infection . Design , biosafety and biocontainment equipment used by international field-deployed laboratories in Africa differ from laboratory to laboratory [29–31 , 36–37] . Currently there is no approved or recommended international standard for a mobile laboratory for the diagnosis of filovirus infections . Security , safety and health risks are always present in mobile missions , but can be mitigated by careful planning , mission preparation and team training [37] . However , field-deployed laboratories should not only enable rapid detection of EBOV or other highly dangerous pathogens , but their operation should only be permitted under high-security conditions with strict biosafety measures to ensure safe diagnosis of BSL-4 agents in developing and poorly- resourced countries . One of the biocontainment features of the SA FEDL that was highly appreciated and valued was the “hot laboratory” equipped with an IsoArk and a negatively pressurised glovebox that local staff accepted as maximised and a highly efficient physical protection against biohazardous clinical materials as well as protection of the environment . As a portable , modular compact unit with all the components fitting into one robust transportation container , the IsoArk can be safely moved to a place where it is needed . One of the advantages of the IsoArk is that it is specifically designed for rapid setup , allowing the conversion of any room or space into a biologically contained area for the isolation of infected or contaminated people and materials within a day upon arrival at a deployment site , and can be used for laboratory work with potentially highly infectious substances . The high standards of bio-safety and portability of the SA FEDL , offered a cost-effective and practical alternative for rapid deployment and operation of a high-level biocontainment facility at the scene of a formidable outbreak in a country that had little capacity for laboratory diagnosis of dangerous pathogens . The modular and portable biocontainment equipment used provided highly efficient primary and secondary biocontainment measures . Of the total of 25 operators who worked in the SA FEDL over 2 years none became infected with EBOV despite processing of large numbers highly infectious clinical specimens from EVD patients . The location of mobile laboratories in areas where EVD was spreading uncontrollably , significantly reduced the time between the collection of biological specimens and the return of results , thus making them much more effective than centralised reference laboratories located distantly from the “hot spots” of the outbreak . The shorter the delay in obtaining a laboratory result , the earlier the confirmed cases can be managed , thus facilitating the reduction in viral transmission [31 , 36–38] . The EVD epidemic in West Africa created an unprecedented challenge not only for the affected countries with weak health care systems and no previous experience with EVD , but also for the most experienced and resourceful countries , institutions and networks . It highlighted the need to maintain well organised laboratory systems and networks that can be effectively managed , including implementation of new diagnostic strategies and laboratory services in response to large-scale public health emergencies . The EVD crisis exposed systemic weaknesses and emphasised the need for better strategies to streamline the development and evaluation of new diagnostic platforms , transfer of material and specimens between countries and organisations , and more effective processes for the rapid deployment of health workers with specific laboratory expertise [39] . It also highlighted the necessity for essential reforms to better govern the global system for preventing and responding to formidable disease outbreaks [22 , 40] and building citizen trust in African health systems [23] . In conclusion , the bio-safety standards and the portability of the SA FEDL , offered a cost-effective and practical alternative for the rapid deployment and operation of a field-operated high biocontainment facility . The Western Urban and Western Rural Areas of Sierra Leone , where the SA FEDL operated , remained a “hotspot” of the EVD epidemic for several months and was the only Ebola diagnostic facility to respond to the overwhelming demand for EVD diagnosis for several weeks during the initial phases of the EVD crisis in the capital city , Freetown . Rapid molecular laboratory confirmation of EVD cases was crucial for the management of EVD cases , contact tracing and secure burial practices in Sierra Leone . The deployment of the SA FEDL capacity in this country also contributed to the overall international efforts in bringing the EVD outbreak in West Africa under control . This initiative became a nucleus of rolling deployments that produced much-needed medical specialists trained in the molecular diagnosis of EVD under exceptionally trying circumstances . It also constitutes the largest Ebola outbreak response in the history of the NICD and South Africa on foreign soil . The SA FEDL teams demonstrated that it is not only possible but highly beneficial to train the national staff in the course of formidable disease outbreak and accomplished their full integration into all operational and diagnostic aspects of the laboratory . The major advantages of incorporating national staff into the SA FEDL operation included not only important contributions to the EVD diagnostic service , but also the utilisation of their knowledge of local settings , including demographics , geography and custom . This facilitated the proper receiving of specimens , accurate collection of patient data , information capture , follow-up on missing or inaccurate specimen submission forms , communication with SL MoHS , the reporting of results , and the procurement of local goods and services . This culminated in successful hand over of the laboratory to the SL MoHS who subsequently effectively managed the laboratory and provided the required EVD diagnostic services until the country was declared free of EVD and during the post EVD outbreak enhanced surveillance .
|
In response to Ebola virus disease outbreak in West Africa , the National Institute for Communicable Diseases in South Africa established a modular high-biosafety field Ebola diagnostic laboratory ( FEDL ) near Freetown , Sierra Leone . This was the sole diagnostic capacity available to respond to the overwhelming demand for Ebola diagnosis for several weeks in the Western Area of Sierra Leone . The deployment of the FEDL capacity contributed to the overall international efforts in bringing the Ebola outbreak in West Africa under control .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[] |
2017
|
South African Ebola diagnostic response in Sierra Leone: A modular high biosafety field laboratory
|
Human African trypanosomiasis ( HAT ) is an important public health problem in sub-Saharan Africa , affecting hundreds of thousands of individuals . An urgent need exists for the discovery and development of new , safe , and effective drugs to treat HAT , as existing therapies suffer from poor safety profiles , difficult treatment regimens , limited effectiveness , and a high cost of goods . We have discovered and optimized a novel class of small-molecule boron-containing compounds , benzoxaboroles , to identify SCYX-7158 as an effective , safe and orally active treatment for HAT . A drug discovery project employing integrated biological screening , medicinal chemistry and pharmacokinetic characterization identified SCYX-7158 as an optimized analog , as it is active in vitro against relevant strains of Trypanosoma brucei , including T . b . rhodesiense and T . b . gambiense , is efficacious in both stage 1 and stage 2 murine HAT models and has physicochemical and in vitro absorption , distribution , metabolism , elimination and toxicology ( ADMET ) properties consistent with the compound being orally available , metabolically stable and CNS permeable . In a murine stage 2 study , SCYX-7158 is effective orally at doses as low as 12 . 5 mg/kg ( QD×7 days ) . In vivo pharmacokinetic characterization of SCYX-7158 demonstrates that the compound is highly bioavailable in rodents and non-human primates , has low intravenous plasma clearance and has a 24-h elimination half-life and a volume of distribution that indicate good tissue distribution . Most importantly , in rodents brain exposure of SCYX-7158 is high , with Cmax >10 µg/mL and AUC0–24 hr >100 µg*h/mL following a 25 mg/kg oral dose . Furthermore , SCYX-7158 readily distributes into cerebrospinal fluid to achieve therapeutically relevant concentrations in this compartment . The biological and pharmacokinetic properties of SCYX-7158 suggest that this compound will be efficacious and safe to treat stage 2 HAT . SCYX-7158 has been selected to enter preclinical studies , with expected progression to phase 1 clinical trials in 2011 .
Human African trypanosomiasis ( HAT ) , commonly known as sleeping sickness , is a neglected disease caused by the kinetoplastid parasite Trypanosoma brucei and is fatal if left untreated . [1] , [2] The parasite is transmitted through the bite of the tsetse fly , and is endemic in sub-Saharan Africa , where it is estimated that 50 , 000 people become infected every year . [3] , [4] , [5] The disease progresses through two distinct stages , an initial acute stage ( stage 1 ) where the parasitic infection is restricted to the hemolymphatic system , and a second stage ( stage 2 ) where the parasites have migrated across the blood-brain barrier and are resident in brain tissue . [6] This latter CNS stage is particularly difficult to treat , as the two drugs available for this purpose , melarsoprol and eflornithine , are toxic , have limited ability to cross the blood-brain barrier , and their activity is dependent upon complex parenteral administration procedures . [7] , [8] Treatment failures are therefore quite common with these drugs . There are currently no orally active treatments for HAT . A quite interesting recent advance in clinical treatment of HAT has been the development of a treatment regime which employs nifurtimox and eflornithine in combination . [9] , [10] , [11] , [12] , [13] This combination still suffers from some of the drawbacks associated with the individual components ( e . g . parenteral administration , toxicity , cost ) , but significantly reduces the complexity and duration of treatment in the clinical setting . Research efforts to discover new treatments for HAT have increased over the past several years , and have begun to deliver new biochemical targets and lead compounds . [14] , [15] We have previously reported that screening of a library of benzoxaboroles from Anacor Pharmaceuticals ( CA , USA ) in a whole cell T . brucei viability assay revealed that these compounds are effective inhibitors of parasite growth at concentrations as low as 0 . 02 µg/mL . [16] From this initial screening effort , benzoxaborole 6-carboxamides were identified as attractive leads , as they exhibited good in vitro potency , activity in stage 1 mouse models of HAT and promising in vitro and in vivo pharmacokinetic properties . [17] This program yielded SCYX-6759 ( see Figure 1 for chemical structures ) - the first compound with sufficient potency , pharmacokinetic properties and blood-brain barrier permeability to provide complete cures in a stage 2 murine model of HAT . While SCYX-6759 was fully efficacious in the stage 2 mouse model following twice-daily intraperitoneal administration at a dose of 50 mg/kg ( 100 mg/kg/day ) for 14 days , it exhibited only partial efficacy in the same model following twice-daily oral administration at 50 mg/kg for 7 days , and was not active at lower doses or when administered in a once-daily paradigm . This pharmacodynamic relationship was consistent with the in vivo pharmacokinetics of SCYX-6759 measured in mice , which demonstrated that , although this compound was well absorbed following oral administration and provided drug concentrations in plasma well in excess of the in vitro minimum inhibitory concentration ( MIC ) , drug concentrations in the brain fell below the MIC by approximately 12 h post-dose following either single oral doses or a 7-day repeat oral dosing regimen matched to the stage 2 efficacy model . Consequently , further optimization of the lead series focused on improvement of the pharmacokinetic properties of the series , with particular emphasis on improving the extent and duration of brain exposure . Of several strategies pursued to improve brain exposure of the benzoxaborole 6-carboxamides , the approach which provided the best balance of potency and pharmacokinetics involved the installation of substituents at the C ( 3 ) position of the benzoxaborole scaffold . In particular , the C ( 3 ) -dimethyl analog SCYX-7158 ( Figure 1 ) exhibited a profile supportive of progression to preclinical and clinical studies .
Animal studies were carried out in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health . All pharmacokinetic ( PK ) studies were performed by AAALAC accredited facilities following an internal approval by their IACUC boards . Pharmacokinetic studies in rodents , beagle dogs and cynomolgus monkeys were performed by Vivisource Laboratories Inc ( Waltham , MA; USDA# 14-R-0185 , OLAW# A4543-01 ) , Sinclair Research Center , LLC ( Columbia , MO; USDA# 43-R-0122 , OLAW# A4333-01 ) or SNBL USA ( Everett , WA; USDA# 91-R-0053 , OLAW# A4261-01 ) , respectively . Blood collection from mice was performed under anesthesia; brain collection was performed following euthanasia . Collection of serial blood samples from rats was performed from tail vein venipuncture , or via a venous access port . Efficacy studies were conducted in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health . The protocol was approved by the Institutional Animal Care and Use Committee of Pace University ( Animal Assurance Welfare Number: A3112-01 ) . Surviving animals are euthanized by carbon dioxide asphyxiation in sealed containers as approved by the American Veterinary Association . All efforts were made to minimize suffering; for example , in efficacy studies conducted at Pace University , animals were checked for parasitemia once per week and were immediately removed from cages and euthanized if parasites were found in a tail vein blood sample . SCYX-6759 [4-fluoro-N- ( 1-hydroxy-1 , 3-dihydrobenzo[c][1] , [2]oxaborol-6-yl ) -2-trifluoromethylbenzamide] and SCYX-7158 [4-fluoro-N- ( 1-hydroxy-3 , 3-dimethyl-1 , 3-dihydro-benzo[c][1] , [2]oxaborol-6-yl-2-trifluoromethyl benzamide] were prepared as 10 mg/mL stocks in dimethyl sulfoxide ( DMSO ) for in vitro biological and absorption , distribution , metabolism and excretion ( ADME ) assays . The bloodstream-form T . b . brucei 427 strain was obtained from Dr . K . Stuart ( Seattle Biomedical Research Institute , Seattle , WA ) and was used for routine assessment of compound sensitivity in vitro . Additional strains used in vitro include T . b . rhodesiense STIB900 , isolated in 1992 from a patient in Tanzania , and T . b . gambiense 108R , isolated from a patient in Democratic Republic of Congo in 2005 . T . b . brucei EATRO 110 was a generous gift from the late William Trager of the Rockefeller University . T . b . brucei TREU 667 strains were kindly provided by F . W . Jennings at the University of Glasgow . Both T . b . brucei EATRO 110 and TREU 667 were cultured according to previously described conditions and used for in vivo studies . [18] , [19] T . b . brucei was cultured in complete HMI-9 medium , [20] which contains 10% fetal bovine serum ( FBS ) ( Invitrogen , Carlsbad , CA ) , 10% Serum Plus medium ( SAFC Biosciences , Lenexa , KS ) , and 100 units/mL penicillin and 0 . 1 mg/mL streptomycin . The trypanosomes were propagated in T-25 vented cap flasks ( Corning Inc . , Lowell , MA ) at 37°C and 5% CO2 with humidity . To ensure log growth phase , trypanosomes were sub-cultured at appropriate dilutions ( typically 1∶100 ) every 2–3 days in fresh HMI-9 medium . T . b . rhodesiense parasites were grown in MEM medium with Earle's salts supplemented with 15% horse serum . For T . b . gambiense , the MEM medium was supplemented with 5% heat-inactivated fetal calf serum ( FCS ) and 15% human serum . [21] L929 mouse fibroblast cells ( ATCC CCL1 , American Type Culture Collection , Rockville , MD ) were used to determine parasite versus mammalian cell selectivity . Cells were cultured in Dulbecco's Modified Eagle Medium ( DMEM ) , supplemented with 10% FBS , L-glutamine and 100 units/mL penicillin and 0 . 1 mg/mL streptomycin . MDCKII-hMDR1 cells were generously provided by P . Borst at the Netherlands Kancer Institute . Cells were cultured in DMEM with Glutamax , 10% ( v/v ) FBS , and 100 units/mL penicillin and 0 . 1 mg/mL streptomycin . Cell monolayers were fed with cell culture media 24 h after seeding and used for permeability studies 3 days later . Compounds to be tested were serially diluted in DMSO and added to 96-well plates to give final concentrations ranging from 5 to 0 . 01 µg/mL . T . b . brucei parasites in the log phase of growth were diluted in HMI-9 media and added to each well for a final concentration of 1×104 parasites per well . For the sensitivity assays using T . b . rhodesiense and T . b . gambiense , pararasites were cultured in MEM supplemented with Baltz components ( 8 ) , diluted in the aforementioned culture media , and added to each well at a density of 1×103 cells/well . The final concentration of DMSO was 0 . 5% and the total volume was 100 µL/well . After a 72 h incubation , resazurin ( Sigma-Aldrich , St . Louis , MO ) was added to each well [20 µL of 25 mg/100 mL stock in phosphate buffered saline ( PBS ) ] and incubated for an additional 4–6 h . [22] To assess cell viability , fluorescence was quantified using an EnVision Multilabel Plate Reader ( Perkin Elmer , Waltham , MA ) at an excitation wavelength of 530 nm and emission of 590 nm . Triplicate data points were averaged to generate sigmoidal dose-response curves and determine IC50 values using XLfit curve fitting software from IDBS ( Guilford , UK ) . The IC50 is defined as the amount of compound required to decrease parasite or cell viability by 50% compared to those grown in the absence of the test compound . The MIC , defined as the lowest concentration of compound that completely inhibits visible parasite growth , was determined by visual inspection of 96-well plates after 48–72 h of incubation with the test compounds . To evaluate the effects of serum on trypanocidal activity , assays were performed in the presence of increasing concentration ( 2 . 5% to 50% ) of fetal calf serum ( Invitrogen , Carlsbad , CA ) . The results were expressed as a fold-change in IC50 values relative to standard conditions ( 10% FCS ) . An evaluation of mammalian cell cytotoxicity was carried out in parallel with the trypanosome sensitivity assays . L929 mouse fibroblast cells were seeded at 2×103 per well and handled as described above for the trypanosome sensitivity assay . The assessment of oxaborole-mediated killing of T . b . brucei in vitro was conducted using the CellTiter Glo kit ( Promega , Inc . , Madison , WI ) to measure trypanosome ATP content as a real-time indicator of viability . Test compounds were serially diluted from 5 to 0 . 01 µg/mL into white wall-clear bottom 96-well plates ( Corning Inc . , Lowell , MA ) containing HMI-9 media . 1×104 trypanosomes were added to each well . At specified intervals , the CellTiter Glo reagent was added to lyse the parasites and the plates were incubated for 10 min in the dark . Luminescence was quantified using an EnVision plate reader . All determinations were done in duplicate . Time-kill parameters were determined from plots of parasite viability versus incubation time for each concentration tested . To establish the time and concentration required to cause persistent or irreversible effects by oxaboroles , T . b . brucei parasites were assessed for their ability to recover from transient exposure to test compounds . Trypanosomes were seeded in clear 96-well V-bottom plates at a density of 1×105 parasites per well and incubated with serially diluted test compound ( from 10 to 0 . 02 µg/mL ) . One plate was prepared for each time point . At the designated time , a plate was removed from the incubator and spun at 2 , 600 g for 5 min to sediment the parasites . The supernatant was aspirated and 100 µL of warmed HMI-9 media was added to each well . The wash was repeated twice more . The parasites were resuspended in 100 µL of warmed media and 20 µL of this suspension was added to 80 µL of HMI-9 media in triplicate plates . Following a 72 h incubation , resazurin was added and trypanocidal activity determined as described for the in vitro sensitivity assay . For efficacy studies against acute infections , groups of 3–5 female Swiss Webster mice ( Ace Animals , Boyertown , PA ) were injected intraperitoneally ( i . p . ) with freshly drawn infected rat blood containing 2 . 5×105 trypanosomes ( T . b . brucei EATRO 110 strain ) . The infection was allowed to progress for 24 h before treatment . Test compound was then given daily by bolus i . p . injection or oral gavage . Oxaborole compounds were formulated in 2% ethanol/5% dextrose and were given at a dose volume of 200 µL ( containing 1 . 25 to 50 mg/kg ) for a 25 g animal . Animals were monitored daily for signs of compound toxicity and clinical signs of trypanosomiasis for a period of 30 days . Mice were checked for parasitemia once per week by microscopic examination of smears prepared from the tail vein blood of animals . Animals remaining parasite free for more than 30 days beyond the end of the treatment period were considered cured . Control untreated animals typically succumbed to the infection within 4–5 days following i . p . inoculation with parasites . For evaluation against late-stage CNS infections , a chronic disease inducing strain ( TREU 667 ) of T . b . brucei was used . [19] Mice were divided into groups of ten and each animal was infected i . p . with 1×104 parasites ( 200 µL ) . As a positive control , a group of mice were treated with a single 10 mg/kg dose of berenil ( Sigma-Aldrich , St . Louis , MO ) administered by i . p . injection on day 4 after infection . Infection in the remaining groups of animals was allowed to proceed for 21 days before treatment with either berenil ( single i . p . dose at 10 mg/kg ) or test compound given i . p . or orally twice daily for 7 or 14 days . Animals were checked for parasitemia once per week and were immediately removed from the cages and euthanized upon recrudescence . Animals were considered to be cured of a CNS infection if they were aparasitemic for at least 180 days after the end of the treatment period . Additionally , brain homogenates or blood of oxaborole-cured animals failed to generate infection when injected into mice immunosuppressed with cytoxan ( 40 mg/kg , QD x 2 d ) . Binding to human or mouse plasma proteins or mouse brain homogenate was determined by rapid equilibrium dialysis ( RED ) ( Pierce , Rockford , IL ) using a 48-well plate-based format according to the manufacturer's instructions . Briefly , test compound at the required concentrations was added to fresh human or mouse plasma ( Bioreclamation , Liverpool , NY ) or freshly prepared mouse brain homogenate . Duplicate aliquots of each sample were transferred into the sample chambers of the RED devices , and dialysis buffer ( BupH PBS ) was added to the buffer chambers . The plates were sealed and incubated at 37°C for 4 h . After dialysis , samples collected from the buffer and tissue chambers were treated with ice-cold methanol ( 3 volumes for plasma , 4 volumes for brain ) to precipitate proteins . The treated samples were centrifuged for 10 min at 3×103 g at 15°C . The supernatants were assayed for test compound by LC-MS/MS . Calibration standards and quality control samples were prepared in matched matrix and assayed with samples . Values for unbound and bound fractions and mass balance were calculated . Concordance of binding for each batch of plasma was confirmed by assay of warfarin , imipramine and carbamezapine . Acceptance criterion for mass balance was 70–120% . Metabolic stability was evaluated using mixed gender CD-1 mouse or human liver microsomal fractions ( XenoTech , Lenexa , KS ) . Compounds ( 1 µM ) were incubated with microsomes ( 1 . 05 mg/mL protein ) for 0 , 10 , 15 and 30 min at 37°C in an oxygen and humidity enriched environment in the presence of an NADPH-regenerating system . Each compound was also incubated , under the same conditions , with CD-1 mouse , Sprague-Dawley rat , beagle dog , cynomolgus monkey , and human liver S9 sub-cellular fractions ( 2 mg/mL protein ) [23] for 0 , 15 , 30 and 60 min . At the end of each incubation period , the reactions were quenched with 3 volumes of ice-cold methanol . Supernatants from the incubation mixtures were analyzed for parent compound by LC-MS/MS . Metabolic competencies of microsomal and S9 fractions were confirmed using control compounds 7-ethoxycoumarin , propranolol and verapamil . Intrinsic clearance ( CLint ) and half-life ( t½ ) values were determined for each compound . The propensity of the compound to cross the blood-brain barrier was examined using an in vitro MDCKII-hMDR1 transwell assay . [24] MDCKII-hMDR1 cells were seeded at density of 3×105 cells per well onto microporous polycarbonate membranes in 12 well Costar Transwell plates ( Corning Inc . , Lowell , MA ) . The cells were used for permeability studies 3 days later . Trans-epithelial resistance ( TEER ) was measured for each insert to ensure the integrity of the monolayer ( acceptable TEER >50 Ωcm2 ) . The permeability and propensity for P-gp-mediated efflux was evaluated by adding each compound at a concentration of 3 µM , in the presence or absence of 2 µM GF120918 , to the apical compartment . Competency of the P-gp efflux transporter was confirmed by assay of propranolol ( non-substrate ) and amprenavir ( substrate ) . Cell monolayers were incubated in triplicate with shaking ( 160 rpm ) at 37°C in a 5% CO2-enriched humidified atmosphere for 1 h . Samples were removed from the apical and basolateral compartments after incubation and assayed for test compound concentrations by LC-MS/MS . Values for mass balance , PappA→B , PappA→B+GF918 , and absorption quotient ( AQ ) were calculated for each compound . [25] , [26] , [27] Acceptance criterion for mass balance was 70–120% . The stability of each oxaborole in the CNS tissue was evaluated by incubation in fresh rodent brain homogenate . SCYX-7158 was incubated in freshly prepared homogenates of mouse brain tissue for 5 h at 37°C with gentle shaking . After incubation , the reactions were quenched with 4 volumes of ice-cold methanol . Supernatants from the incubation mixtures were analyzed for parent compound by LC-MS/MS . The potential for compound to exert drug-drug interactions that are mediated through inhibition of cytochrome P450 activities was assessed using P450-Glo assay kits ( Promega Inc . , Madison , WI ) . Assays for human cytochrome P450 isoforms 3A4 , 1A2 , 2C9 , 2C19 , and 2D6 were performed in triplicate according to the manufacturer's instructions over the concentration range 1–100 µM ( n = 6 levels ) . Briefly , SCYX-7158 was added to membrane preparations containing human recombinant CYP450 enzymes together with luminogenic substrates specific to each isoform . Specific CYP450 inhibitors were included as positive controls ( ketoconazole , CYP3A4; alpha-napthoflavone , 1A2; sulfaphenazole , CYP2C9; troglitazone , 2C19 , and quinidine , CYP2D6 ) . Reactions were initiated by addition of NADPH-regenerating solution and were allowed to incubate for an additional 15–30 min ( isoform dependent ) . Luciferin detection reagent , containing the luciferin-labeled probe substrate , was then added . Luminescence was recorded on an EnVision Multilabel Plate Reader . Calculation of IC50 concentrations for each enzyme was determined using GraphPad Prism ( V5 . 01 ) . The potential for drug-drug interactions was characterized as high ( IC50 <1 µM ) , moderate ( 1 – 10 µM ) , or low ( IC50 >10 µM ) . [28] The blood distribution , CNS disposition and pharmacokinetics of SCYX-7158 were evaluated in infected or non-infected rodents following intravenous or oral administration of compounds . Pharmacokinetics and bioavailability of SCYX-7158 were also evaluated in non-infected beagle dogs and cynomolgus monkeys . In-life phases of non-infected rodent studies were performed at Vivisource ( Waltham , MA ) ; studies in non-naïve cynomolgus monkeys were performed by SNBL USA ( Everett , WA ) . Male CD-1 mice ( ∼25 g ) , male Sprague-Dawley rats ( ∼225 g ) , or male cynomolgus monkeys ( ∼3–5 kg ) were administered test article by either bolus intravenous injection ( IV ) or oral gavage . Animals in the IV groups received a single 2 mg/kg IV dose . Animals received oral doses of test articles as either single or twice daily doses ranging from 8 mg/kg to 50 mg/kg . All doses were administered as clear colorless solutions in either 50% ( v/v ) PEG 400: 20% ( v/v ) ethanol: 30% ( v/v ) carboxymethylcellulose ( 0 . 5% w/v in sterile water for injection ) or 2% ( v/v ) ethanol: 5% ( w/v ) dextrose in sterile water for injection . Doses were administered in a volume of 4 mL/kg , 2 mL/kg or 1 mL/kg for mice , rats and cynomolgus monkeys , respectively . For pharmacokinetic analysis , blood samples were collected from mice via cardiac puncture under terminal anaesthesia . Serial blood samples were collected from rats via a vascular access port located in the lateral saphenous vein . Mice and rats were euthanized in a CO2 chamber before collection of terminal blood samples or brain tissue . Blood samples were collected into polypropylene tubes containing K2EDTA anticoagulant and stored on ice until centrifuged for the preparation of plasma . Plasma was stored at −70°C . Whole brains were collected following decapitation , blotted dry , placed into polypropylene containers and then immediately frozen at approximately −70°C . CSF was collected from rats via cisterna magna puncture , placed in sterile polypropylene tubes , and stored at approximately −70°C . Samples of plasma ( 25 µL ) and CSF ( 10 µL ) were treated with 3 volumes of ice-cold methanol ( containing 25 ng/mL of 2-chloro-4-fluoro-N- ( 1-hydroxy-1 , 3-dihyrobenzo[c][1] , [2]oxaborol-6-yl ) benzamide or 25 ng/mL d6-SCYX-7158 as an internal standard ) to precipitate proteins . Treated samples were gently mixed at room temperature for 10 min , and then centrifuged at approximately 3000×g for 15 min at 15°C . The supernatants were transferred to 96-well plates or HPLC vials for analysis by LC-MS/MS . Pharmacokinetic parameters were calculated from composite mean plasma or tissue data using non-compartmental ( oral and intravenous routes ) and bi-exponential ( intravenous route ) analyses in Microsoft Excel .
In whole cell assays , SCYX-7158 exhibited potent activity against representative T . b . brucei , T . b . rhodesiense and T . b . gambiense strains . Parasite-mediated reduction of the pro-fluorescent dye resazurin was used as indicator for trypanosome viability . As shown in Table 1 , IC50 values for SCYX-7158 were approximately 0 . 07 µg/mL to 0 . 37 µg/mL following incubation of the parasite strains with the compound for 72 h . In the T . b . brucei S427 strain , the MIC value for SCYX-7158 was 0 . 6 µg/mL , approximately two times the IC50 measured for this strain . In the T . b . brucei S427 assay , the primary oxidative metabolite SCYX-3109 ( vide infra ) was inactive , as it exhibited no inhibition of parasite growth at a concentration of 10 µM . In contrast to the potent activity of SCYX-7158 against trypanosomes , no significant inhibition of cell proliferation was observed in an in vitro mammalian cell ( L929 mouse cell line ) assay at drug concentrations up to 50 µg/mL . We next evaluated the in vitro time-kill relationship for SCYX-7158 . [17] , [29] , [30] In these experiments , trypanosomes were exposed to continuous drug pressure and survival of the parasites was measured at several time points over 24 h . The results of this experiment , which measured parasite ATP content as an indicator of viability , are presented in Figure 2A . SCYX-7158 displayed concentration-dependent trypanocidal activity characterized by rapid onset , with greater than 50% reduction in viability within 8 h of exposure at a concentration ( 1 . 25 µg/mL ) about 2 times the MIC ( 0 . 6 µg/mL ) . At these concentrations , >99% of the parasites were killed within 24 h of exposure to the compound . The final aspect of activity of SCYX-7158 that was evaluated in vitro was the irreversibility of the trypanocidal effect . In these experiments ( Figure 2B ) , we demonstrated that a short exposure ( 10–12 h ) to the compound was sufficient to produce irreversible effects on trypanosome survival , albeit at a concentration of about 5 times the IC50 . As was observed in the time-kill experiments , increasing the concentration of compound above this threshold value did not significantly alter the robustness or speed of its trypanocidal activity . Our initial in vivo experiments were conducted in a mouse model of acute trypanosomiasis . In this model , mice were infected with 2 . 5×105 parasites of the T . b . brucei EATRO 110 strain . This strain produces a robust infection in mice with rapid increases in parasitemia , ultimately leading to death of untreated mice within 4–5 days following infection . [18] Starting 24 h after infection , mice were administered an oral dose of SCYX-7158 once daily for 4 days . Mice treated with doses as low as 5 mg/kg/day ( administered as either 5 mg/kg once daily ( QD ) or 2 . 5 mg/kg twice daily ( BID ) ) exhibited a 100% cure rate ( Table 2 ) . Animals were monitored weekly for parasitemia through 30 days after infection . Those animals that remained parasite-free through the entire course of the experiment were considered cured . [18] As the in vitro time-kill and irreversibility assays suggested that complete parasite clearance could be obtained within 12–24 h , we examined whether a single high dose of SCYX-7158 could produce a similar effect in vivo . In this experiment , a single 25 mg/kg i . p . dose of SCYX-7158 produced 100% cure by the criteria described above . Lower doses by the i . p . route , or oral administration of SCYX-7158 at 25 mg/kg , while not fully efficacious , increased the survival time of mice compared to untreated controls ( Table 2 ) . The most important need for a clinically relevant HAT drug is the ability to cross the blood-brain barrier and kill parasites in the brain of the patient . In order to assess the potential of SCYX-7158 to address this need , a model of the stage 2 CNS HAT was employed . [31] In this model , mice are infected with 1×104 parasites of the T . b . brucei TREU 667 strain , which produces a persistent hemolymphatic infection where parasites migrate across the blood-brain barrier within 21 days after infection . [19] Two control groups were included in each study – a positive control group in which animals were treated with berenil at day 4 after infection , and a negative control group in which animals were treated with berenil at day 21 after infection . Berenil does not cross the blood-brain barrier , so although it is very effective in clearing parasites from the hemolymphatic system , it is unable to clear parasites from the CNS . Consequently , animals treated on day 4 are cured as parasites have not yet crossed the blood-brain barrier; however , animals receiving berenil 21 days after infection do not sustain cures and recrudesce because parasites that have infected the CNS are able to migrate back across the blood-brain barrier and re-infect the hemolymphatic system generally around days 35–42 after infection . [32] We examined SCYX-7158 at several doses in this model of Stage 2 HAT . When administered as a once daily oral dose of 12 . 5 mg/kg over 7 days starting on day 21 after infection , SCYX-7158 produced an 80% cure rate of the T . b . brucei infection , with a 100% cure rate observed following 7 daily oral doses of 25 mg/kg ( Figure 3 ) . In this model , cure is defined as lack of parasitemia in the blood for 180 days after the last dose , together with a lack of infection measured throughout 30 days in fresh animals inoculated with the blood and/or brain homogenates collected from the SCYX-7158 treated animals . Concurrent with our studies to explore the biological activity of SCYX-7158 , we conducted extensive in vitro ADMET experiments to help understand the ability of this compound to be effective in the treatment of stage 2 HAT ( Table 3 ) . SCYX-7158 was metabolically stable when incubated with liver sub-cellular fractions from rodents , beagle dog , cynomolgus monkey and humans . Values for intrinsic clearance and half-life with all species were less than 5 µL/min/mg protein and longer than 350 min following incubation with microsomes and S9 fractions , respectively . When incubated with primary hepatocytes from rat or dog for extended periods ( 6–12 h ) , we observed the formation of two minor metabolites . Phase 1 oxidative deboronation of SCYX-7158 yielded the diol SCYX-3109 ( Figure 1 ) as the primary metabolite; although less than 5% of the SCYX-7158 was metabolized over 12 h . Subsequent glucuronidation of SCYX-3109 was observed to a small degree . In vitro SCYX-7158 showed low clearance predictive of good exposure during in life studies . The potential for SCYX-7158 to inhibit cytochrome P450 enzymes was evaluated using P450-Glo assays ( Promega ) for the human isoforms 3A4 , 1A2 , 2C19 , 2C9 and 2D6 . The IC50 values for SCYX-7158 in these assays were all above 10 µM , suggesting that this compound has low risk for CYP450-based drug-drug interactions . [28] We also evaluated the primary oxidative metabolite , SCYX-3109 , in the CYP inhibition assays , where it exhibited IC50 values above 100 µM for the human isoforms 1A2 , 2C19 , 2C9 and 2D6 . An IC50 of 7 . 5 µM for inhibition of human CYP 3A4 was measured . The mechanism by which SCYX-7158 is trypanocidal is currently unknown . In order to assess potential biochemical targets ( or target classes ) that might be implicated in the mechanism of action of SCYX-7158 , and to concurrently identify any potential off-target activities of SCYX-7158 that may hold potential for toxicity to the mammalian host , we evaluated SCYX-7158 in an array of in vitro receptor binding and enzyme inhibition assays . At a test concentration of 10 µM , SCYX-7158 did not exhibit any significant binding to , or inhibition of , any of the >100 biochemical targets tested . While these results did not identify the potential biochemical targets , we were encouraged because they predict a low risk of mechanism-based toxicity in our animal safety studies , which is also supported by the observation that SCYX-7158 was well tolerated by mice at doses up to 100 mg/kg BID ( e . g . 200 mg/kg/day , 8 times the effective dose ) which represents the highest dose administered during efficacy studies . This dose was based on practical considerations ( e . g . solubility of compound in vehicle and dosing volume limitations ) . By way of comparison , in the closely related GVR35 stage 2 HAT model , melarsoprol must be dosed at 10–15 mg/kg QD × 5 days to effect cure , but is toxic to mice at 20 mg/kg QD x 5 days ( R . Brun , personal communication ) . Preliminary toxicological evaluation of SCYX-7158 in rats and dogs is ongoing , and will be reported as part of a regulatory filing prior to initiation of clinical trials . Due to implication of the hERG channel in potential cardiovascular toxicity , specifically long QT syndrome , [33] we evaluated SCYX-7158 in several hERG potassium channel assays . SCYX-7158 did not exhibit any significant binding to this channel in a radioligand binding assay ( −6% at 10 µM ) , and we confirmed this result by investigation of the ability of SCYX-7158 to block the hERG channel expressed in HEK-293 cells using a whole-cell patch clamp technique . [34] , [35] At test concentrations of 30 and 100 µM , SCYX-7158 produced a mean fractional block of 0 . 108±0 . 037 and 0 . 198±0 . 018 , respectively , indicative of a hERG IC50 >100 µM . The final in vitro toxicology study performed on SCYX-7158 was a bacterial reverse mutation ( Ames ) assay . In this assay , the potential for mutagenicity of SCYX-7158 was evaluated by exposure of several test strains of Salmonella typhimurium and Escherichia coli , both in the presence and absence of rat liver S9 fraction . [36] At a maximum dose of 5000 µg per plate , no positive mutagenic response was observed in any of the test strains employed , leading to the conclusion that SCYX-7158 is classified as an Ames-negative compound . Binding of SCYX-7158 to human and mouse plasma proteins was determined by rapid equilibrium dialysis ( RED , Pierce ) . Binding to plasma proteins was concentration dependent where the unbound fraction ( fu ) of SCYX-7158 in mouse plasma at the MIC ( ∼0 . 6 µg/mL ) was 0 . 3% rising to 3 . 2% at plasma concentrations equivalent to Cmax at steady-state ( ∼15 µg/mL , 25 mg/kg doses ) . Protein binding was modestly weaker in human plasma where unbound fraction was 1 . 3% and 5 . 5% with 1 µg/mL and 50 µg/mL SCYX-7158; the corresponding values for mouse plasma were 0 . 32% and 4 . 63% , respectively . Binding to mouse brain tissue was independent of concentration ( fubrain ∼5% ) . In these experiments , SCYX-7158 was also stable when incubated with plasma or brain homogenate for 4 h at 37°C . To assess the possible impact of protein binding on potency , mouse or bovine serum was added to the in vitro T . b . brucei inhibition assay . The resulting in vitro IC50 was attenuated by less than 3–4 fold in the presence of 25% mouse serum or 50% bovine serum . These results suggest low affinity for protein binding ( data not shown ) . In contrast , the in vitro potency of suramin , the current standard of care for stage 1 HAT ( hemolymphatic stage ) , when tested under the same conditions , was attenuated by greater than 25 fold . Successful treatment of stage 2 HAT requires therapeutically relevant exposure in the CNS compartment . To achieve this , SCYX-7158 needed to readily cross the blood-brain barrier and not be a substrate for the P-glycoprotein ( P-gp ) efflux transporter . An early prediction of the ability of SCYX-7158 to cross the blood-brain barrier was obtained by the assessment of the permeability in the MDCKII-hMDR1 monolayer transport assay . [24] In this assay , MDCKII cells over-expressing P-gp were grown to confluence on transwell membranes and incubated with SCYX-7158 ( 1 . 1 µg/mL ) . Apparent permeability in the apical to basolateral direction ( PappA→B ) was 776 nm/s , consistent with high blood-brain barrier permeability , where PappA→B values >150 nm/s are considered indicative of CNS exposure . [37] Furthermore , PappA→B values >50 nm/s predict rapid and complete absorption following oral administration . [26] When the permeability assay was performed in the presence of the known P-gp inhibitor GF120918 , the PappA→B+918 value was similar at 853 nm/s . The absorption quotient ( AQ ) , calculated from these two values ( as described in the Methods section ) was 0 . 09 , indicating that SCYX-7158 is not a P-gp substrate . [25] Overall , the high PappA→B value and low AQ value predict that SCYX-7158 should readily cross the blood-brain barrier and , in light of good plasma pharmacokinetics , achieve therapeutically relevant exposures in brain tissue following oral dosing . Given the attractive in vitro ADME profile of SCYX-7158 , we were confident that the compound would exhibit good in vivo pharmacokinetics . This expectation was confirmed in mice , rats , and monkeys following both intravenous and oral administration ( Figure 4 ) . In uninfected mice , 4 . 3 mg/kg intravenous dose of SCYX-7158 showed an apparent elimination half-life ( t1/2 ) of 26 . 6 h; systemic clearance ( CL ) of 0 . 089 L/h/kg; a volume of distribution ( Vdss ) of 1 . 69 L/kg and area under the concentration-time curve ( AUC0–24 h ) of 48 h•µg/mL . Following an oral dose of 13 . 4 mg/kg , which corresponds to the lowest efficacious dose in the murine stage 2 HAT model , SCYX-7158 was rapidly absorbed , as a Cmax of 6 . 96 µg/mL was achieved in plasma at 6 h after dose , with an oral clearance ( Cl/F ) value of 0 . 163 L/h/kg , an AUC0–24 h of 82 h•µg/mL and absolute oral bioavailability of 55% . After a 26 mg/kg oral dose , which corresponds to the dose giving a 100% cure rate in the murine stage 2 HAT model , Cmax increased to 9 . 8 µg/mL and the AUC0–24 h was 113 h•µg/mL . CNS exposure was determined in homogenates prepared from whole brain tissues . SCYX-7158 exhibited good permeability across the blood-brain barrier and achieved measurable levels after both intravenous and oral doses . Following intravenous administration at 4 . 3 mg/kg , SCYX-7158 demonstrated a brain AUC0–24 h of 5 . 57 h•µg/mL . In the study where the compound was dosed orally at 13 . 4 mg/kg or 26 mg/kg , the values for Cmax in brain tissue were 4 . 62 µg/mL and 8 . 1 µg/mL , respectively . The corresponding values for AUC0–24 h were 31 . 4 h•µg/mL and 68 h•µg/mL . At these therapeutic doses , the brain to plasma ratios of SCYX-7158 were 38% and 60% based on AUC0–24 h following the 13 . 4 mg/kg and 26 mg/kg oral doses , confirming that SCYX-7158 readily enters the CNS compartment and that disposition occurs in a dose-dependent manner . In uninfected rats , following oral administration of SCYX-7158 at a nominal dose of 25 mg/kg ( dose affording a 100% cure rate in mice ) , Cmax increased approximately 2 fold more than that in mice ( Cmax = 18 . 2 µg/mL ) and AUC0–24 h , and hence oral clearance , improved approximately 4 fold ( AUC0–24 h 291 h•µg/mL and CL/F = 0 . 092 L/kg/h ) . The time to maximum concentration was similar to that in mice ( tmax = 8 h ) . . In the same experiment , brain and cerebrospinal fluid ( CSF ) pharmacokinetics parameters were as follows: Cmax = 8 . 5 µg/mL in the brain , 1 . 08 µg/mL in the CSF and AUC0–24 h = 129 h•µg/mL in the brain and 14 . 6 h•µg/mL in the CSF . The brain to plasma AUC0–24 h ratio was 44% , which is similar to that in mice . The pharmacokinetic properties of SCYX-7158 were also evaluated in non-human primates following administration by intravenous and nasogastric ( NG ) routes . In this study , uninfected male and female cynomolgus monkeys were treated with SCYX-7158 at 2 mg/kg ( IV ) on study day 1 and 10 mg/kg ( NG ) on study day 8 . Following each dose , blood samples were taken at 11 time points between 0 . 17 and 72 h , and CSF was collected at 2 , 6 , 12 , 18 and 24 h in an off-set sparse sampling paradigm . As observed in both the mouse and rat studies , SCYX-7158 exhibited excellent plasma pharmacokinetics , with CL of 0 . 022 L/h/kg; Vdss of 0 . 656 L/kg and area under the concentration-time curve 78 . 8 h•µg/mL , and 94 . 4 for AUC0–24 h and AUC0–inf , respectively , following intravenous administration . Pharmacokinetic properties were independent of gender . In the oral phase of the study , SCYX-7158 exhibited a Cmax of 11 µg/mL at 9 . 5 h after dose , an oral clearance ( Cl/F ) value of 0 . 025 L/h/kg , an AUC0–inf of 460 h•µg/mL , corresponding to absolute oral bioavailability of 89% . Concentrations of SCYX-7158 in the CSF of non-human primates were approximately 5% of the plasma concentration at each time point . In this study , we also measured concentrations of the oxidative metabolite SCYX-3109 in plasma . In the 10 mg/kg NG group , the concentration of SCYX-3109 was approximately 1% of the SCYX-7158 concentration at all time points ( e . g . Cmax = 0 . 14 µg/mL and AUC0-inf = 6 . 4 h•µg/mL ) . Though not measured due to technical limitations , we can estimate that the maximum concentration of boric acid ( borate ) would be no greater than 0 . 02 µg/mL , as it is generated on an equimolar basis to SCYX-3109 . This concentration is below the background concentration ( 0 . 03 – 0 . 10 µg/g ) reported for boric acid in the human population , and well below the concentration ( 1 . 27 µg/g ) reported as a NOAEL in rats . [38] , [39] We have also measured plasma concentrations of the oxidative metabolite SCYX-3109 in separate studies in mice and rats . In mice , following a oral dose of 56 mg/kg SCYX-7158 , the plasma concentrations SCYX-3109 were approximately 1% of the SCYX-7158 concentration at all time points , where the Cmax for SCYX-7158 was 18 µg/mL and for SCYX-3109 was 0 . 27 µg/mL . Rats were administered an oral dose of 10 mg/kg of SCYX-7158 , and concentrations of SCYX-3109 were 3-5% of the SCYX-7158 concentration at each time point . The Cmax for SCYX-7158 was 14 . 3 µg/mL and for SCYX-3109 was 0 . 5 µg/mL . As in the non-human primate studies , boric acid concentrations were not measured in these studies , but the estimated maximum concentration could be calculated as 0 . 05 µg/mL and 0 . 09 µg/mL in mice and rats , respectively . As with the non-human primate studies , these concentrations of boric acid are well below the NOAEL in rats . [38] , [39] Steady-state pharmacokinetic properties of SCYX-7158 were determined in T . brucei-infected mice after 7 daily oral doses of 6 , 12 . 5 , 25 , 50 or 100 mg/kg . SCYX-7158 showed dose-related increases in exposure , although exposure increased in a less than proportional manner ( Figure 5 ) . Brain exposure was impressive in these mice as well . For example , following a 12 . 5 mg/kg oral dose , a Cmax of 6 . 54 µg/mL and an AUC0-inf of 34 . 2 h•µg/mL were achieved in mouse brain tissue; at 25 mg/kg , Cmax of 11 . 98 µg/mL and an AUC0-inf of 81 . 5 h•µg/mL were observed . Most importantly , brain concentrations of SCYX-7158 were maintained at or above the MIC ( 0 . 6 µg/mL ) for close to 24 h following the 25 mg/kg dose , consistent with the efficacy observed in the stage 2 HAT assay . Interestingly , efficacy correlated with total brain exposure rather than the unbound concentration . Consistent with minimal attenuation of in vitro potency in the presence of serum , this likely reflects weak , non-restrictive , binding of SCYX-7158 to brain tissue complemented with rapid distribution into trypanosomes . The pharmacokinetic properties and efficacy of SCYX-7158 in the stage 2 murine HAT model have demonstrated that a once-daily oral regimen is possible during clinical application . Efficacy in the stage 2 model was dependent on maintaining brain concentrations above the MIC for approximately 24 h or achieving brain concentrations above 3 times the MIC for shorter periods of time , as depicted in Figures 2B and 5 . If pharmacokinetic clearance scales from pre-clinical species to humans based on body weight or hepatic blood flow are likely , clinical dose would be approximately 2 . 5 mg/kg , although this must await confirmation in clinical trials . These observations along with in vivo pharmacodynamic relations observed in the mouse stage 2 HAT experiments , where maintenance of drug concentration at or above the trypanocidal MIC for 14–20 h was sufficient to demonstrate cures in this model , suggest that SCYX-7158 could be a once-daily oral treatment for HAT . The optimization of a series of benzoxaboroles discovered to exhibit parasitical activity against T . brucei has culminated in the identification of SCYX-7158 . We have demonstrated that this compound is potent in in vitro trypanocidal assays and has attractive in vitro physicochemical and ADME properties . In animal models of HAT , SCYX-7158 exhibits significant activity following oral administration , including cure of a CNS T . brucei infection following 7 days administration at a dose of 25 mg/kg . The in vivo pharmacokinetic characterization of SCYX-7158 reveals that this compound is highly bioavailable across species , and can cross the blood-brain barrier to achieve therapeutically-relevant concentrations in the brain and cerebrospinal fluid of rodents . Based on these properties , SCYX-7158 has been selected to enter IND-enabling preclinical studies , with expected progression to phase 1 clinical trials in 2011 .
|
Human African trypanosomiasis ( HAT ) is caused by infection with the parasite Trypanosoma brucei and is an important public health problem in sub-Saharan Africa . New , safe , and effective drugs are urgently needed to treat HAT , particularly stage 2 disease where the parasite infects the brain . Existing therapies for HAT have poor safety profiles , difficult treatment regimens , limited effectiveness , and a high cost of goods . Through an integrated drug discovery project , we have discovered and optimized a novel class of boron-containing small molecules , benzoxaboroles , to deliver SCYX-7158 , an orally active preclinical drug candidate . SCYX-7158 cured mice infected with T . brucei , both in the blood and in the brain . Extensive pharmacokinetic characterization of SCYX-7158 in rodents and non-human primates supports the potential of this drug candidate for progression to IND-enabling studies in advance of clinical trials for stage 2 HAT .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"and",
"Discussion"
] |
[
"medicinal",
"chemistry",
"chemical",
"biology",
"chemistry",
"biology",
"microbiology",
"parasitology",
"parasite",
"physiology"
] |
2011
|
SCYX-7158, an Orally-Active Benzoxaborole for the Treatment of Stage 2 Human African Trypanosomiasis
|
Severe malaria occurs predominantly in young children and immunity to clinical disease is associated with cumulative exposure in holoendemic settings . The relative contribution of immunity against various stages of the parasite life cycle that results in controlling infection and limiting disease is not well understood . Here we analyse the dynamics of Plasmodium falciparum malaria infection after treatment in a cohort of 197 healthy study participants of different ages in order to model naturally acquired immunity . We find that both delayed time-to-infection and reductions in asymptomatic parasitaemias in older age groups can be explained by immunity that reduces the growth of blood stage as opposed to liver stage parasites . We found that this mechanism would require at least two components – a rapidly acting strain-specific component , as well as a slowly acquired cross-reactive or general immunity to all strains . Analysis and modelling of malaria infection dynamics and naturally acquired immunity with age provides important insights into what mechanisms of immune control may be harnessed by malaria vaccine strategists .
Plasmodium falciparum ( Pf ) malaria in holoendemic areas is characterized by high level parasitaemia and symptomatic infections in early childhood , followed by the development of semi-protective immunity that allows the persistence of low level asymptomatic infections and appears to reduce the likelihood of becoming infected if bitten by an infective mosquito . The mechanisms mediating anti-infection and anti-disease immunity are complex but are thought to include innate and adaptive immune responses that limit both the liver and blood stage of the parasite life cycle in the human host [1] , [2] . One approach to understanding acquired immunity to Pf-malaria has been to study correlates of protection by measuring point-prevalence levels of immunity and prospectively assessing the infection status and clinical disease . A number of studies have elaborated upon this approach by first treating patients with anti-malarial drugs to eliminate blood-stage malaria infection , and then observing the time to natural ( re ) -infection in an endemic setting [3] . By measuring immune parameters at baseline , and observing their association with time-to-infection , it may be possible to identify the immune responses most important for protection from malaria . A significant difficulty with these studies is that most immune responses to malaria increase with age and after cumulative exposure to malaria antigens , and so it is often unclear whether measured responses actually mediate protection or are merely a historical marker of past exposure [4] , [5] . Both antibody specificity and isotype may play a role in protection [6] , [7] and indeed the specific assay used to measure immune function can lead to contradictory or inconsistent conclusions . For example , antibodies that are reactive in an ELISA assay tend to increase with age , but are often not correlated with protection when corrected for age [8] . Antibodies detected using a functional assay that measures inhibition of parasite growth show an increase with age , and an association with protection from clinical disease , but not from infection [9] . Surprisingly , growth inhibitory antibodies ( that can restrict parasite growth in vitro ) are associated with a delay in time-to-infection for individuals within a given age group , but the level of inhibition decreases with age [3] , [10] . Independent of these experimental studies , modeling of malaria infection has also attempted to understand the possible cross-reactivity and molecular targets of malaria immunity , using a heuristic approach based on a qualitative assessment of the data An alternative approach to understanding anti-malarial immunity is to study the dynamics of infection and then predict how these dynamics would be influenced by acquired immunity . That is , by comparing the infection dynamics observed in susceptible ( children ) in contrast to resistant ( adult ) populations , we can use a reverse-engineering approach to understand the differences observed in infection and growth , and predict what immune mechanisms are compatible with such an outcome . Here we use such a modeling approach to understand the effects of naturally acquired immunity on the dynamics of P . falciparum malaria infection in a cohort of 197 people from a holoendemic region of western Kenya .
The details of cohort study have been described elsewhere [3] . Briefly , upon entry into the study ( day 0 ) study participants ( n = 201 ) were treated with Coartem® , which is expected to eradicate blood stage infection but is not effective against liver stage parasites [11] . Therefore , if a study participant was parasitaemic at week two post-treatment , this was considered an emerging liver stage infection and they were eliminated from further study . After treatment , blood smears were monitored weekly for 11 weeks by examination of thick and thin blood smears using light microscopy for the presence of Pf-malaria parasites . In addition , if weekly samples were not collected after week two post-treatment , then the study participant was eliminated; thus leaving 197 remaining for analysis . The cohort was divided into four age groups based on the immuno-epidemiology of malaria; C1 ( children 1–4 years old ( y . o . ) were the least likely to have developed anti-malarial immunity and thus most susceptible to rapid infection; C2 ( 5–9 y . o . ) have begun to develop partially protective immunity though may have limited strain-specific immunity; C3 ( 10–14 y . o . ) have developed anti-malarial immunity that should begin to contain a broader repertoire of strain-specific immunity due to cumulative exposure; and A ( adults>15 y . o . ) who have developed anti-malarial immunity that decreases the rate of infection , parasite growth and provides coverage across many strains . Fig . 1A shows the infection curves for each age group over the 11 weeks of the study . The early stages of infection involve an infected bite , infection of a liver cell , and maturation of the liver stage parasite over 6–7 days [12] . After this , the infected liver cells rupture , releasing approximately 20 , 000 merozoites [13] . These merozoites then initiate the blood stage infection , and successive rounds of blood stage replication follow , each lasting approximately 2 days . These earliest events during human liver and early blood stage infection are not measured in our cohort , and we first detect parasites when their density in blood reaches our detection threshold ( ∼40 parasites/µl ) . Both the rate of initiation of blood stage infection as well as the rate of growth of blood stage immunity will affect the time to first detection of infection ( Fig . 1B ) . Liver stage immunity acts to block some infected bites from reaching the blood stage , and thus the initiation of the blood stage infection will occur with a frequency less than or equal to the real infected biting rate , and occur approximately seven days after the infected bite . Once blood stage infection is initiated , parasites grow at a rate determined by how many merozoites successfully invade new RBCs , and grow to maturity over the two-day life-cycle . We might consider this the ‘parasite multiplication rate’ ( PMR ) , which reflects the fold increase in parasitaemia over the two day life cycle of an infected RBC . Thus , the concentration ( C ) of parasites during their growth in the blood stage can be described by formulawhere A is the initial concentration of parasitized RBC in blood produced by the liver stage ( estimated elsewhere [13] ) , r is the PMR and t is the time in days passed after initiation of blood stage . Incorporating the rate of initiation of blood stage infection ( k ) , we expect the proportion of uninfected individuals remaining at a given time ( S ( t ) ) to follow the equation: ( 1 ) where T is the detection threshold and τ is the first possible moment of blood stage infection after treatment . In our cohort , the youngest children are the most-susceptible to infection ( have the least immunity ) , and thus we use this group as a baseline from which to observe the effects of age-acquired immunity in the older cohorts . The infection curve of the youngest children seems to conform well with the simple dynamics described in equation ( 1 ) ( Fig . 1 . C ) . Thus , assuming that the earliest initiation of the blood stage ( τ ) is day 7 ( due to the pharmacodynamics of lumafantrine ) , we found the rate of initiation of blood stage infection ( k ) at 0 . 066/day , with 95% confidence interval ( CI ) of ( 0 . 056 , 0 . 076 ) , and PMR ( r ) of 5 . 9 , CI ( 3 . 145 , 8 . 671 ) for the youngest age group . Using this fit to the infection of the youngest children , we then attempt to find what mechanisms of immunity can alter this infection curve to produce the curves observed in older age groups .
Previous work suggests that immunity may act anywhere in the pre-erythrocytic stage to block sporozoite invasion [14] or kill infected liver cells . Wherever it acts , the primary results of pre-erythrocytic immunity will be to ( i ) reduce the proportion of infected bites that lead to blood stage infection ( reduce k ) , and/or ( ii ) to reduce the number of infected liver cells that successfully mature to release merozoites and initiate the blood stage ( reduce A ) . From a modeling perspective , anything that reduces the initiation of successful blood stage infection has the same effect – it simply decreases the slope of our reinfection curve ( which remains in exponential form ) . Fig . 2 . A . illustrates the predicted effects of decreasing the rate of initiation of blood stage infection and shows the best fits of a model that incorporates this change to the cohort data . The fits are essentially a family of curves commencing at the same time , and decreasing at different rates . The best fit parameters and goodness of fit statistics are in the [Text S4] , ( model 1 ) . This is quite different from the observed infection rates in two respects: ( a ) it does not capture the greater delay until infection is observed in older children and adults , and ( b ) it does not capture the convex part of the adults' and older children's infection curve . One may think that greater delay in detection of blood stage infection in adults may be explained by other factors such as increased blood volume in adults , or a decrease in the number of infected liver cells surviving to maturity in adults ( due to the effects of cytotoxic T cell immunity killing [13] or growth inhibition [15] ) . However , this is not adequate to explain the observed delay . The number of infected liver cells is rarely higher then 10–14 , [13] , [16] , [17] , and the blood volume differs by only a factor of approximately 5 , thus even combined they could explain at best a delay of 4 . 7 days ( using the PMR of the youngest age group 5 . 9 ) . It is clear that , in addition to underestimating the delay , pre-erythrocytic immunity alone fails to account for the change in shape of the curve associated with age and cumulative exposure . Once the blood stage of infection is successfully initiated , immunity may act on either the growth of the parasite within the infected RBC or directly on the released merozoites to reduce the rate of successful subsequent invasion events . Immunity acting at these stages will have the effect of reducing the parasite multiplication rate , and reduced PMR will lead to a greater delay until infection is detected . However , reducing the PMR alone affects the infection curves simply by shifting the curves to the right . That is , if the rate of successful initiation of the liver stage is unaffected , the effect is only to change the time until parasite density grows above the threshold but not to affect the infection rate . We attempted to fit our model assuming only the PMR changed with age ( Fig . 2 . B ) . The best fit parameters and goodness of fit statistics are in the [Text S4] , ( model 3 ) . . It is once again clear that although growth-inhibiting immunity can account for the delay in infection , it fails to capture the shape of the infection curve . Even combining the effects of liver and blood stage immunity fails to capture the observed infection curves . For example , in the adult infection curves , no combination of liver and blood stage immunity can account for the delayed , slowly starting shoulder of the infection curve , followed by an early period of high infection , followed by a further slowing in infection rate ( Fig . 1 ) . The simple models above consider the effects of a fixed growth rate within a given age group . However , the growth rate of parasites is likely to vary between individuals as a result of a combination of parasite strain and host factors such as intrinsic immune maturation [18] , [19] . Therefore , rather than consider all infections in an age group having the same growth rate [PMR = 6 , for example] , we investigated the effects of variation in the parasite growth rate around the population mean within each group . For simplicity we have assumed a normal distribution in PMR , and the same standard deviation ( as a proportion of the mean ) for each age group . We note that it is not essential that all groups have the same coefficient of variation , however fitting the same parameter to all reduces the number of parameters fitted by three ( See [Text S1] for a complete description of the mathematical model ) . Using a simple assumption that all age groups have the same biting rate and the same type of positive normal distribution of growth rates ( with the standard deviation the same proportion of the mean PMR ) we then fitted a theoretical infection function ( 2 ) to the observed infection proportion , allowing only the mean PMR to vary between age groups ( note that a higher biting rate in adults has been suggested by some [20] and can also be accommodated in the model ) . ( 2 ) g ( . ) is the probability density function of the delay to detection , ( 3 ) fN ( μ , pμ ) ( x ) is the probability density function of a normal distribution with parameters μ and pμ , and μ is the mean PMR in a given age group . Constant p is a positive number same for all age groups . FE ( k ) ( x ) is the cumulative density function of an exponential distribution with parameter k – the average number of bites per day , which is also the same for all age groups . The derivation of the theoretical infection function that incorporates distribution in the time to infective bites and PMR within the host for different age groups is shown in [Text S1] . The best fit parameters and goodness of fit statistics are in the [Text S4] , ( model 4 ) . Remarkably , allowing only the average PMR to vary between age groups captures the main features of the natural infection profiles . We see both the increasing delay with age , as well as the early rapid increase in infection rates followed by an apparent slowing down of the rate of infection ( Fig . 3A ) . With this model of a distribution in PMRs , we can now understand the unusual shape of the adult infection curve: The early slow shoulder of the curve represents the small fraction of infections that grow rapidly , and are detected early ( the right hand side of the distribution in Fig . 3B ) . The rapid phase of infection is around the mean of the PMR curve , and the apparent slowing represents the very slowly growing infections , which are not detected during the 11 weeks of analysis . Because of the distribution of growth rates , there is a proportion of infections where the PMR <1 ( shaded in Fig . 3B ) , implying the number of parasites decreases at each round of infection ( each currently infected RBC infects less than one RBC in next round ) . In children , with mean PMR of ≈3 . 8 , only a small proportion of infections have PMR <1 . However , for the adults , with mean PMR ≈1 . 35 , a large proportion of bites ( ≈24% ) have PMR <1 , which is why we see an apparent slowing of the infection rate later in the study . We note that by contrast simply allowing a distribution of the level of liver stage immunity alone does not improve the fit of the liver stage model ( see [Text S3] for a detailed description of the model ) . Our analysis suggests that a model that can fit the data well is one that assumes a distribution in PMRs within each age group , and a decrease in mean PMRs with age . A key question that follows is – could such a distribution with age arise from a known mechanism of immunity ? Similarly , we have considered above only the relationship between PMR and time to detection of infection . However , decreased PMR would also decrease the observed levels of parasitaemia with age . Is the change in PMR required to produce the reinfection curves consistent with the change in parasitaemia levels with age ? To answer this , we developed a stochastic model of within-host immunity and parasite growth to explore the effects of naturally acquired immunity . We focused on the impact of immunity affecting parasite growth , and further allowed that such immunity might either be strain-specific , or ‘general’ ( affecting all strains equally ) . In this context , a strain-specific immunity may be defined by expression of variable proteins such as var genes or immunologic targets proteins such as merozoite or sporozoite surface antigens ) that can vary across strains . General immunity is a response ( be it immune or physiological – such as hyper-splenism ) that acts equally on all strains . A number of previous models of malaria immunity have considered the effects of partially cross-reactive immune responses [21] , [22] , [23] , [24] . However , we effectively model only the extremes of ‘completely strain-specific’ or ‘completely cross reactive’ responses for simplicity . In our model , strain-specific immunity arises as the result of the infection with a certain strain and neutralises the parasites of this strain only . General immunity arises as the result of infection with any strain and it can neutralise parasites of any strain . The acquisition rate of both types of immunity is proportional to concentration of parasites in the blood . Without infection , the existing level of immunity decays at some constant rate . Thus the model has four parameters of immunity: for both strain-specific immunity and general immunity we have a rate of increase in immunity per unit parasite ( denoted by α and γ respectively ) , and a rate of loss of immunity in the absence of parasite denoted ( denoted by β and δ respectively ) . We also require three parasite parameters: the mean number of bites per day k ( biting rate ) , the baseline multiplication rate r of the parasites , and the number of different strains n . The basic equations of parasite – immunity dynamics of the model are: ( 4 ) Here Pi is the concentration of parasites of the strain i , Si is the strength of the i th strain specific immunity , . G is the strength of the general immunity . The function h ( . ) is either one ( if the concentration of parasites<threshold ) or zero ( if the concentration of parasites≥threshold ) , and allows for the elimination of parasites and the decay of immunity when the level of parasites drops below a threshold Z ( in this case 0 . 005 parasites/microlitre ) . A detailed description of the mathematical model is in the [Text S2] . Using this model we can simulate the ‘life history’ of an individual in a malaria endemic area . In the model , we start with the assumption that all individuals are identical and lack immunity at birth ( t = 0 ) , and are then exposed to bites from different strains arriving at random times , with a random order of strains ( choosing from our 50 notional strains ) , to which they then develop immunity ( ignoring the transient contribution of maternally-derived immunity ) . Thus , the timing of bites and which strain is inoculated is stochastic , but between bites the dynamics of parasites and immunity is deterministic ( as in equation 2 ) . Fig . 4 shows the dynamics of parasite infection and acquisition of strain specific and general immunity by an individual , starting from birth to one year [left panel] and five to six years [right panel] . The top panels show the dynamics of parasite infection , as different parasite strains ( indicated by different colours ) initiate blood stage infection , grow , and then induce strain-specific immunity , leading to their clearance . Given the long half-life of strain-specific immunity and the absence of within-host parasite antigen variation in our model , each new parasitemia peak represents infection with a new strain . In addition to inducing strain-specific immunity ( bottom panels , coloured ) , these infections also induce general immunity ( solid black line ) , which accumulates over time . By simulating the life history of a small population of individuals ( n = 50 ) , we can then apply the concept of ‘treatment-time-to-infection’ trials to the simulated individuals . That is , by removing all blood stage parasites of individuals in different age groups and observing the time until parasite levels reach our detection threshold , we can simulate ( re ) -infection ( Fig . 5 and 6 ) . Fig . 5 shows the dynamics of parasitaemia during ( re ) -infection from the field study data ( top ) and the simulation ( bottom ) , for four subjects from each of the different age groups . Remarkably , the simulation captures a number of the factors observed in our observational study; firstly , the natural infection curves show increasing delay with age and an increasing proportion of individuals remaining uninfected . Secondly , the observed reduction in parasite levels in blood with age is also captured in the model , indicating that the decreased PMR required to produce the reinfection curves is consistent with the decreased PMR required to produce the observed reduction in parasitaemia with age . ( as higher immunity means parasites are controlled at a lower parasitaemia ) ( Fig . 5 ) . Fig . 6A shows the infection curves for the whole simulated population , and Fig . 6B compares the mean parasitaemia for the field study data and simulation for different ages . Importantly , the major factor that can account for both delayed infection and lower parasitaemia in adults is simply a reduced average growth rate of parasites with age and naturally acquired immunity .
A number of modelling studies have previously been applied to understanding the dynamics of malaria infection and the impact of immunity . Some of these studies have focused upon the dynamics of experimental infection of neurosyphilis patients in the USA during the early 20th century with both P . falciparum [25] and P . vivax [26] . Such infections with P . falciparum often show a dynamic of repeated recrudescence , in the absence of reinfection . Modelling both infection rates and infection dynamics has also been applied to more recent data on infection in naturally exposed individuals [27] , [28] . The recent discovery of the molecular mechanisms of antigenic variation in P . falciparum has also driven studies modelling the dynamics of immune interaction with an antigenically variable pathogen [21] , [22] , [23] , [24] . In our study , we utilise field data on malaria infection dynamics after a short treatment course in a holoendemic region of Kenya . The estimated P . falciparum infection rates are extremely high , especially amongst young children , but consistent with the high entomological infection rates measured in this area [29]and with the blood-stage infection rates previously observed by others in similar studies [30] , [31] . The prior treatment of patients gives an ideal opportunity to study malaria incidence in the absence of recrudesence from prior infection . However , there is a potential problem equating the measured time to reinfection in treated , asymptomatic individuals with incidence rates of new blood-stage infection in individuals living in highly endemic areas , and then using these findings to make general statements about acquired immunity to malaria . However , given that children in this area are treated on average twice a year [32] , this is likely as natural an infection dynamic as can be readily studied . Fitting of a simple model of infection dynamics to the field data on P . falciparum infection demonstrates that the observed rates of infection with age are consistent with a form of immunity that reduces the PMR of the parasite in blood . This mechanism of blood-stage immunity reducing PMRs is consistent not only with the dynamics of infection after treatment in different age groups , but also with the observed reduction in peak parasitaemia with older age groups ( Fig . 6 ) . Although clinical malaria was not studied in this model , one might imagine a reduction in parasite replication with age may also play some role in both the observed reduction in episodes of severe malaria as well as an apparent shift from acute to chronic infections with age . The fact that the model fits well without invoking a need for liver-stage immunity does not exclude a role for liver stage immunity , but suggests it is not a major force shaping the observed infection dynamics . By contrast , liver-stage immunity alone would be expected to have relatively little impact on either the delay or peak parasitaemia . The estimated PMR in this study and their change with age are very consistent with published studies of parasite growth in unexposed adults in the UK ( estimated to have a high PMR of approx 8–14/cycle [12] , [13] ) , compared with the reported very low PMRs in malaria endemic areas ( 1 . 6–3/cycle ) [33] , [34] . Although the initial model was developed on the basis of a number of heuristic assumptions , the model provides some useful insights into the possible mechanisms of blood-stage immunity induced from natural malaria infections . Firstly , ‘parasite clearing’ immunity must be highly strain-specific: That is , we often see one parasite peak being cleared immediately before another peak arises ( Fig . 5 , C1 , top ) . Given the time taken for clearance and growth , this means that we must have one parasite being cleared while another grows . Thus the immune response mediating clearance of individual parasite peaks must have the characteristics of being both highly strain-specific , and rapidly induced . Although this strain-specific response must be rapidly induced , the duration of strain-specific immunity may be varied in the model , depending on the number of strains used and the biting rate . That is , if there are only a small number of immunologically distinct strains of parasite , then the duration of immunity needs to be short – to allow reinfection . However , if there are a large number of immunologically distinct strains , the half-life of strain-specific immunity could be considerably longer and still consistent with the data . Thus we are not able to speculate on the relative duration of strain-specific versus general immunity , as has been discussed elsewhere [21] , [22] , [23] , [24] . The long timescale over which age-related immunity is acquired suggests that whatever factor mediates this must have a long half-life ( since the time taken for the immunity level to reach its maximum is related to the half-life of immunity ) . If specific immunity were the only mechanism , then it would need to be long-lived , and the age-associated acquisition of immunity would simply be the progress from having immunity to a few strains , to having immunity to most strains . However , the intermediate phenotype here ( having immunity to a proportion of strains ) is incompatible with the age and parasitaemia data . That is , if we had immunity to 50% of strains , and no immunity to the other 50% , this would mean that 50% of bites would be blocked ( changing the reinfection curve as in Fig . 2A ) and that parasitaemia of the remaining strains would be ‘normal’ . This is clearly not consistent with the data , suggesting that although strain-specific immunity is required to explain the clearance of individual parasite peaks , it is not sufficient to explain age-associated changes in reinfection and parasitaemia . The alternative to a strain-specific immune response to individual infection episodes is a general anti-parasite response that decreases the growth of all strains . In contrast to specific immunity , general immunity takes many years to acquire and accumulates after repeated infections . Thus , each infection should both induce only a small rise in general immunity , and it must be long-lived . It is also clear that general immunity alone cannot produce the observed infection dynamics , as this could not simultaneously drive one parasite strain down while the other grows . In broad terms , we might consider that an increase in general immunity with time drives down the mean PMR of an age group . The distribution in growth rates within an age group arises due to strain-specific immunity , with individual strains ( in different individuals ) having different growth rates due to the variable levels of strain-specific immunity , which wanes over time after infection with that strain . Clinical studies of immune correlates of protection from malaria are often complicated and difficult to reconcile ( reviewed in [8] ) . Overall , antibody levels have been found to generally increase with age and exposure , when measured by the ability to bind parasite antigens in ELISA . It has generally been difficult to demonstrate a clear correlation with protection from parasitaemia , as this is so strongly confounded by age . These antibody levels are also long-lived , appearing to persist with a half-life up to 5–10 years [35] . However , it is difficult to differentiate between binding antibodies exerting a direct protective effect , or merely acting as biomarkers of past infections . Our model does not address the particular mechanisms or specificity of immunity . However , one prediction of the model is that each patent infection is cleared by a highly strain-specific response , as it is common that the clearance of one strain occurs while infection with another strain is still growing . This is consistent with early studies suggesting that growth-inhibitory antibodies may be highly strain-specific when tested in clinical isolates [36] . Interestingly , some growth-inhibitory antibodies appear as a correlate of protection within an age group , but paradoxically decrease with age [3] , [10] . It is interesting to note that strain-specific immunity also paradoxically decreases with age in our model , due to the increase in general immunity . We note some difficulties with defining the term ‘strain’ in malaria , as recent deep sequencing studies have suggested a wide range of genotypes [37] . Moreover , even once the genetic structure of malaria strains is resolved , the immunological specificity and cross-reactivity of responses may be highly complex . We do not use the term ‘cross-reactive immunity’ when describing a factor that limits the growth of all strains , but instead prefer the term ‘general immunity’ . The concept of cross reactivity may tend to imply specific antibodies that cross-react between different strains . However , we do not wish to exclude other factors such as an enlarged spleen may tend to decrease the PMR of all strains , and may also play a role in protection . The majority of cases of severe malaria occur amongst young children in endemic areas , and age and persistent exposure provide some level of immunity to both clinical symptoms , as well as measured parasitaemia [38] . Although this age-associated acquisition of immunity is known , clinical studies of immune responses to malaria often struggle to identify the mechanisms of protection [38] . We have employed an alternative strategy in an attempt to deconstruct the mechanisms of naturally acquired immunity based on the dynamics of infection in otherwise healthy asymptomatic individuals . We find that simple pre-erythrocytic immunity acting alone is inconsistent with the data , as is a highly strain-specific form of blood-stage immunity , as both would have the overall effect of merely decreasing the number of successful infections . Instead , the observed dynamics of infection and parasitaemia levels in a cohort of individuals from a holoendemic area of western Kenya are consistent with the reduction in parasite multiplication rate with age , leading to both a delay in the time until infection is detected , as well as reduce the peak parasitaemia . In terms of the clinical outcomes observed , our study suggests that the reduced rate of clinical malaria and reduced rate of apparent infection ( measured at a particular threshold of parasitaemia detection ) may both be driven by a reduction in parasite growth rate , independent of any real change in infection rate . Further studies are required to deconvolute immunity that inhibits parasite growth distinct from its ability to infect hepatocytes and erythrocytes .
|
Human malaria infections resulting in serious complications and death occur predominantly in young children , and resistance is gradually acquired with repeated exposure . Malaria parasites have two major stages within the human host during its life cycle: an initial liver stage , and the subsequent blood stage , where parasites replicate in and destroy red blood cells . The mechanisms of acquired resistance to severe malaria may involve immunity to both the liver and blood stage parasites . However the relative contribution of each type of immunity is not yet understood . To gain novel insight , we have analysed data from a malaria exposed cohort from western Kenya . We used mathematical modeling to understand what form of immunity is consistent with the observed rates of reinfection in adults and children seen in the field study data . We found that the reinfection pattern can be completely explained by blood stage immunity . Moreover , the blood stage immunity must consist of rapidly-induced strain-specific immunity that clears individual infections , and general immunity that accumulates slowly and decreases the average parasite growth rate with age . Understanding the dynamics of naturally acquired immunity and infection provides important insights for effective vaccine development .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"population",
"modeling",
"biology",
"computational",
"biology",
"infectious",
"disease",
"modeling"
] |
2012
|
The Dynamics of Naturally Acquired Immunity to Plasmodium falciparum Infection
|
Urinary tract infections ( UTIs ) afflict over 9 million women in America every year , often necessitating long-term prophylactic antibiotics . One risk factor for UTI is frequent sexual intercourse , which dramatically increases the risk of UTI . The mechanism behind this increased risk is unknown; however , bacteriuria increases immediately after sexual intercourse episodes , suggesting that physical manipulation introduces periurethral flora into the urinary tract . In this paper , we investigated whether superinfection ( repeat introduction of bacteria ) resulted in increased risk of severe UTI , manifesting as persistent bacteriuria , high titer bladder bacterial burdens and chronic inflammation , an outcome referred to as chronic cystitis . Chronic cystitis represents unchecked luminal bacterial replication and is defined histologically by urothelial hyperplasia and submucosal lymphoid aggregates , a histological pattern similar to that seen in humans suffering chronic UTI . C57BL/6J mice are resistant to chronic cystitis after a single infection; however , they developed persistent bacteriuria and chronic cystitis when superinfected 24 hours apart . Elevated levels of interleukin-6 ( IL-6 ) , keratinocyte cytokine ( KC/CXCL1 ) , and granulocyte colony-stimulating factor ( G-CSF ) in the serum of C57BL/6J mice prior to the second infection predicted the development of chronic cystitis . These same cytokines have been found to precede chronic cystitis in singly infected C3H/HeN mice . Furthermore , inoculating C3H/HeN mice twice within a six-hour period doubled the proportion of mice that developed chronic cystitis . Intracellular bacterial replication , regulated hemolysin ( HlyA ) expression , and caspase 1/11 activation were essential for this increase . Microarrays conducted at four weeks post inoculation in both mouse strains revealed upregulation of IL-1 and antimicrobial peptides during chronic cystitis . These data suggest a mechanism by which caspase-1/11 activation and IL-1 secretion could predispose certain women to recurrent UTI after frequent intercourse , a predisposition predictable by several serum biomarkers in two murine models .
Nearly nine million people present each year to primary care physicians with a urinary tract infection ( UTI ) , costing nearly $2 billion yearly [1] , [2] . Women suffer the majority of these infections , with the lifetime risk approaching 50% [3] . Furthermore , 25–40% of these women will suffer recurrent UTI ( rUTI ) , with 1 . 5 million women referred to urology clinics and often requiring prophylactic antibiotics to prevent recurrence [4]–[6] . Uropathogenic E . coli ( UPEC ) are responsible for>80% of community acquired UTI and 50% of nosocomial UTI [7] , [8] . In the absence of antibiotic therapy , up to 60% of women experience symptoms and/or bacteriuria lasting months after initial infection [9]–[12] , implying that cystitis is not always self-limiting . Furthermore , if the infection persists without adequate treatment , the organisms have the capacity to ascend the ureters , causing pyelonephritis and sepsis [13] . Antibiotic resistant organisms further complicate infection and threaten to increase the likelihood of chronic UTI , pyelonephritis and potentially bacteremia [14] , [15] . UTIs are increasingly being treated with fluoroquinolones , which in turn has led to a rise in resistance and the spread of multi-drug resistant microorganisms globally , which is a looming worldwide crisis [16] , [17] . It is therefore imperative to understand the molecular mechanisms that underlie this problematic disease in order to develop novel therapies . Sexual intercourse is one of the most significant risk factors predisposing otherwise healthy women to UTI . Early studies demonstrated that sexual intercourse led to a 10-fold increase in bacteria/ml of urine and a subsequently increased predisposition to developing a UTI within 24 hours thereafter [5] , [18]–[21] . More recent studies have shown that the frequency with which a woman has sexual intercourse dramatically impacts the likelihood of developing both acute and rUTI [4] , [22] , [23] . Scholes et . al found a direct association between the number of episodes of sexual intercourse in a given month and the risk of developing rUTI . However the significance of the timing between these episodes of sexual intercourse is unknown . Are evenly spaced episodes associated with an equal risk or , instead , does an episode prime the bladder for rUTI if another insult follows within a sensitive period ? To address this question , we developed a model of sequential infection in mice to explore the hypothesis that a sensitive period exists after an initial bacterial insult to the bladder in which the likelihood of developing severe , chronic infection is dramatically increased . Murine models of UTI have been used to decipher complexities of this disease in naïve individuals . UPEC are capable of colonizing multiple body habitats and niches , including both intracellular and extracellular locations within the bladder , as well as in the gastrointestinal ( GI ) tract and the kidneys . Selective pressure and bacterial population bottlenecks during colonization impact the ultimate fate of disease [24]–[27] . Adhesive pili assembled by the chaperone/usher pathway ( CUP ) , such as type 1 pili , contain adhesins at their tips that function in adherence and invasion of host tissues and in biofilm formation on medical devices . Upon introduction of UPEC into the bladder , bacteria bind to either mannosylated uroplakin plaques or β1-α3 integrin receptors on the epithelial surface of the bladder via the type 1 pilus FimH adhesin [28]–[30] . Upon internalization , UPEC can be exocytosed as part of a TLR4 dependent innate defense process [31] . In addition to expulsion of individual bacteria , the host can exfoliate superficial facet cells to shed attached and invaded bacteria into the urine for clearance [29] . A small fraction of invaded bacteria escape into the host cell cytoplasm , where they are able to subvert expulsion and innate defenses by replicating into biofilm-like intracellular bacterial communities ( IBCs ) [24] , [32] . UPEC eventually flux out of these communities with a substantial proportion existing as neutrophil resistant filaments [33] , [34] . Importantly , evidence of IBCs and bacterial filaments have been observed in women suffering acute UTI , one to two days post self-reported sexual intercourse , but not in healthy controls or infections caused by Gram-positive organisms , which do not form IBCs [21] . IBCs have also been observed in urine from children with an acute UTI [35] . Additionally , IBC formation and the innate immune response of cytokine secretion and exfoliation have been observed in all tested mouse strains , but the long-term outcome of infection differs [36]–[38] . There are two main , mutually exclusive , outcomes to acute infection in C3H/HeN mice: either chronic bacterial cystitis ( chronic cystitis ) , which is characterized by persistent high titer bacteriuria ( >104 CFU/ml ) and high titer bacterial bladder burdens ( >104 CFU ) two or more weeks after inoculation , accompanied by chronic inflammation [37] , [39] , or resolution of bacteriuria [37] . Mice that resolve infection may harbor small populations of dormant UPEC called Quiescent Intracellular Reservoirs ( QIRs ) [40] . Other mouse strains exhibit varied proportions of these two outcomes . C57BL/6J mice resolve bacteriuria within days and thus are resistant to chronic cystitis , but are susceptible to QIR formation [40] , [41] . In contrast , other TLR4-responsive C3H background sub-strains and closely related CBA/J and DBA/2J mice experience persistent high-titer bacteriuria and bladder colonization by UPEC in the presence of chronic inflammation lasting at least four weeks post-infection ( wpi ) . During chronic cystitis , persistent lymphoid aggregates and urothelial hyperplasia with lack of superficial facet cell terminal differentiation accompany luminal bacterial replication [37] . These same histological findings of submucosal lymphoid aggregates and urothelial hyperplasia have been observed in humans suffering persistent bacteriuria and chronic cystitis [42] . Since murine chronic cystitis predisposes to recurrent chronic UTI after antibiotic-mediated bacterial clearance , this is also a relevant model to interrogate the mechanism of recurrent cystitis [37] . In mouse models of UTI , mice initially experience urinary frequency and dysuria as determined by reaction to noxious stimuli and nerve responses during acute infection [43] , [44]; however , during chronic cystitis bacterial replication may exist in an asymptomatic carrier state as studies have not been conducted to determine whether dysuria persists . Interestingly , higher serum levels of interleukins ( IL ) 5 and 6 , keratinocyte cytokine ( KC/CXCL1 ) , and granulocyte colony-stimulating factor ( G-CSF ) in C3H/HeN mice at 24 hours post infection ( hpi ) predicted the development of persistent bacteriuria and chronic cystitis thereafter , suggestive of a host-pathogen checkpoint during acute infection that predicts long term outcome [26] , [37] . In women with an acute UTI , increased amounts of serum CXCL1 , M-CSF , and IL-8 correlated with subsequent rUTI , suggesting a similar checkpoint [45] . In this manuscript , we developed a superinfection model to mimic the clinical scenario of frequent sexual intercourse whereby sequential inocula are introduced within a brief period of time . C57BL/6J mice are resistant to chronic cystitis when singly infected; however , 30% of C57BL/6J mice developed chronic cystitis when superinfected 24 hours after the initial infection . Serum elevations of IL-6 , KC , and G-CSF prior to superinfection predicted the development of persistent bacteriuria in C57BL/6J mice similar to singly infected C3H/HeN mice . Superinfecting C3H/HeN mice 1–6 hours after the initial inoculation increased the proportion of mice experiencing chronic cystitis . In order for this elevation to occur , we found that the initial UPEC inoculum ( the “priming” inoculation ) must be alive , invasive , capable of intracellular replication , and able to regulate hemolysin expression . Inhibition of the caspase 1/11 inflammasome prior to priming reduced bacterial CFU at four wpi relative to DMSO-treated mice . Microarray analysis of mouse bladders four wpi revealed that both C57BL/6J and C3H/HeN mice secreted antimicrobial peptides and IL-1 during chronic infection . In contrast to C3H/HeN mice , immunoglobulin expression was upregulated in C57BL/6J mice experiencing chronic cystitis . This immunoglobulin expression was absent in C57BL/6J mice that resolved infection and in C3H/HeN mice . Our data suggest mechanisms whereby certain women may be susceptible to rUTI after frequent sexual intercourse dependent on intracellular bacterial replication and the host immune response .
Studies suggest that a host-pathogen checkpoint within the first 24 hpi determines UTI outcome in C3H/HeN mice [26] , [37] . In addition , the chronic inflammation observed in mice experiencing chronic cystitis was found to predispose to rUTI after re-infection [37] . Thus , we hypothesized that superinfecting mice during this period of acute inflammation would increase the proportion of mice experiencing chronic cystitis . We transurethrally infected 7–8 week old female C3H/HeN mice with 107 CFU UTI89 or PBS as the priming inoculation and superinfected them 1–2 , 6 , or 24 hours thereafter . Enumeration of bacterial CFU at one wpi as an initial screen revealed a dramatic increase in the proportion of mice experiencing chronic cystitis in mice superinfected 1–6 hours after priming compared to singly infected or PBS treated mice ( Fig . 1A ) . We used a cutoff of 106 CFU to demarcate mice experiencing high-titer bacterial infection at one week . Importantly , we did not observe a significant increase in CFU when a single inoculum was doubled ( 2×107 CFU ) . Superinfection at 24 hpi had no effect on bacterial titers at one week , suggesting that the factors predisposing to increased susceptibility to chronic cystitis upon superinfection wane over time [26] . However , inoculation with PBS followed by UTI89 24 hpi did lead to high titers in 60% of mice . While this result is perplexing , it possibly reflects that sacrifice six days post infection was not sufficient to delineate the typical bimodal distribution of outcomes [37] . The process of catheterization also induces inflammation , which may not have resolved by 6 dpi [46] . We conducted all subsequent C3H/HeN superinfections one hour after priming . Since early severe inflammatory responses predispose to chronic cystitis [37] , we hypothesized that the initial inoculum primed the bladder by initiating an innate immune response to intracellular bacteria that predisposed to a higher proportion of mice experiencing chronic cystitis upon superinfection . We utilized a panel of UTI89 mutants in fimH , ompA , and kps that have been shown to differ in their ability to: i ) invade and form IBCs and ii ) persist during chronic cystitis in co-infection experiments [47] , [48] . Mature IBCs caused by WT bacteria are clonally derived from a single invasive event [24] . The mannose-binding pocket of FimH is invariant among sequenced UPEC [47] , and the binding pocket mutant , FimH::Q133K , is defective in mannose-binding and can neither invade the bladder epithelium nor form IBCs . FimH undergoes compact and elongated conformational changes wherein the receptor binding domain bends approximately 37° with respect to the pilin domain . The mannose-binding pocket is deformed in the compact conformation whereas the elongated conformation is mannose binding proficient [49] , [50] . Several residues outside the mannose-binding pocket ( positions 27 , 62 , 66 and 163 ) are under positive selection in clinical UPEC isolates compared to fecal strains [47] and have been shown to function in modulating the conformational changes between the elongated and compact states [48] . FimH::A27V/V163A predominantly adopts a high-mannose binding , elongated conformation . Its expression results in: i ) a 10-fold reduction in intracellular CFU one hpi and ii ) a defect in the ability to form IBCs at six hpi . FimH::A62S shifts the equilibrium towards the compact conformation . Expression of this allele results in: i ) a 10-fold reduction in intracellular CFU one hpi and ii ) a 10-fold reduction in IBC formation compared to WT UTI89 [47] , [48] . UTI89ΔompA forms half the number of IBCs as UTI89 [51] , and UTI89Δkps is defective in IBC formation . UTI89Δkps can replicate intracellularly and the IBC defect can be rescued by co-inoculation with WT UTI89 , which results in mixed strain , non-clonal , IBCs [52] . We primed mice with these strains and superinfected one hpi with WT UTI89 and assessed bacteriuria at days 1 , 7 , 14 , and 21 and enumerated bladder titers at 28 dpi . Mice were designated as having chronic cystitis if they had urine bacterial titers greater than 104 CFU/ml at each time point and bladder titers greater than 104 CFU at sacrifice [37] . We found that the FimH::A27V/V163A allele was incapable of priming the bladder for the development of chronic cystitis ( p<0 . 05 relative to WT superinfection ) . In contrast , FimH::A62S did not significantly differ from PBS or WT superinfection; therefore , it may be capable of priming , though to a lesser degree . UTI89ΔompA and UTI89Δkps were both able to prime the bladder for enhanced chronic cystitis relative to PBS when superinfected one hpi with WT UTI89 ( p<0 . 05 and p<0 . 01 respectively; Fig . 1C ) . We also primed with heat-killed UTI89 and found that live , but not heat killed , UTI89 were capable of priming the bladder indicating that bacterial products such as LPS were insufficient ( Fig . 1B ) . These data indicate that live and invasive UTI89 capable of at least some degree of intracellular replication are required for the priming to enhance the incidence of chronic cystitis upon superinfection of UTI89 . Taken together these data suggest that priming begins during invasion and early IBC formation . One of the most potent host defenses to eliminate adherent and invaded UPEC is superficial facet cell exfoliation [29] . The process of exfoliation is activated in part by the bacterial expression of hemolysin ( HlyA ) [53] ( Nagamatsu et al . in review ) . UTI89ΔcpxR overexpresses HlyA , leading to exfoliation and attenuation in our murine model of cystitis ( Nagamatsu et al . in review ) . The UTI89ΔcpxRΔhlyA double mutant was not attenuated , suggesting that the in vivo defect was due to increased hemolysin expression ( Nagamatsu et al . in review ) . The ability of UPEC to rapidly build up in numbers in the form of IBCs and then disperse to neighboring cells may be part of a mechanism to subvert an exfoliation response . Thus , fine-tuning the expression of HlyA during acute bladder infection may serve to maximize UPEC persistence and give UPEC a fitness edge against the host innate inflammatory response . Interestingly , in C3H/HeN mice , UTI89 ΔhlyA is not attenuated throughout infection and causes chronic cystitis comparable to UTI89; however , other reports suggest deletion of HlyA in UPEC CFT073 decreases virulence [54] . We investigated the role of hemolysin in priming the bladder for chronic cystitis upon superinfection by utilizing UTI89ΔhlyA or UTI89ΔcpxR as the initial inoculation followed by WT UTI89 one hpi . Both of these strains were statistically significantly different when compared to WT UTI89 as the priming inoculum . Therefore , we conclude that neither was capable of priming the bladder for enhanced chronic cystitis ( Fig . 1D ) . Thus , too high or low expression of hemolysin abolished the ability of UTI89 to prime for enhanced chronic cystitis implying that an optimal level of hemolysin expression is critical for priming the bladder for enhanced chronic cystitis . HlyA-mediated exfoliation is in part due to its ability to trigger degradation of paxillin , a scaffold protein that modulates the dynamics of cytoskeletal rearrangements [55] . HlyA can also trigger cell death in human bladder epithelial cells and release of IL-1α via caspase-4 ( the murine ortholog is caspase-11 ) activation and caspase-1-dependent IL-1β secretion via activation of the NLRP3 inflammasome pathway , which orchestrates additional cell death ( Nagamatsu et al . in review ) . We hypothesized that inflammasome and caspase 1/11 activation were essential for superinfection . Thus , mice were treated intravesically with a dose of caspase 1/11 inhibitor or DMSO one hour prior to priming and a second dose with the priming inoculum to test this hypothesis ( Fig . 2A ) . Providing two doses of the inhibitor was previously shown to be effective in dampening in vivo inflammatory responses . In vitro , the inhibitor dramatically reduced downstream elements of inflammasome activation , IL-1α and IL-1β secretion , when bladder cells were infected with UTI89 ( Nagamatsu et al . in review ) . Caspase 1/11 inhibition significantly reduced median bladder titers at four weeks after superinfection relative to the DMSO control group ( Fig . 2B ) . We also saw a trend of caspase 1/11 inhibition in reducing the proportion of WT superinfected mice experiencing chronic cystitis to single infection levels ( Fig . 2B ) . DMSO also reduced the proportion of mice experiencing persistent bacteriuria and chronic cystitis , but to a lesser degree than caspase 1/11 inhibition ( Fig . 2B vs . Fig . 1B–D ) , suggesting an anti-inflammatory role of DMSO alone . Intriguingly , DMSO was recently found to inhibit the NLRP3 inflammasome [56] . Taken together , these data implicate hemolysin and the NLRP3 inflammasome in the priming response to enhanced chronic cystitis . We further investigated whether chemical exfoliation could enhance the proportion of mice experiencing chronic cystitis prior to a single infection . We utilized the cationic protein , protamine sulfate , which has previously been used to exfoliate the superficial facet cell layer of the urothelium [40] , [57] . A 10 mg/mL dose delivered intravesically in 50 µL PBS was shown to exfoliate 65% of the facet cell layer 12 hours after treatment while an additional booster dose of 50 mg/mL led to 95% exfoliation [40] . We utilized these concentrations to initiate , but likely not complete , the process of exfoliation one hour prior to infection with UTI89 . We did not observe a significant increase in the proportion of mice experiencing chronic cystitis over PBS pretreatment ( Fig . 2C ) . Thus , these data suggest that at least partial IBC formation in conjunction with caspase 1/11 activation primes the bladder for enhanced chronic cystitis , but chemical initiation of exfoliation is not sufficient . Taken together , these data suggest that exfoliation per se might not play a significant role in impacting the likelihood of enhanced chronic cystitis but instead may reflect a downstream marker of the priming event . C57BL/6J mice typically rapidly resolve bacteriuria and are resistant to chronic cystitis upon single inoculation with UPEC [37] , [38] . Five to ten percent of the time after inoculation with UTI89 , C57BL/6J mice experience persistent bacteriuria , but this is generally due to kidney infection without concomitant high titer bladder infection [37] , [41] . This degree of kidney infection is not infectious dose dependent and therefore likely due to ureteric reflux of the bacteria during experimental inoculation [37] . We investigated whether superinfecting C57BL/6J mice during acute infection would stimulate an immune response leading to chronic cystitis . We inoculated bladders with PBS or 107 CFU of UTI89 followed by superinfection with UTI89 1 , 6 , 24 , 48 hours or one week after initial infection and collected urine at days 1 , 7 , 14 , and 21 dpi followed by enumeration of bladder and kidney titers at 28 dpi ( Fig . 3 ) . A 24 hpi superinfection resulted in 35% of mice sustaining persistent bacteriuria with bladder titers >104 CFU at four weeks compared to 0% in the singly infected group ( Fig . 3A ) . Kidney titers were also increased in the mice with persistent bacteriuria , but we did not observe a significant increase in the proportion of mice with kidney infection greater than 104 CFU ( Fig . 3B ) . These data suggest that at 24 hours after infection the bladders of C57BL/6J mice were primed to develop chronic cystitis upon superinfection . We investigated whether an ascending kidney infection plays a role in predisposing these mice to chronic cystitis by inoculating PBS into the bladder , either 24 hours before or after infection with UTI89 , to stimulate a bladder and ureter stretch response or potentially increase reflux of bacteria into the kidneys , respectively . We determined the percentage of mice with persistent bacteriuria and those with bladder and kidney titers greater than 104 CFU at sacrifice ( Table 1 ) . We found in all conditions that persistent bacteriuria was a 100% predictor of kidney titers>104 CFU at four wpi . Persistent bacteriuria also predicted bladder titers greater than 104 CFU at four wpi in C57BL/6J mice superinfected 24 hpi with UTI89 . For the group of mice inoculated with PBS before the initial UTI89 infection , persistent bacteriuria did not correlate with high bladder titers suggesting these bacteria were only replicating in the kidneys . Serially infecting with two inocula of UTI89 trended towards increased persistent bacteriuria and chronic cystitis compared to the group inoculated with UTI89 followed by PBS at 24 hpi ( P = 0 . 066; Table 1 and Fig . 4A ) . Kidney titers of UTI89 superinfected mice were significantly higher than when PBS was used to prime or superinfect perhaps suggesting that repeat infection may also increase susceptibility to pyelonephritis ( Fig . 4B ) . Thus , a 24 hpi superinfection of WT UTI89 led to increased rates of persistent bacteriuria and chronic cystitis; however , bladder/ureter stretch or kidney ascension at 24 hpi may contribute to this increase . C3H/HeN mice that progress to chronic cystitis upon single inoculation can be predicted by elevated serum levels of IL-5 , IL-6 , KC , and G-CSF at 24 hpi [37] . We hypothesized that similar elevations would predict sensitization to chronic cystitis in C57BL6/J mice if they were subsequently superinfected . Thus , we determined levels of 23 serum cytokines from C57BL/6J mice 24 hrs after initial inoculation with PBS or UTI89 prior to superinfection . We then superinfected a subset of the mice initially infected with UTI89 ( superinfection in Fig . 5 ) leaving the other mice untouched ( UTI89 group ) . All mice were evaluated with urine titers over 28 d and sacrificed to enumerate bladder titers . We stratified the superinfected mice based on outcome four weeks later as determined by persistent bacteriuria and chronic cystitis . We found elevations of serum KC ( Fig . 5A ) , IL-6 ( Fig . 5B ) , and G-CSF ( Fig . 5C ) in mice that progressed to chronic cystitis relative to those that resolved infection or were mock-infected with PBS . Therefore , higher levels of these cytokines correlate with chronic cystitis that develops later if mice are superinfected . At the time we obtained serum , the single infection and superinfection groups were identical , and no statistical differences existed among them . These data demonstrate that a subset of C57BL/6J mice respond to an initial infection in a way that results in higher specific serum cytokine levels and primes them to develop chronic cystitis if an additional insult is delivered 24 hpi . During chronic cystitis of singly-infected C3H/HeN mice , the bladder epithelium is hyperplastic and normal terminal differentiation of the superficial facet cell layer , including the expression of surface uroplakins , does not occur [37] . In this environment , the bacteria are able to persist extracellularly by an unknown mechanism . To assess this , we conducted scanning electron microscopy analysis on bladder tissue harvested at four wpi and found that bacteria replicate in the presence of ongoing epithelial exfoliation and neutrophil influx in chronic cystitis of both C3H/HeN and C57BL/6J mice ( S1A–D Fig . ) . This analysis supports previous experiments that have shown that during chronic cystitis the majority of bacteria are extracellular , replicating in the urine or adherent to underlying transitional epithelial cells [24] , [37] . The mechanism by which bacteria adhere in the absence of uroplakins has not been demonstrated in vivo , but in vitro studies have shown that FimH binds integrins and other host proteins such as TLR4 [30] , [58] , [59] . Alternatively additional adhesive factors such as other CUP pili may play a role . Interestingly , during chronic cystitis , neutrophils , which we observed to be actively engulfing bacteria , are insufficient for clearing infection; however , the reason for this is unclear . Mature superficial facet cells could not be discerned at this time point , but were present in mock-infected mice ( S1E Fig . ) . Patients with persistent bacteriuria or rUTI have been reported to have similar histopathology [42] . In order to identify the bladder micro-environment in which UPEC replicate during chronic cystitis , we conducted microarray analysis on RNA extracted from bladders four wpi . C3H/HeN mice were singly-infected and C57BL/6J mice were superinfected to develop chronic cystitis . Mice from each strain inoculated with PBS were used as controls . Depicted in Fig . 6 are the expression profiles relative to the global average with green indicating increased expression and red denoting decreased . C3H/HeN mice experiencing chronic cystitis had a dramatically different expression profile from resolved and mock-infected mice ( Fig . 6A ) . Uroplakins were among the most downregulated genes during chronic cystitis in both mouse models , consistent with the lack of terminally differentiated superficial facet cells ( S1 Fig . ) . Eleven of the 20 ( 55% ) most upregulated genes during chronic cystitis were the same in both mouse strains ( S1 Table ) . The functional categorization revealed that most of the up-regulated genes function in inflammatory response , cytokine release , and ion binding [60]–[62] . Of interest among these genes in both of these mouse models is the inflammasome-related cytokines IL-1 . We have shown that UPEC activate the caspase 4 murine homologue , caspase 11 , during acute infection in a hemolysin-dependent fashion ( Nagamatsu et . al . in review ) . Despite these similarities , interesting differences existed in the ongoing inflammatory response in mice experiencing chronic cystitis ( S1 Table ) . In C57BL/6J mice , the inflammatory response is immunoglobulin- and cytokine-mediated whereas in C3H/HeN mice , we noted a remarkable absence of upregulated immunoglobulin genes . The increased expression of antimicrobial peptides such as RegIIIγ and the calgranulins ( s100a8 and s100a9 ) is interesting because this increased expression is not sufficient to eliminate bacterial replication during chronic cystitis . Interestingly , C3H/HeN mice that were mock infected exhibited a very similar profile to mice that resolve infection ( Fig . 6A ) . Contrary to C3H/HeN mice , C57BL/6J mice that resolved infection differed significantly from either chronic cystitis or mock infected mice , suggesting an element of altered physiology and immunological memory of the infection ( Fig . 6B ) . This information supports research that serially infecting mice that resolve infection makes them less susceptible to recurrent infection [37] , [63] . What is interesting here is that the mechanisms by which this occurs may differ between mouse strains , and possibly by extension , women .
We have developed models of bacterial superinfection of the urinary tract , which may provide insight into the connection between recent and frequent sexual intercourse and the susceptibility to the development of chronic UTI [5] , [22] . Our results demonstrate that superinfection resulted in increased susceptibility to chronic cystitis in both susceptible and resistant mouse genetic backgrounds , but the time window for priming differed between strains . We have previously shown that chronic cystitis predisposes to severe rUTI upon a subsequent infection weeks to months after clearance of the first infection with antibiotics [37] . Clinically , millions of women take post-coital and prophylactic antibiotics so as not to develop rUTI [64] . Therefore , if clinically applicable , our results detailed here may partially explain why frequent sexual intercourse is such a strong risk factor for UTI . The necessity of prophylactic antibiotics could be obviated if the risk factors and bacterial traits identified here can be altered in the clinical population of women suffering chronic rUTIs . Frequent sexual intercourse is among the most important risk factors for rUTI in young women [22] . Peri-urethral carriage of the causal strain and sexual intercourse immediately precede the development of a rUTI [5] . Sexual intercourse likely introduces mixed populations of bacteria into the urinary tract , with E . coli being the most common [18] . In this environment , UPEC invade bladder tissue and replicate , forming IBCs and bacterial filaments , which have been observed in human urine in 40% of patients suffering acute UTI , 24–48 hours after reported sexual intercourse [21] . These data may provide mechanistic insight as to the frequent clinical observation that recent and frequent sexual intercourse over a brief period of time leads to increased rates of rUTI [23] . Furthermore , elevated levels of serum CSF1 , CXCL-1 , and CXCL-8 in women with acute UTI were associated with a higher rate of rUTI [45] . Using C3H/HeN and C57BL/6J mice , we have shown that superinfection during the period of acute infection dramatically increases the proportion of mice that experience chronic cystitis with inoculations of 107 UPEC ( Fig . 1A and 3A ) . The bacterial characteristics responsible for frequent recurrences are beginning to be assessed [65] . Hemolysin is expressed by 50% of UPEC isolates , but is more likely to be associated with symptomatic UTI [66] . It is possible that hemolysin-mediated exfoliation and caspase 1/11 activation leads to UTI-associated symptoms . In our studies , we found that an increase in priming for chronic cystitis correlated with the bacterial ability to invade and replicate within the bladder tissue ( Fig . 1B–C ) , and through hemolysin to activate caspase 1/11 leading to IL-1 secretion and bacterial replication ( Fig . 1D and 2B ) . Activation of caspase 1/11 has been shown to contribute to epithelial cell death in vitro and exfoliation in vivo in C3H/HeN mice , suggesting that caspase-mediated exfoliation may expose the underlying epithelium upon which UPEC replicates during chronic cystitis ( Nagamatsu et . al . in review ) . Inhibition of caspase 1/11 protected superinfected mice from chronic cystitis ( Fig . 2 ) , suggesting a role for cytokines downstream of caspase activation including IL-1α and IL-1β , identified in our microarray of four-week bladders ( Fig . 6; S1 Table ) . A microarray analysis revealed that in C3H/HeN and C57BL/6J mice , 11/20 of the most upregulated genes during chronic cystitis were the same . Differences between the responses to infection in these mouse strains may result from the dramatic increase in kidney infection or QIR presence in C57BL/6J relative to C3H/HeN mice [37] , [40] . Further , this data supports the hypothesis that a muted inflammatory response to UPEC infection is more likely to lead to resolution [26] . Also , our studies suggest that serum biomarkers such as IL-6 , KC , and G-CSF may predict a predisposition to rUTI ( Fig . 5 ) [37] . Recently , it was demonstrated that cytokines involved in immune cell chemotaxis and maturation ( the human homolog of KC included ) during acute UTI enhanced the likelihood of developing rUTI [45] . We have created mouse models that have identified both bacterial and host immune factors that may predispose women to rUTI . Inhibiting caspase-mediated inflammation or downstream effectors may serve to prevent a UTI from becoming a chronic or recurrent UTI . Further work to identify bacterial and host factors that influence the balance between resolution and chronic infection is required to lead to better treatments clinically . The ability of UPEC to invade bladder tissue allows it to transcend stringent bottlenecks during infection [24] , [25] , [27] . The ability to replicate intracellularly also impacts the ability of a second invading strain to proliferate in the bladder environment ( Fig . 1B–C ) . The molecular basis of bacterial colonization of the bladder during chronic cystitis is an area of active investigation . Previously , it has been shown that mannosides are effective in treating chronic cystitis arguing that FimH-mediated binding plays an important role [67] . It has recently been demonstrated that FimH variation outside of the binding pocket affects protein conformation and pathogenicity of UPEC [48] . This variation may impact bacterial adherence and replication during chronic cystitis . Furthermore , because invasion and intracellular replication appear to influence the likelihood to develop chronic cystitis , treatments with soluble compounds such as mannosides that block the ability of UPEC to invade the tissue or compounds that might alter FimH conformation hold promise as effective means to prevent or treat rUTI [67]–[70] . These analyses may allow us to identify high-risk patients for more aggressive therapy and/or anti-virulence compounds to limit this troubling disease .
All WT bacterial strains utilized were derivatives of UTI89 , including tagged , isogenic UTI89 isolates , kanamycin resistant UTI89 attHK022::COM-GFP , kanamycin resistant UTI89 with re-integrated UTI89 FimH , spectinomycin resistant UTI89 attλ::PSSH10-1 , and chloramphenicol resistant UTI89 [24] , [47] , [71] . FimH mutant strains , ΔompA , Δkps , ΔhlyA , ΔcpxR were all previously published [47] , [51] , [52] ( Nagamatsu et al . in review ) . Bacteria for infection were prepared as previously described [72] . Six to seven week old female C3H/HeN ( Harlan ) or C57BL/6J ( Jackson ) were transurethrally infected with a 50 µL suspension containing 5×106–2×107 CFU of UTI89 or relevant mutant in PBS under 3% isofluorane . Protamine Sulfate ( Sigma ) was dissolved in PBS and caspase 1/11 inhibitor Ac-YVAD-CMK ( BACHEM ) was dissolved in DMSO and transurethrally inoculated into the bladder . At indicated timepoints after infection , mice were anesthetized and infected again . Venous blood was obtained at 24 hpi , just prior to re-infection , by submandibular puncture and centrifuged at max speed at 4°C in Microtainer serum separation tubes ( BD ) and stored at −20°C until use . Cytokine expression was measured using the Bio-Plex multiplex cytokine Group I bead kit array ( Bio-Rad ) , which measures 23 cytokines . Urine was obtained by gentle suprapubic pressure and serially diluted and plated on appropriate antibiotic plates . Mice were sacrificed by cervical dislocation under isofluorane anesthesia , and their organs were aseptically removed . Chronic cystitis was determined if animals had urine titers>104 CFU/mL at 1 , 7 , 14 , 21 dpi and bladder titers>104 CFU at sacrifice at 28 dpi [37] . Animals that resolved infection and had a recurrence or had resolved the infection with reservoir titers>104 CFU were marked in red and considered to have resolved the chronic infection . Organ titers shown are the total bacterial burden . The Washington University Animal Studies Committee approved all mouse infections and procedures as part of protocol number 20120216 , which was approved 01/11/2013 and expires 01/11/2016 . Overall care of the animals was consistent with The guide for the Care and Use of Laboratory Animals from the National Research Council and the USDA Animal Care Resource Guide . Euthanasia procedures are consistent with the “AVMA guidelines for the Euthanasia of Animals 2013 edition . ” C3H/HeN or C57BL/6J mice were infected as discussed above . After 28 days , animals that had developed chronic cystitis , resolved the infection , or aged matched PBS controls were sacrificed for RNA isolation . Upon sacrifice , 5 bladders from each condition were immediately pooled and homogenized in Trizol for RNA isolation according to the manufacture's suggested protocol . DNase treatment was performed to remove any contaminating DNA before submission to the Genome Technology Access Center for sample processing and hybridization on Affymetrix Mouse Gene 1 . 0 chips in triplicate . Data was analyzed using the Partek Genomics Suite . Gene lists were compiled using fdr-ANOVA analysis with a significance cut off of p<0 . 001 . Experiments were repeated twice with a representative analysis shown . Microarray data are available in the ArrayExpress database ( www . ebi . ac . uk/arrayexpress ) under accession number E-MTAB-2930 . Mice were infected as described above . Bladders were aseptically harvested , bisected , and splayed . Bladders were fixed in 2 . 0% glutaraldehyde in 0 . 1M sodium phosphate buffer overnight . Bladders were then washed three times with 0 . 1M sodium phosphate buffer and de-ionized water before being fixed in 1 . 0% osmium tetroxide . Bladders were washed and then critical point drying was performed with absolute ethanol and liquid carbon dioxide . Sputter coating was performed with gold-palladium using a Tousimis Samsputter-2a . Images were obtained on a Hitachi S-2600H operated at 20 kV accelerating voltage . Datapoints below the limit of detection ( LOD ) were set to the LOD for graphical representation and statistical analysis . For cytokine data , values out of the range of the instrument were not included for analysis . Fisher's exact test was utilized to determine differences between groups for rates of chronic cystitis . One-way ANOVA was utilized to determine whether any cytokine differences were apparent and pairwise assessment of median values was determined by Mann-Whitney test . Unless otherwise indicated , p<0 . 05 was considered significant . Analyses were performed in Graphpad Prism 5 . 0 .
|
Urinary tract infections ( UTIs ) affect millions of women each year resulting in substantial morbidity and lost wages . Approximately 1 . 5 million women are referred to urology clinics suffering from chronic recurrent UTI on a yearly basis necessitating the use of prophylactic antibiotics . Frequent and recent sexual intercourse correlates with the development of UTI , a phenomenon referred to clinically as “honeymoon cystitis . ” Here , using superinfection mouse models , we identified bacterial and host factors that influence the likelihood of developing chronic UTI . We discovered that superinfection leads to a higher rate of chronic UTI , which depended on bacterial replication within bladder cells combined with an immune response including inflammasome activation and cytokine release . These data suggest that bacterial inoculation into an acutely inflamed urinary tract is more likely to lead to severe UTI than bacterial presence in the absence of inflammation . Modification of these risk factors could lead to new therapeutics that prevent the development of recurrent UTI .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"animal",
"models",
"of",
"infection",
"bacterial",
"diseases",
"infectious",
"diseases",
"medicine",
"and",
"health",
"sciences",
"women's",
"health",
"medical",
"microbiology",
"microbial",
"pathogens",
"biology",
"and",
"life",
"sciences",
"superinfection",
"microbiology"
] |
2015
|
Uropathogenic Escherichia coli Superinfection Enhances the Severity of Mouse Bladder Infection
|
Filarial diseases represent a significant social and economic burden to over 120 million people worldwide and are caused by endoparasites that require the presence of symbiotic bacteria of the genus Wolbachia for fertility and viability of the host parasite . Targeting Wolbachia for elimination is a therapeutic approach that shows promise in the treatment of onchocerciasis and lymphatic filariasis . Here we demonstrate the use of a biodegradable polyanhydride nanoparticle-based platform for the co-delivery of the antibiotic doxycycline with the antiparasitic drug , ivermectin , to reduce microfilarial burden and rapidly kill adult worms . When doxycycline and ivermectin were co-delivered within polyanhydride nanoparticles , effective killing of adult female Brugia malayi filarial worms was achieved with approximately 4 , 000-fold reduction in the amount of drug used . Additionally the time to death of the macrofilaria was also significantly reduced ( five-fold ) when the anti-filarial drug cocktail was delivered within polyanhydride nanoparticles . We hypothesize that the mechanism behind this dramatically enhanced killing of the macrofilaria is the ability of the polyanhydride nanoparticles to behave as a Trojan horse and penetrate the cuticle , bypassing excretory pumps of B . malayi , and effectively deliver drug directly to both the worm and Wolbachia at high enough microenvironmental concentrations to cause death . These provocative findings may have significant consequences for the reduction in the amount of drug and the length of treatment required for filarial infections in terms of patient compliance and reduced cost of treatment .
Filarial parasites belonging to the family Onchocercidae remain a significant global burden , endemic in over 80 countries worldwide , particularly India and sub-Saharan Africa [1] , and infecting up to 120 million individuals [2] . The major human diseases caused by filaria are Lymphatic Filariasis ( LF ) , River Blindness ( RB ) , and Loiasis . Of the Onchocercidae , Onchocerca volvulus causes RB while LF results from the infections of Wuchereria bancrofti and Brugia malayi . Infection is initiated through arthropod transmission of larvae into the skin of a vertebrate host where they feed , mature into fertile adults and reproduce . Adult worms survive within these tissues for several years shedding thousands of microfilariae ( MF ) that migrate through the skin and lymphatics and distribute throughout the body . Morbidity associated with infection stems from both the persistence of the worm in the skin and the lymphatics that creates scarring and fluid accumulation and the MF causing debilitating chronic dermatitis due to irritation of MF antigen [3] . Resident within W . bancrofti and B . malayi parasites at all stages of life are the endosymbiotic bacteria , Wolbachia . These obligate intracellular bacteria are members of the Rickettsialaes order of α-proteobacteria [4] that also contain the mammalian pathogens in the genera Rickettsia , Brucella , Bartonella , Ehrlichia and Anaplasma [5] . Typically , Wolbachia are present within hypodermal cells of the lateral cords of both sexes , however , vertical transmission of the bacteria to MF is facilitated by the presence of bacteria in the ovaries , oocytes and embryos during development in females [6] . Within the last decade increasing concerns have arisen regarding the ability to effectively control and eradicate current infections of the filarial endoparasitic worms that cause onchocerciasis and LF [7] . It is estimated that 120 million people are currently suffering from LF and more than one billion individuals in 73 countries are at risk of developing LF [8] . In 2000 , the WHO initiated a call to eliminate LF as a public-health problem by 2020 [9] . Significant steps have been taken towards this eradication effort but have been mainly limited to the use of mass drug administration of a microfilaricide to interrupt transmission and diminish morbidity . This approach has reduced the reoccurrence in some countries by 46% over the last 13 years [2 , 9] . Drugs with the greatest efficacy toward LF have been limited to trials of annual , bi-annual , single-dose , multiple dose , and combinations of either ivermectin or diethylcarbamazine with albendazole [7 , 10] . Recently , antimicrobial treatments such as doxycycline have been added to antifilarial regimens to target the symbiotic intracellular bacteria Wolbachia [11 , 12 , 13] . Anti-Wolbachia drugs have been shown to reduce the pathogenicity and reproductive capacity of adult filarial worms [14] . The limitations associated with the use of the above drugs are four-fold: specific restrictions on patient age and health status due to cytotoxicity; the inability to reach the deep tissues where the adult filarial nematodes reside; counter-indications in areas endemic for the Loa loa parasite; and poor patient compliance [15] . These limitations have created an urgent need to seek new technologies that can be used to dramatically improve therapy protocols . In the parasite , Wolbachia are vertically transferred from the ovaries of the adult into embryos that gestate and are shed as MF into the surrounding tissue [16] . In the adult female , the bacteria contribute to reproduction and embryo development . Treating adult worms in vitro and in vivo with tetracyclines , including doxycycline , to eliminate Wolbachia has been shown to halt embryogenesis , reduce the amount and viability of MF shed [14 , 17 , 18] in tissues , and lead to death of the adult worm in 6 to 12 months [12 , 19] . Loss of viability in the adult worm through anti-Wolbachia therapy is precipitated by loss of the endosymbiont , triggering apoptosis and tissue disruption [20] . Another antibiotic that is effective against Rickettsiae is rifampicin , which is better tolerated in children and pregnant patients , is able to kill Wolbachia in Onchocerca parasites in vitro [21] . However , rifampicin is less effective than doxycycline in vivo [18 , 22] . Additional therapeutic options for doxycycline are to combine it with other current antiparasitic drugs; however combination therapies present difficulties for drugs that have different half-lives , for widespread drug administration programs that suffer from poor patient compliance , and for long treatment regimens that require multiple years of therapy [23] . Amphiphilic polyanhydride nanoparticles have been studied extensively as carriers for drugs and vaccines [24 , 25 , 26 , 27 , 28] . These biodegradable materials degrade into dicarboxylic acids upon scission of the anhydride bond , rendering them highly biocompatible . These materials are generally hydrophobic and this confers a surface erosion mechanism to devices made of polyanhydrides [29 , 30] . The advantage of such a surface erosion mechanism is that payloads can be released in a sustained and predictable manner , providing a so-called zero order release profile , in which the release rate of the payload is constant . Such release profiles enable single dose therapies , maximize the time over which the in vivo drug concentration is maintained within the therapeutic window , and enhance patient compliance . This nanoscale platform also has the capability to be a delivery system for combination therapies that carry payloads to hard to reach tissues , thereby offering the potential to increase the efficacy of the drug-nematode interaction . The ability of the polyanhydride nanoparticles to slowly erode and release the cargo molecule in a controlled manner can also allow for specificity against both the adult nematode and Wolbachia [11] . In this work , we demonstrate that co-delivery of doxycycline and ivermectin through encapsulation into polyanhydride nanoparticles effectively kills adult B . malayi filarial worms 8-fold faster with up to a 4 , 000-fold reduction in the amount of drug used . We hypothesize that the mechanism behind this enhanced killing of the macrofilaria is the ability of the nanoparticles to penetrate the outer membrane of the B . malayi worm and effectively deliver drug directly to the worm and its symbiotic bacteria Wolbachia at high enough microenvironment concentrations to cause death . The approach described herein holds great promise to interrupt the life cycle of the nematode not only by reducing the number of MF , but also by improving macrofilaricidal activity . These findings have the potential to dramatically enhance the treatment of patients suffering from debilitating filarial diseases such as LF and RB .
Live adult B . malayi females , males , and microfilariae were acquired from the NIH/NIAID Infectious Disease Filariasis Research Reagent Repository Center ( FR3 ) at the University of Georgia ( Athens , GA ) . Adult worms were maintained in non-phenol red Roswell Park Memorial Institute ( RPMI ) 1640 medium supplemented with 10% heat-inactivated fetal bovine serum ( FBS ) and 1% penicillin-streptomycin . The B . malayi were held in an incubator at a temperature of 37°C supplemented with 5% carbon dioxide ( CO2 ) . Female and male worms were stored individually in 48 well microtiter plates containing 1 mL of RPMI-1640 . Previously shed microfilariae were housed in 50 mL conical tubes containing 25 mL of RPMI-1640 . Upon arrival of the filarial worms they were separated and placed into individual 48 well microtiter plates that contained the previously described media . The motility of the adult worms treated with various formulations was observed for 30 seconds by utilizing a 2X objective on a Nikon Microscope and scored utilizing the following 0–5 scoring system: zero percent motility reduction , head and tail uninhibited as well as mid-section unaffected = 5; 1 to 25% motility reduction , ability to visualize both the head and tail movements easily as well as midsection = 4; 26 to 49% motility reduction , showing a partial mid-section paralysis = 3; 50 to 74% motility reduction , showing full mid-section paralysis followed by a substantial reduced movement in the head and tail = 2; 75 to 99% motility reduction , showing full mid-section paralysis followed by either head and or tail paralysis but not limited to occasional movement over a 30 second time period = 1; and a score of 0 represented 100% death ( i . e . , non-motile ) . The loss of motility of the adult worms was compared with that of respective untreated controls . The in vitro release kinetics of ivermectin and doxycycline from 20:80 CPTEG:CPH nanoparticles was determined using high performance liquid chromatography ( HPLC ) . Nanoparticles were suspended in 1 mL of phosphate buffered saline ( PBS , pH 7 . 2 ) , followed by thorough sonication to suspend the particles and incubated at 37°C under constant shaking . At each time point , the tubes were centrifuged at 10 , 000 rcf for 10 min . The supernatant was removed and replaced with fresh PBS and returned to incubation . The ivermectin was separated using a Zorbax C8 4 . 6 x 250 cm chromatography column using a mobile phase of tetrahydrofuran , acetonitrile , and water in a 40:40:20 volume ratio , respectively . The flow rate was set at 1 mL/min and ivermectin eluted at a retention time of 5 . 3 min . The ivermectin was quantified with fluorescent detection using an excitation wavelength of 365 nm and an emission wavelength of 475 nm with a gain of 10 and attenuation of 64 [33] . The doxycycline was quantified using a Varian Pursuit XRs 3 C18 150 x 2 mm column . The mobile phase consisted of water with 0 . 1% formic acid and acetonitrile at a 98:2 volume ratio and at a flow rate of 0 . 2 mL/min . After one minute the mobile phase was ramped to 100% acetonitrile over 4 min . Doxycycline was quantified with tandem mass spectroscopy ( MS/MS ) using the following conditions: Q1 to Q3 transitions of 445 to 154 at 30 V collision energy and 445 to 427 . 9 at 16 V collision energy [34] . The encapsulation efficiency of the particles was determined as described previously [35] and found to be approx . 100% for both drugs and for the rhodamine . Ivermectin ( 22 , 23-dihydroavermectin B1 ) and doxycycline hyclate ( C22H24N2O8 · HCl · 0 . 5H2O · 0 . 5C2H6O ) were dissolved in DMSO at a final concentration of 0 . 02% v/v . RPMI-1640 medium was prepared such that the final concentrations of each drug in nanoparticles were 195 , 49 , 10 , 5 , 1 . 95 , 0 . 049 , 0 . 01 , 0 . 005 , and 0 . 001 μM , respectively . Control medium contained 0 . 02% DMSO and rhodamine B , but no drug . Individual worms that were previously placed into 48 well flat bottom culture plates upon arrival had new fresh media that contained 1 mL of RPMI-1640 , 0 . 01% streptomycin and penicillin and 10% heat-inactivated fetal calf serum . B . malayi were incubated at 37° C and 5% CO2 for 1–2 hours for acclimatization to evaluate base line motility before the experiment began . The nanoparticle-encapsulated drugs and standard antifilarials ( i . e . , free drug with no nanoparticle encapsulation ) were added as previously described above . For example , 2 mg of the 195 μM drug concentration nanoparticles were added to each well and correspondingly lower amounts of nanoparticles were added as the drug concentrations decreased . Controls received equal amounts of media but lacked the nanoparticles and standard antifilarials . B . malayi were incubated at 37°C and 5% CO2 and monitored for 14 days post treatment . Reversibility assays where conducted 24 h following a 9 day course of treatment and or after the final individual within the soluble group reached a motility score of 0 . Individual worms where transferred to a new 48 well microtiter plate that contained 1 mL of the previously described media and were placed into an incubator at a temperature of 37°C supplemented with 5% carbon dioxide ( CO2 ) . After 24 h a motility score was recorded . Upon completion of reversibility adult female worm viability was assessed utilizing a standard 3- ( 4 , 5-dimethylthiazol-2-yl ) -2 , 5-diphenyltetrazolium bromide ) ( MTT ) -formazan colorimetric assay . The same female worms used in the motility assay were gently blotted and transferred to a 96 well microtiter plate that contained 0 . 2 mL of 1mg/mL MTT in phosphate-buffered saline ( pH 7 . 2 ) and incubated for 2 h at 37°C . The formazan formed was extracted in 0 . 15 mL of DMSO for 4 h at 23°C and the absorbance was measured at 570 nm using a spectrophotometer ( FLUOstar Omega version 1 . 01 ) . The mean absorbance of the formazan from the treated worms was compared with that of the controls . The viability of the treated worms was assessed by calculating the percent inhibition in motility and MTT reduction over the DMSO-treated control worms . The effect of treatments on MF shedding and subsequent motility was assessed using two experimental approaches . The direct effect on MF was performed by collecting MF from the spent media of healthy untreated female worms . Spent media was centrifuged for 10 min at 200 rcf , the MF were re-suspended in fresh media and enumerated by counting using a 14 . 80 mm² grid plate . Aliquots containing at least 250 MF were dispensed into wells of a 96 well plate and treated in triplicate with either soluble drug , blank nanoparticles or nanoparticles containing drugs as indicated . Wells were then monitored and scored daily for MF movement . Motility of MF scoring was carried out using a 10X Nikon microscope and individual scores from five separate fields of view ( FOV ) were averaged to yield a motility score for each well . Scoring was adapted from a previously reported method by observing the movements of the anterior and posterior body over one min [36] . Data were analyzed by ANOVA and graphs were assembled using GraphPad Prism ( v6 . 05 ) . In separate experiments , the effect of treating adult female Brugia and their ability to release MF was monitored and recorded for 14 days . The motility of the shed MF was also recorded . Each experiment , the various treatment groups , and the different doses of the drugs were repeated for a minimum of three times . To understand the interaction between the nanoparticles and the filarial worms , recently deceased adult worms ( motility score of “0” ) from treated and untreated experimental groups were fixed and prepared for confocal microscopy . Worms were removed from treatment wells , washed twice in pH 7 . 4 PBS and fixed with 4% paraformaldehyde for 1 h at room temperature . Worms were removed , rinsed with fresh PBS and mounted carefully onto glass slides using Prolong Gold with DAPI ( Life Technologies ) . Slides were then stored in the dark at 4°C prior to imaging . Imaging was performed using an Olympus IX-71 laser-scanning confocal microscope ( Fluoview 2000 , v2 . 0 ) equipped with 405 nm , 488 nm , 559 nm and 635 nm lasers , spectral photomultiplier detection and a piezo-controlled stage . Laser power and exposure settings were set according to the negative ( worms with no rhodamine ) and positive ( rhodamine containing nanoparticles ) and maintained constant over all the experiments . Image stacks were collected simultaneously for bright field ( transmitted light—grey scale ) , cell nuclei ( DAPI—blue ) and nanoparticles ( rhodamine—red ) . Bright field and cell nuclei were used to identify various tissues , organs and structures in the worm such as oral opening , pharynx , uterus , the nerve ring , extracellular secretory apparatus , esophageal track and anal pore . Rendering of Z-stack data in three dimensions and minimal image processing was performed as described previously by our laboratory [37] using ImageJ v 1 . 49a ( NIH ) . Software parameters for image processing remained constant for all image/data sets presented .
After encapsulation of the anti-filarial drugs by nanoprecipitation , the particle morphology and size were examined using scanning electron microscopy . Surface morphology was found to be consistent with previous work and the scanning electron photomicrographs are shown in Fig 1 ( panel A ) [26 , 27 , 38] . Additionally , particle size was also found to be consistent with previous work [37 , 39] . The mean diameter of the drug-containing 20:80 CPTEG:CPH nanoparticles was 218 ± 56 nm . Using a HPLC assay as described in the Methods section , the release kinetics of ivermectin and doxycycline from the nanoparticles was quantified , as shown in Fig 1 ( panel B ) . The 20:80 CPTEG:CPH nanoparticle formulation provided sustained release of both drugs over one week . That data showed that the doxycycline release profile from the hydrophobic 20:80 CPTEG:CPH nanoparticles showed a small burst effect , which is consistent with previous work [40] . It was also observed that there was a much larger burst release of doxycycline as compared to that of ivermectin . The effectiveness of the polyanhydride nanoparticle-based delivery platform was compared to that of the standard soluble treatment in terms of the overall percent survival of B . malayi over the duration of the study with a single treatment of the previously described groups ( Figs 2 and 3 , panel A ) . The effect of blank ( i . e . , no drug ) nanoparticles containing only rhodamine was assessed with 2 mg/mL of nanoparticles , the amount corresponding to that of the highest drug concentration administered ( i . e . , 195 μM ) . Similarly , untreated controls did not contain active drugs , but were supplemented with rhodamine and DMSO corresponding to the amounts present in the highest treatment group ( i . e . , 195 μM drug concentration ) . The motility of the worms treated with the blank nanoparticles did not vary significantly from that of the untreated worms for either female or male worms ( Figs 2 and 3 and S1 Fig ) . The results clearly demonstrate the effectiveness of the nanoparticle formulations for all the treatment groups when compared to the soluble treatment group for drug concentrations as low as 1 nM . Another parameter to quantify the difference between the effectiveness of the soluble dual ivermectin/doxycycline treatment and that of the nanoparticle-based delivery platform was the average number of days to death . We conducted a fourteen-day in vitro study that quantified the average days to death of female ( Fig 2 , panel B ) and male ( Fig 3 , panel B ) B . malayi . Random testing of viability of worms exhibiting a motility score of 0 for metabolic activity was performed using the MTT reduction assay throughout the studies and for all treatment groups and a representative of % MTT inhibition analysis is shown in S2 Fig for the 5 μM drug concentration . Inhibition of metabolic activity greater than 75% of that of untreated controls indicated that worms were non-viable and declared dead . In addition , separate worms with a motility score of 0 were subjected to reversibility testing whereby after 24 h of no motility , worms were removed from the test media , washed three times and placed in fresh media with no drugs . Recovery of movement was interpreted as the worms being temporarily paralyzed and those worms were removed from the ATD , except where indicated† ( Figs 2 and 3 , panel A ) . We observed reversible paralysis only when high doses of soluble drug ( 49 μM—195 μM ) were used in approximately 25% of all worms treated in these groups . Strikingly , reversible paralysis was not observed in any nanoparticle-treated group of any drug concentration , showing that the nanoparticle formulations effectively killed the worms . The data indicated that the nanoparticle-based delivery platform significantly ( p < 0 . 001 ) lowered the overall time to death when compared directly to the soluble dual ivermectin/doxycycline treatment . Treatment of female worms with 195 μM of the soluble drugs resulted in an average time of death in excess of nine days and 63% of the worms were killed ( Fig 2 , panels A and B ) . In contrast , treating the worms with the nanoparticle-based delivery system with the same drug concentration sharply reduced the average time to death to less than 1 . 2 days , and more significantly , 100% of the worms were killed . We also performed dose titration studies and systematically lowered the concentration of the two drugs from 195 μM to 1 nM . Female worms treated with the 1 nM concentration of soluble ivermectin/doxycycline resulted in an average time to death that was >14 days , and more significantly , only 33% of the worms were killed at this concentration ( Fig 2 , panel A ) . In contrast , the worms treated with the nanoparticle-based delivery system containing the 1 nM drug concentration had an average time to death of <6 days , and more significantly , 100% macrofilarial death was observed upon treatment with the low concentration of ivermectin/doxycycline . To determine if nanoparticles changed the reproductive capacity of female worms the shedding of MF was observed for fourteen days ( Fig 2 , panel D ) . Both the soluble ivermectin/doxycycline and the nanoparticle-based delivery groups significantly reduced the overall microfilariae shed at the high drug concentration of 195 μM . But as the concentration of ivermectin/doxycycline was systematically reduced to 5 nM , significant differences ( p<0 . 01 ) were observed in the MF shed by the worms treated with the soluble ivermectin/doxycycline groups compared to the worms that were treated with the nanoparticle-based delivery system , with the nanoparticles being more effective at lower drug concentrations . The effect of the various drug treatments on the motility of B . malayi MF was observed for 14 days ( Fig 4 ) . Similar to the effects observed for adults , the average time to death as well as the motility of the MF treated with the soluble drug was significantly ( p < 0 . 001 ) higher in comparison to that of the MF that were treated with the nanoparticle-based delivery system , especially at low drug concentrations . For the B . malayi males ( Fig 3 , panels A and B ) treatment with drug-loaded 20:80 CPTEG:CPH nanoparticles showed a significant difference ( p<0 . 05 ) in terms of survival when compared to the survival of worms treated with an equivalent soluble dose of ivermectin/doxycycline . In terms of average time to death , male worms treated with a 195 μM concentration of soluble ivermectin/doxycycline died after seven days and more significantly , only 30% of them were killed at this dose ( Fig 3 , panel A ) . In sharp contrast , all the male worms treated with the nanoparticle-based delivery system containing the same dose of ivermectin/doxycycline died in less than one day . When the treatment concentration was reduced to 0 . 049 μM , the average time to death of worms treated with soluble ivermectin/doxycycline was >14 days with a death rate of only 20% . On the other hand , all the male worms treated with the nanoparticle-based delivery system containing the 0 . 049 μM concentration of ivermectin/doxycycline died within ten days . In conjunction with the average time to death assay we also monitored the motility over time to characterize the overall effectiveness of the nanoparticle treatments ( panel C of Figs 2 and 3 ) . With both male and female B . malayi worms , an exponential reduction in overall motility was observed in the worms administered the nanoparticle-based delivery treatments in comparison to that of the worms that received the soluble drug treatment . To visualize the interactions of polyanhydride nanoparticles with B . malayi female worms , animals were treated with equivalent amounts of a cocktail containing rhodamine red dye with ivermectin/doxycycline either in soluble form or encapsulated within the 20:80 CPTEG:CPH nanoparticles . At regular intervals worms were washed three times in PBS to remove surface-bound nanoparticles , fixed , and then imaged using confocal microscopy to localize nanoparticles within tissues . Detection of focal , intense rhodamine staining is consistent with size and staining patterns of intact nanoparticles . In comparison , diffuse red staining is indicative of free rhodamine . Worms treated with 195 μM of the soluble drug cocktail did not accumulate rhodamine within tissues over a period of 96 hours ( Fig 5 , panel A ) . Over this same time interval , the much lower dose of 5 μM of ivermectin/doxycycline encapsulated within the nanoparticles that also contained rhodamine , extensively labeled the interior structures of female worms ( Fig 5 , panel B ) . Diffuse red staining reveals the rapid distribution of the released dye throughout the body of the worm and focal , intact 20:80 CPTEG:CPH nanoparticles are easily discernable deep within worm tissues by confocal microscopy ( Fig 5 , panels B and C ) . Three-dimensional reconstruction of confocal Z-stacked images revealed the penetration of nanoparticles and release of rhodamine throughout the inner tissues of the worm ( Fig 5 , panel C ) . Similar rhodamine and nanoparticle staining patterns were evident throughout the length of the treated worms at all concentrations tested . Internal staining of worms treated five hours post treatment with 195 μM of the soluble drug cocktail were identical to the rhodamine intensity of worms treated with 5 μM of the drug cocktail encapsulated within the nanoparticles for 96 hours . These observations offer insights into why the nanoparticle-based delivery treatments of B . malayi are highly effective , especially at low doses .
Filarial diseases represent a significant social and economic burden in areas that are endemic with the filarial endoparasite B . malayi , and its symbiotic bacteria Wolbachia . We report the use of a polyanhydride nanoparticle-based drug delivery platform for the co-delivery of the antiparasitic drug , ivermectin , to reduce macro and microfilarial burden and the antimicrobial , doxycycline , to eliminate the symbiotic Wolbachia . The co-delivery of doxycycline and ivermectin in the context of polyanhydride nanoparticles effectively killed adult female and male B . malayi filarial worms with up to a 4 , 000-fold reduction in the amount of drug used . Further , the time to death of the macrofilaria was significantly reduced when the anti-filarial drug cocktail was delivered in the context of the polyanhydride nanoparticles . Confocal microscopy experiments suggest that the polyanhydride nanoparticles behave like a Trojan horse and penetrate the outer membrane of the worm cuticle . The nanoparticles effectively deliver the drugs at high enough microenvironment concentrations to vital areas within the worm , thus significantly enhancing their effectiveness at killing the worms . These findings may have significant consequences for the reduction in the amount of drug and the length of treatment required for filarial infections and provide a paradigm-changing technology to combat filarial disease .
|
Infection with the filarial endoparasites Brugia malayi and its symbiotic bacteria Wolbachia represent a significant burden to both humans and animals . Current treatment protocols include use of multiple drugs over a course of months to years , resulting in high costs , undesirable side effects , and poor patient compliance . By encapsulating two of these drugs , ivermectin and doxycycline , into biodegradable polyanhydride nanoparticles , we report the ability to effectively kill adult B . malayi with up to a 4 , 000-fold reduction in the amount of drug used . These results demonstrate a promising role for the use of nanoscale drug carriers to reduce both the course of treatment and the amount of drug needed to increase affordability of lymphatic filariasis treatment and enhance patient compliance .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
Polyanhydride Nanoparticle Delivery Platform Dramatically Enhances Killing of Filarial Worms
|
Age-related defects in stem cells can limit proper tissue maintenance and hence contribute to a shortened lifespan . Using highly purified hematopoietic stem cells from mice aged 2 to 21 mo , we demonstrate a deficit in function yet an increase in stem cell number with advancing age . Expression analysis of more than 14 , 000 genes identified 1 , 500 that were age-induced and 1 , 600 that were age-repressed . Genes associated with the stress response , inflammation , and protein aggregation dominated the up-regulated expression profile , while the down-regulated profile was marked by genes involved in the preservation of genomic integrity and chromatin remodeling . Many chromosomal regions showed coordinate loss of transcriptional regulation; an overall increase in transcriptional activity with age and inappropriate expression of genes normally regulated by epigenetic mechanisms was also observed . Hematopoietic stem cells from early-aging mice expressing a mutant p53 allele reveal that aging of stem cells can be uncoupled from aging at an organismal level . These studies show that hematopoietic stem cells are not protected from aging . Instead , loss of epigenetic regulation at the chromatin level may drive both functional attenuation of cells , as well as other manifestations of aging , including the increased propensity for neoplastic transformation .
Somatic stem cells replenish many tissues throughout life . In general , they have slow turnover and reside in specialized niches , protected from the environment , so that only a few are activated at a time . Thus , stem cells are a defense against aging , replacing cells lost through attrition . If the rejuvenating effect of stem cells were perfect , senescing cells would be replaced indefinitely; but even in highly regenerative tissues such as the skin , the gut , and the hematopoietic system , age-related decline in function is well established [1] . Still unclear are the effects of aging on the stem cells themselves , which could contribute to inferior tissue repair . Hematopoietic stem cells ( HSCs ) continuously replenish the blood and immune system throughout life . Data from mice support an age-related decline in stem cell function [1] , suggesting that older HSCs are inadequate to cope with the demands of blood production . When limited numbers of aged hematopoietic progenitors are transplanted into young recipients under competitive conditions , they show an overall reduction in long-term repopulating potential [2]; in particular , lymphopoiesis is deficient , whereas myelopoiesis is enhanced [3 , 4] . Paradoxically , however , the total number of primitive progenitors has been reported to increase with age in the C57Bl/6 mice [2 , 5] . A recent study of aged hematopoietic stem and progenitor cells suggested that increased expression of particular proto-oncogenes may underlie some of these observed changes [4] . Although the previous studies varied widely , the findings provide compelling evidence for major age-related alterations in HSC function . To gain insight into the molecular mechanisms that underlie these deficits , we examined gene expression in HSC as a function of age on a genome-wide scale in normal and an early-aging p53 mutant strain . These data provide a comprehensive molecular portrait of aging in HSC , and show that stem cell aging mirrors the aging of other tissues , marked by a dramatic inflammatory response , stress responses , and substantial alterations in the regulation of chromatin structure .
The number of whole bone marrow ( WBM ) progenitors defined by cell surface markers in C57Bl/6 mice increases with age relative to total WBM cellularity [2 , 5] . To assess whether this property extends to HSCs purified using the side population ( SP ) cells defined by their ability to efflux Hoechst 33342 dye [6] , we examined the SP cells in C57Bl/6 mice ranging from 2 to 21 mo of age . The results demonstrate a 9-fold increase in the number of SP cells with age ( Figure 1A ) , with the most primitive SPlow cells [7] showing the greatest increase . These SP cells also exhibit the surface phenotype of HSC regardless of age ( c-kitpos , lineageneg , and Sca-1pos [SParKLS] ) , consistent with their high degree of purity and homogeneity ( Figures 1B and S1 ) . Aged HSCs were uniformly CD48neg , which is one of the markers recently described to mark differentiated hematopoietic cells from both young and aged mice [8 , 9] ( unpublished data ) . Thus , murine HSCs defined by multiple phenotypes increase 9-fold in WBM over approximately 2 y . The increase in cell number was not a result of greater proportion of S-phase HSCs , as determined by propidium iodide staining ( Figure 1C ) , suggesting an alternative mechanism for the increase in HSC number . Limiting dilution bone marrow transplantation can measure the ability of HSCs to reconstitute recipients under competitive conditions and the functional purity of HSC [10] . HSCs ( SParKLS ) were therefore purified from either 2- or 21-mo-old mice and transplanted into lethally irradiated recipients , along with competitor bone marrow . Progeny from donor HSC were distinguished from competitor and recipient cells using the CD45 allelic system [11] . The proportion of peripheral blood progeny derived from purified young and old HSC was monitored at 4 , 8 , and 16 wk post-transplantation . Four weeks after transplantation , there was little difference in the contribution of young versus old HSCs ( Figure 1D ) , but at 8 and 16 wk post-transplant , the contribution from the old HSCs was significantly reduced , but still multilineage ( Figure S2 ) . This finding argues that HSCs acquire a defect in long-term , but not short-term , repopulating potential with increasing age . This deficit represents roughly a 3-fold loss in functional activity per purified stem cell . With a 5- to 10-fold numerical increase in HSC , this indicates that the total stem cell activity remains fairly constant with age , which is consistent with other reports [2 , 12] . To identify transcriptional changes in aged HSCs that correlate with the observed functional deficit , we examined the expression of more than 14 , 000 genes , using Affymetrix MOE430a microarrays and HSCs purified from 2- , 6- , 12- , and 21-mo-old mice . A quadratic trend line ( parabola ) was fit for each gene over the 19-mo test period , which showed that the genes generally either increased or decreased in expression in a time-dependent fashion . We used a linear contrast model based on the entire observation course to determine which genes had the largest changes in expression over time . This revealed 1 , 600 genes that were up-regulated at 21 mo ( “Up-with-Age” group ) , and 1 , 500 that were down-regulated ( “Down-with-Age” group ) , which is summarized as a heat map in Figure 2 . A small hand-picked list is shown in Table 1; the entire list of differentially expressed genes is supplemented in Tables S1 and S2 , and a searchable database of all genes on the array can be found at http://rd . plos . org/pbio . 0050201 . Expression changes of a subset of these genes were validated by real-time quantitative PCR in duplicate on independently purified HSC ( Figure S3 ) . We also compared transcriptional profiles for WBM versus HSCs to identify HSC-specific transcripts; surprisingly , only a modest overlap of genes was found with those that were up-regulated or down-regulated with age , suggesting that the HSC-specific transcriptional programs remain relatively stable as the organism ages ( Figure 2 ) . A remarkable overlap was found between genes up- and down-regulated with age in this study and a previous study of HSC aging , with the top ten genes being identical [4] . We next sought to identify biological processes that were enriched in age-induced or age-repressed genes , compared to chance alone . For this purpose , we used Gene Ontology ( GO; http://www . geneontology . org ) to group genes on the basis of a particular biological process [13] , and identified GO categories that were enriched with statistical significance by a method previously reported [14] . When applied to the Up-with-Age gene list , the analysis revealed a large number of enriched categories that have been linked to aging in general , such as NO-mediated signal transduction , the stress response ( protein folding ) , and the inflammatory response , whereas categories enriched for Down-with-Age genes often included those involved in the preservation of genomic integrity , such as chromatin remodeling and DNA repair ( Figure 3A ) ( the entire GO results can be found at http://rd . plos . org/pbio . 0050201 ) . A link between aging and inflammation has been demonstrated in several vertebrate models and in humans [15] , and we found evidence for the age-dependent regulation of several stress-related genes in HSCs . One of the most highly up-regulated of these genes expresses P-selectin , a cell surface adhesion molecule that serves as a marker for physiological stress states , including inflammation [16] , aging [17] , and cardiovascular disease [18] . P-selectin expression in HSCs , was of particular interest because it mediates the leukocyte–vascular endothelium interaction important for leukocyte extravasation [16] and therefore has implications for HSC migration . Flow cytometric analysis demonstrated increasing levels of P-selectin on the surface of HSCs isolated from 24- to 28-mo-old mice ( 21%–81% , Figure 4A ) , in contrast to scant levels ( 3% ) on HSCs from young mice . We hypothesized that the p65 isoform of NF-κB , which transcriptionally regulates P-selectin [19] would be activated in aged HSCs . To test this , we purified HSCs from 2- and 22-mo-old mice , and examined them for p65 localization by immunofluorescence . In contrast to only 3% of 2-mo-old HSCs , 71% of 22-mo-old HSCs showed enhanced nuclear localization of p65 protein ( Figure 4B and 4C ) . These results implicate NF-κB activation as the mechanism of increased P-selectin expression in aged HSCs , most likely reflecting a time-dependent rise in inflammation . The time course of data allowed us to examine the timing of changes in age-regulated gene expression . We determined when the trend line for each given gene achieved half its maximum change over the full time course ( T1/2-max ) , then grouped the genes by GO category and plotted the results for those categories that had a significant enrichment in the previous analysis ( Figure 3A ) , creating a GO-timer . As shown in Figure 3B , genes that participate in NO-mediated signal transduction were the first to be up-regulated during HSC aging , followed closely by those contributing to the stress response and the regulation of lymphocyte proliferation . Inflammatory-response genes were not activated until late in the aging process , after up-regulation of NF-κB signaling , strengthening our hypothesis that inflammation exerts a strong influence on HSC aging through stimulation of the NF-κB pathway . Complete GO-timer results can be found at http://rd . plos . org/pbio . 0050201 . In Saccharomyces cerevisiae , the chromatin regulatory factor Sir2 , a NAD-dependent histone deacetylase , suppresses recombination and silences transcription at multiple genomic loci [20]; its loss is associated genomic instability and aging . Since genes involved in chromatin remodeling and transcriptional silencing were excessively down-regulated in our GO enrichment analysis , we predicted global dysregulation of transcriptional activity . We reasoned that this would be evidenced by finding regions of chromosomes in which genes that were physically clustered together changed coordinately with age . To test this idea , genes were ordered by their chromosomal position , and age-induced and age-repressed genes were mapped using a density-based statistical approach . The result was a single curve across each chromosome , with peaks representing regions of coordinate up-regulation , and valleys regions of coordinate down-regulation ( Figure 5A ) . Chromosomal loci with significant coordinate changes in gene expression were identified as centers of regulated expression ( COREs; Figure 5A , red lines ) . Using this method , we found more than 100 such COREs among the 19 mouse autosomes ( Table S3 ) . Importantly , there were twice as many CORE peaks as there were valleys , indicating a predominance of a loss of transcriptional silencing throughout the genome . Two recent linkage analyses in BXD mice indicated 22 chromosomal regions , or quantitative trait loci ( QTLs ) , containing genes regulating hematopoietic progenitor number and organism lifespan [21 , 22] . We therefore asked whether any of the 17 mapped QTL loci might correspond to COREs . Eight of the QTLs mapped to seven COREs on Chromosomes 2 , 4 , and 7 ( Figure 5C ) , consistent with the possibility that the QTL-associated genes on these chromosomes may play a critical role in regulating lifespan and HSC number . When compared to other non–HSC-related QTL datasets obtained from WebQTL [23] , we observed a much greater overlap with the Henckaerts , et al . dataset [22] ( p-value = 0 . 06 ) . With such dramatic changes in chromatin regulation , we expected to find inappropriate expression of specific genes in aged HSC . Indeed , with age , we observed a striking increase in expression of the immunoglobulin ( Ig ) heavy chain gene ( IgH; ∼37-fold ) and the immunoglobulin kappa ( IgK ) light chain gene ( ∼50-fold ) . Ig genes are not normally active in HSC , but become highly expressed in B cells . We thus decided to examine age-related transcription of the highly studied IgK locus in more detail . The possibility that B cells were contaminating the purified HSCs was excluded using antibodies against two markers of early and mature B cells , CD19 and Il-7r ( Figure S4A ) . In addition , a PCR-based assay to detect Ig recombination [24] showed no evidence of V ( D ) J recombination ( Figure S4B ) . Transcripts from an unrecombined IgK locus , termed germline ( GL ) transcripts because they are removed during V ( D ) J recombination , have been sighted previously in a pre–pro-B cell line [25]; however , this occurs only after epigenetic modification , including DNA demethylation and subsequent histone modification [26] . Using reverse transcriptase PCR ( RT-PCR ) to detect GL IgK transcripts in poly-A RNA extracted from HSC from 2- and 24-mo-old mice , GL transcripts were readily detectable in HSC from old mice , but were absent in HSC from young mice ( Figure 5D ) . Sequencing confirmed their GL structure , which included the entire first joining minigene ( J1 ) and the constant region with a poly-A tail . These data are consistent with a loss of epigenetic regulation at the IgK locus with age in aged HSC . Such epigenetic changes may enable access by NF-κB and other factors that drive IgK gene expression , leading to inappropriate expression in aged HSC . The tumor suppressor p53 is activated by stress and has been implicated in the balance between longevity and tumor formation [27] . Mutant mice expressing a truncated allele of p53 ( m allele ) display an augmented p53 response , several early-aging phenotypes , and hematopoietic defects , including a blunted HSC proliferative response [28] and a lack of sufficient stem cell potential throughout life to maintain longevity [29] . In contrast , mice with half the amount of p53 ( p53+/− ) would be expected to live longer than wild-type ( WT ) mice , except that they develop lethal tumors . Consistent with the role of p53 in regulating cell proliferation , the p53+/− mice exhibit increased stem cell proliferation [29] . We examined HSC from p53+/m and p53+/− HSC to gain insight into the role of stem cells in organismal aging . HSC from 12-mo-old p53+/m and p53+/− mice were isolated and their gene expression profiles determined as above . These two strains studied constitute the available extremes of p53 activity; the p53+/− mouse has half the p53 dose of WT mice , and the p53+/m allele mice have more p53 activity than WT . A pairwise comparison between the p53+/m and the p53+/− HSC data ( Tables S4 and S5 ) confirmed up-regulation of the GO categories DNA repair , Response to DNA damage stimulus , and Apoptosis in the p53+/m HSC ( Figure 6A , Tables S6 and S7 ) . Using the trend lines from the WT HSC time course , we calculated for each gene a predicted age ( in months ) based on the level of expression in both the p53+/m and p53+/− 12-mo-old mice . Interestingly , many of the genes from Table 1 exhibit a “younger” expression pattern in the p53+/m mice , including up-regulated genes such as Icam-1 ( −4 . 3 mo ) , CatnB ( −3 mo ) , and Dnaja2 ( −2 . 8 mo ) , as well as down-regulated genes such as Madh4 ( −6 . 3 mo ) , Hdac6 ( −4 . 6 mo ) , Dnmt3b ( −5 mo ) , and Sirt7 ( −6 . 4 mo ) . Genes were then grouped on the basis of GO , and the categories with a significant shift in predicted age ( Wilcoxon t-test , p-value < 0 . 05 ) were identified . Overall , the p53+/m mice appeared younger by 1–5 mo compared to the p53+/− mice in the vast majority of significant GO categories ( 84/87 ) ( Figure 6B and Table S8 ) . Only three categories were “older” in p53+/m mice , the most striking being that of the inflammatory response , indicating that this can be uncoupled from the other manifestations of aging . Examples are P-selectin and Cox2 , which are older in the p53+/m compared to the p53+/− HSC by +4 . 6 and +6 mo , respectively . The notion of relative youth of p53+/m stem cells is supported by the observation that HSC from p53+/m mice do not exhibit an age-related increase in number as is observed in their WT counterparts [30] , and have attenuated activity after 5FU treatment [28] , which can be explained by the lower HSC proliferative capacity . This is also consistent with our observation that HSC from the p53+/m mice form in vitro colonies at a similar frequency as WT HSC , but with reduced size , indicating a reduced proliferative capacity ( Figure S5 ) . These results suggest that although p53+/m mice appear older at a tissue and organismal level , the genetically imposed slower rate of proliferation uncouples the manifestations of aging in HSC , causing them to remain relatively young at a molecular level despite an inflammatory and aged tissue milieu .
Limitations in stem cell number or function have been proposed to restrict longevity . However , we show here that in WT mice , stem cells decline in function , when measured per HSC , with a concomitant increase in their number , resulting in a minimal net change in overall HSC activity , strongly suggesting that stem cells are not likely to be a factor limiting hematopoietic regeneration with age . However , their functional deficits do show that HSCs are impacted by the forces of aging in a manner similar to that of differentiated cells [31–34] . In our molecular analysis , we identified global age-related changes in gene expression in murine HSCs , with a view to identifying mechanisms that could be responsible for these age-associated declines in HSC function . Genes involved in inflammatory and stress responses dominated the group of up-regulated genes , whereas those participating in chromatin regulation and DNA repair were prominent among down-regulated genes . Many of the observed expression changes are corroborated by a recent study of HSC aging [4] . Although the number of genes that go up or down with age was similar between our study and the one by Rossi et al . [4] , the precise overlap between the genes is modest: only 343 genes were differentially expressed in both lists . We believe this is primarily due to a different purification strategy and array analysis that was used in our study . Remarkably , however , the top ten genes in both lists of genes that were Up-with-Age as well as those that were Down-with-Age were identical , and the magnitude of the changes was striking . For example , P-selectin was found highly up-regulated with age , for a mean fold-change of approximately 60 between the two studies . Likewise , both studies observed the up-regulation of specific protooncogenes , such as Runx1 , that we believe could contribute to the increased incidence of myeloid leukemia with age . In both studies , the inflammation markers P-selectin and clusterin ( both regulated by NF-κB ) and protein-folding genes such as Hspa8 and Dnajc3 were found up-regulated . Additionally , we found genes involved in DNA repair and chromatin maintenance ( Xab2 , Rad52 , Polb , and Lmna ) to be down-regulated in both studies , supporting the notion that HSC become epigenetically dysregulated . The marked stress response exhibited by the HSC suggests that a proinflammatory microenvironment exists within the aging marrow . Studies of aging arteries , brain , and kidney have also observed up-regulation of inflammatory markers [33–35] , but could not distinguish between immune cell infiltration and intrinsic inflammatory response of the tissue . The up-regulation of TLR4 ( Toll-like receptor 4 ) ( Table 1 ) is especially intriguing because one of its principal signaling targets is NF-κB , which regulates a variety of genes involved in the inflammatory response , including prostaglandin E2 synthase ( Cox-2 ) and the adhesion molecules ICAM1 and P-selectin [36] ( Table 1 ) . This molecular profile resembles the elevated inflammatory state previously described in older mice ( “inflamm-aging” [37] ) ; however , our mice were raised in specific-pathogen–free conditions , and thus the inflamm-aging can be independent of antigenic load . Strikingly , Inflammatory Response was one of the few GO categories found “older” in the p53+/m mice compared to the p53+/− mice . This correlates with the early-aging phenotype observed at the organismal level in p53+/m mice , and suggests that inflammation is an intrinsic effect of age . The up-regulated inflammatory genes in aging HSCs may have mechanistic implications for the decline of HSC function . The ability of HSCs to engraft in bone marrow is influenced by their homing properties; since cell adhesion plays a critical role in the engraftment of HSCs , aberrant up-regulation of genes encoding P-selectin and ICAM1 might affect the ability of older HSCs to home to the bone marrow , disrupting repopulation . Likewise , down-regulation of genes in the TGFβ signaling pathway such as SMAD4 , endoglin , and spectrin b2 ( Table 1 ) , may explain the increased number of HSCs , because the TGFβ pathway influences HSC pool size [38] . Equally striking was the increased expression of clusterin ( clu ) and amyloid beta precursor protein ( App ) in aged HSCs ( Table 1 ) , because these proteins aggregate within plaques in the brains of Alzheimer's patients [39] . Cathepsin B , involved in APP [40] processing , also increased in aged HSCs ( Table 1 ) . Clusterin , containing chaperone homology , along with several chaperonin subunits ( beta , gamma , zeta , eta , and theta ) increased with age , as did stress-response genes Hsc70/Hspa8 , HSP90/Hspca , Bip/Hspa5 , and several Hsp40 homologs ( Dnaja1 , Dnaja2 , Dnajb6 , Dnajb10 , and Dnajc3 ) ( Tables 1 and S1 ) . These changes , as well as the approximately 6-fold enrichment in NO-mediated signaling genes ( Figure 3 ) , may be in response to oxidative stress , a common feature of aged cells [41] . Some of these genes were previously noted to increase with age in murine hearts and skeletal muscle , and to decrease in animals on dietary restriction [31 , 32] . Although inflammatory and stress-response genes increased with age , genes that ensure transcriptional fidelity declined with age . For instance , lamin A declined with age and is linked to normal human aging [42] and Hutchinson-Gilford progeria [43] ( Table 1 ) . Likewise , Bloom syndrome homolog ( Blm ) declined , mutations in which cause genomic instability , hypermutability , and cancer predisposition [44] . Multiple DNA repair genes , including Rad52 , Xrcc1 , and Xab2 were also down-regulated . Because DNA damage is thought to be a driving factor in aging [45] , a blunted DNA damage response could retard DNA repair , increasing the risk of retaining mutations , leading to malignant transformation , one of the hallmarks of old age . It was surprising that HSCs from 12-mo-old early-aging p53+/m mice appeared molecularly younger than age-matched WT and p53+/− HSC . This suggests that a genetically imposed lower rate of stem cell proliferation , as seen in the p53+/m HSC , can reduce the apparent age of the HSC , despite their residence in an environment that exhibits other outward manifestations of aging [28] . But this reduced proliferative capacity also results in poorer hematopoietic regeneration activity , when the stem cells are examined at the population level [28] . In other words , lower HSC proliferation results in a more youthful stem cell , but poorer tissue regeneration , and consequently an aged phenotype; this indicates that stem cell proliferation and tissue regeneration are finely balanced to maximize longevity , so that cell cycle disruption results in an uncoupling of tissue and organismal aging from the aging of the resident stem cell . Finally , three lines of evidence in our work indicate broad changes in epigenetic regulation with age . Several GO categories ( Figure 3A ) and specific genes involved in transcriptional silencing via chromatin regulation are down-regulated with age , such as the SWI/SNF-related chromatin remodeling genes ( Smarca4 and Smarcb1 ) , as well as three histone deacetylases ( Hdac1 , -5 , and -6 ) and a DNA methyltransferase ( Dnmt3b ) . Because these changes largely occur mid-way through life ( Figure 3B ) , they could easily be envisioned to underlie inappropriate expression of additional genes . In addition , the CORE analysis revealed many chromosomal regions coordinately changing with age , and suggested an overall loss of transcriptional silencing . Finally , inappropriate transcription from the IgK locus , known to be driven by NF-κB activity following epigenetic modification enabling accessibility of the locus , was observed in old , but not young , HSCs ( Figure 5 ) . Together , these data suggest an epigenetic view of aging that readily explains how so many diverse effects of age are evident at molecular , cellular , and organismal levels , and contrasts with the assumption that accumulation of lesions in genomic DNA or mitochondria accounts for the major effects of aging . Chromatin dysregulation could be a primary force in aging; epigenetic changes in otherwise normal cells could drive the loss of overall cellular functionality , as well as lay a fertile ground for secondary genetic events that lead irreversibly to oncogenic transformation . In this model , inappropriate expression of protooncogenes , or down-regulation of tumor suppressors , could result in a pre-transformed state , similar to myelodysplastic syndrome , a notion corroborated by another study of aging in murine HSC [4] in which Runx1 , Pml , and other protooncogenes were up-regulated with age . Likewise , the increased transcriptional accessibility of some loci may enable interactions between otherwise distant chromatin domains , enhancing the likelihood of chromosomal translocations . Of particular note , the IgK locus that we show is transcriptionally active in aged HSC , is a well-established translocation partner with the Myc protooncogene in the generation of hematopoietic malignancies [46] . Moreover , IgK alleles have been shown to be differentially located within the nucleus depending on their state of activation [47] , which could result in their juxtaposition to oncogenes , increasing the likelihood of translocation [48] . The role of epigenetic changes in cancer formation is increasingly recognized [49]; here , we suggest that chromatin dysregulation is a natural result of environmental insults with age , and a primary driver of secondary effects of age , including malignancy . A systems approach to the ways in which inflammation , stress response , and epigenetic regulation are linked may be essential to understanding aging and cancer .
All mice were housed in a specific-pathogen–free barrier and fed autoclaved acidified water and mouse chow ad libitum . C57Bl/6 CD45 . 1 mice were allowed to age to 2 , 6 , 12 , and 21–28 mo . HSCs were isolated as those cells that displayed the SP phenotype of Hoechst 33342 efflux [50] and were SParKLS . Hoechst and antibody staining was performed as previously described [11] . Sca-1 magnetic enrichment was performed with a Sca-1-biotin antibody ( eBioscience , http://www . ebioscience . com/ ) and antibiotin microbeads ( Miltenyi Biotec , http://www . miltenyibiotec . com/en/default . aspx ) on an AutoMACS ( Miltenyi ) . The Sca-1–enriched cells were suspended at a concentration of 108 cells/ml and incubated on ice for 15 min with strepavidin-Alexa488 ( Molecular Probes , http://probes . invitrogen . com ) , c-Kit-phycoerythrin ( PE; eBioscience ) , and antibodies against the lineage markers: PE-Cy5 conjugated Mac-1 , Gr-1 , CD4 , CD8 , B220 , and Ter119 ( eBioscience ) . Flow cytometric analysis was performed on a triple-laser instrument ( MoFlow; Cytomation , http://www . dako . com ) . For cell cycle analysis , HSCs were sorted into deionized water containing 0 . 1% sodium citrate and 50 μg/ml propidium iodide ( PI ) and analyzed with a FACScan flow cytometer ( BD Biosciences , http://www . bdbiosciences . com ) . Forty CD45 . 2 mice were transplanted with 3 , 10 , 30 , 100 , or 500 CD45 . 1 HSCs from either 2- or 21-mo-old mice along with 1 × 105 CD45 . 2 WBM competitor cells . Engraftment was measured by peripheral blood contribution 4 , 8 , and 16 wk post-transplantation . After sorting , the HSCs were resuspended in 50% FCS in HBSS at a concentration of 25 , 000 cells/ml , and 0 . 2 ml of cell suspension was spun onto glass slides with a cytocentrifuge ( Wescor , http://www . wescor . com ) at 800 rpm for 4 min . Fixed slides were blocked for 1 h with 30% goat serum in PBS , stained for 1 h with polyclonal anti-p65 antibody ( 1:100 in PBS; eBioscience ) , and stained with secondary antibody ( 1:250 and DAPI 1:1 , 000 ) . Texas Red background immunofluorescence was established with controls containing no primary antibody at an exposure time longer than 5 s . Texas red immunofluorescence images were photographed for 0 . 8 s . Total RNA was isolated from 2 . 5–5 × 104 HSCs ( 2–5 mice pooled per microarray ) with an RNAqeuous kit ( Ambion , http://www . ambion . com ) , treated with DNAseI , and precipitated with phenol:chloroform . This RNA was linearly amplified in two rounds of T7-based in vitro transcription ( MessageAmp kit; Ambion ) , and labeled in the last round with biotin-conjugated UTP and CTP ( Enzo Biochem , http://www . enzo . com ) . Amplified biotinylated RNA ( 20 μg ) was diluted in fragmentation buffer ( 5X; 200 mM Tris-acetate [pH 8 . 1] , 500 mM KOAc , and 150 mM MgOAc in DEPC water ) to a final volume of 40 μl , incubated at 94 °C for 25 min , and stored at −80 °C . A sample was run on a 4% nondenaturing agarose gel to confirm an RNA fragment length of approximately 50 base pairs . The labeled RNA was hybridized to MOE430A chips according to standard protocols ( Affymetrix , http://www . affymetrix . com ) ; the chips were then washed and counterstained with PE-conjugated streptavidin and a biotinylated anti-streptavidin antibody . The raw image and intensity files were generated with MAS 5 . 0 software ( Affymetrix ) . All microarrays passed several quality controls as previously described [14] . Normalization and model-based expression measurements were performed with GC-RMA [51] ( http://www . bioconductor . org ) . To identify genes whose expression varied significantly over time , we fit smooth curves to gene expression profiles by regression analysis . A quadratic polynomial was fit to the profile of each gene by use of LIMMA [52] , enabling us to analyze progressive changes as well as peaks and valleys in expression over time . The Up-with-Age and Down-with-Age gene lists were identified by applying a linear contrast model and modified t-statistics from an empirical Bayes procedure [52] coupled with a linear step-up [53] multiple-testing correction ( to estimate and minimize the false discovery rate to less than 5% ) and a fold-change criterion of at least 2-fold for the 19-mo study period . In short , this contrast model is a composite score of reproducibility ( t-statistic , p-value ≤ 0 . 05 ) and fold-change ( >2-fold ) , which simply conveys a degree of difference ( contrast ) in gene expression over time . To investigate the biological significance of the gene lists described above , we used GO ( http://www . geneontology . org ) . GO is a controlled vocabulary that describes gene biological roles and is arranged in a quasi-hierarchical structure from more general terms to the more specific . After mapping each gene in the two lists to the GO tree structure , we determined the number of genes at or below any given node in the GO hierarchy and the amount of statistically significant enrichment ( Fisher exact p-value ) for each GO node relative to chance observation , using a previously developed procedure [54] . To assess the emergence and disappearance of enriched GO categories , we defined the time of half-maximal expression change ( T1/2-max ) for each gene in each category over the time interval . For genes whose maximal expression values were outside the 2- to 21-mo interval , the T1/2-max was determined as the intermediate expression value between the expression at 2 and 21 mo . For genes whose extreme expression values were within the interval , the T1/2-max was determined as the intermediate expression value between the expression at 2 mo and that extrema . Genes were grouped by GO category , yielding reliable estimates of time of induction and reduction for a given biological process . We conducted this analysis separately for the Up-with-Age and Down-with-Age gene lists . To identify COREs , we obtained the genome coordinates of the Affymetrix MOE430A array from the MM5 build of the UCSC Genome Browser . To compare the locations of age-induced or age-repressed genes , we divided all genes into two disjoint classes based on the sign of the 21-mo versus 2-mo contrast . Redundant probe sets were removed by grouping all probe sets by Entrez Gene annotation . Because not all probe sets for a single Entrez Gene identifier have the same sign for the contrast score , the sign of the mean value was assigned to the Entrez Gene identifier . To compare the positions of these locations , we constructed a Gaussian kernel density estimate by chromosomal position for genes that increased with age and genes that decreased with age . We then calculated a ratio of these density estimates for the two groups . This ratio represents the density of genes that increase ( peak ) or decrease ( valley ) with age . A permutation test was performed to estimate the p-value where gene locations were randomly swapped along each chromosome , maintaining gene density but randomizing direction of regulation ( up/down ) , Density estimate ratios were calculated based on 1 , 000 random permutations . This calculation enabled us to estimate a threshold of statistical significance such that peaks and valleys ( high densities of age-induced and age-repressed genes ) exceeding the 0 . 025 and 0 . 975 permutation-based quantiles were judged to be statistically significant at an estimated p-value of 0 . 05 . Purified cells were sorted into lysis/PCR buffer , and PCR was performed as previously reported [24] . For GL RT-PCR , approximately 20 , 000 HSCs from either young or old mice were sorted into HBSS , and RNA was isolated by the RNAqueous kit ( Ambion ) . RT-PCR was performed with an oligo-dT primer and SuperScript II ( Invitrogen , http://www . invitrogen . com ) followed by 50 cycles of PCR . RT-PCR fragments were purified , cloned into the Topo 2 . 1 vector ( Invitrogen ) , and sequenced . IgH recombination primers are previously published [24] . IgK GL transcript primers include 5′-CTTCAGTGAGGAGGGTTTTTG-3′ ( forward 1 ) , 5′-ACTATGAAAATCAGCAGTTCTC-3′ ( forward 2 ) , and 5′-CGTTCATACTCGTCCTTGGTC-3′ ( reverse ) . To assess the age-related expression differences between the p53+/m and p53+/− mice , genes with best-fitting trend lines ( R2 > 0 . 50 ) from the WT HSC aging time course were selected , and a predicted age ( in months ) was extrapolated for each gene based on the level of expression for both the p53+/m and p53+/− 12-mo-old mice . Genes were grouped on the basis of GO for both phenotypes and the categories with a significant shift in age ( Wilcoxon t-test ) between the p53+/m and p53+/− mice were identified by a p-value ≤ 0 . 05 and a median aged difference of greater than 1 mo . Mice used in these experiments have been back-crossed onto the C57Bl/6 background for four or more generations . All data can be downloaded from our Web site http://rd . plos . org/pbio . 0050201 . In addition , all microarray data files have been deposited in the Gene Expression Omnibus ( accession number GSE6503 ) .
Entrez Gene ( http://www . ncbi . nlm . nih . gov/sites/entrez ? db=gene ) ID accession numbers for the genes discussed in this paper are App ( 11820 ) , Blm ( 12144 ) , CatnB ( 12387 ) , Cct6a ( 12466 ) , CD150 ( 6504 ) , CD19 ( 12478 ) , CD45 ( 19264 ) , CD48 ( 12506 ) , c-Kit ( 16590 ) , Clu ( 12759 ) , Cox2 ( 19225 ) , Ctsb ( 13030 ) , Ctsc ( 13032 ) , Ctss ( 13040 ) , Dnaja1 ( 15502 ) , Dnaja2 ( 56445 ) , Dnaja2 ( 56445 ) , Dnajb10 ( 56812 ) , Dnajb6 ( 23950 ) , Dnajc3 ( 19107 ) , Dnmt3b ( 13436 ) , Dnmt3b ( 13436 ) , Eng ( 13805 ) , Hdac1 ( 433759 ) , Hdac5 ( 15184 ) , Hdac6 ( 15185 ) , Hdac6 ( 15185 ) , Hspa5 ( 14828 ) , Hspa8 ( 15481 ) , Hspca ( 15519 ) , Icam1 ( 15894 ) , IgH ( 111507 ) , IgK ( 243469 ) , Il-7r ( 16197 ) , Lmna ( 16905 ) , Madh4 ( 17128 ) , p53 ( 22059 ) , p65 ( 19697 ) , Pml ( 18854 ) , Rad52 ( 19365 ) , Runx1 ( 12394 ) , Sca-1 ( 110454 ) , Selp ( 25651 ) , Sirt2 ( 64383 ) , Sirt3 ( 64384 ) , Sirt7 ( 209011 ) , Smarca4 ( 20586 ) , Smarcb1 ( 20587 ) , Spnb2 ( 20742 ) , Tlr4 ( 21898 ) , Xab2 ( 67439 ) , and Xrcc1 ( 22594 ) .
|
Aging is marked by a decline in function of the entire organism . The effect of age on the regenerative capacity of adult stem cells , which should rejuvenate tissues throughout life , is poorly understood . Bone marrow stem cells , also known as hematopoietic stem cells ( HSCs ) , continuously regenerate the cells that comprise the blood , including the immune system , which fails with age . Here , we show that older HSCs were less able to regenerate the blood system than young HSCs . Paradoxically , the HSC number increased concomitantly , leading to no major difference in overall blood production , even though the immune system did exhibit some defects . To determine why these changes occurred , we looked at global patterns of gene expression in young versus old HSC . Stem cells exhibited an elevated inflammatory response and a decline in factors , called chromatin regulators , that orchestrate DNA accessibility and gene expression . Additional evidence supports the idea that loss of overall gene regulation ( epigenetic regulation ) is a major event during aging . Whereas much of aging research is concentrated on accumulation of mutations in DNA rather than on global regulatory mechanisms , we speculate that these epigenetic changes could drive many of the manifestations of age . This view also may explain the increased incidence of cancer with age .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] |
[
"cell",
"biology",
"immunology",
"mus",
"(mouse)",
"hematology"
] |
2007
|
Aging Hematopoietic Stem Cells Decline in Function and Exhibit Epigenetic Dysregulation
|
The Middle East respiratory syndrome coronavirus ( MERS-CoV ) is a zoonotic betacoronavirus that was first detected in humans in 2012 as a cause of severe acute respiratory disease . As of July 28 , 2017 , there have been 2 , 040 confirmed cases with 712 reported deaths . While many infections have been fatal , there have also been a large number of mild or asymptomatic cases discovered through monitoring and contact tracing . New Zealand white rabbits are a possible model for asymptomatic infection with MERS-CoV . In order to discover more about non-lethal infections and to learn whether a single infection with MERS-CoV would protect against reinfection , we inoculated rabbits with MERS-CoV and monitored the antibody and inflammatory response . Following intranasal infection , rabbits developed a transient dose-dependent pulmonary infection with moderately high levels of viral RNA , viral antigen , and perivascular inflammation in multiple lung lobes that was not associated with clinical signs . The rabbits developed antibodies against viral proteins that lacked neutralizing activity and the animals were not protected from reinfection . In fact , reinfection resulted in enhanced pulmonary inflammation , without an associated increase in viral RNA titers . Interestingly , passive transfer of serum from previously infected rabbits to naïve rabbits was associated with enhanced inflammation upon infection . We further found this inflammation was accompanied by increased recruitment of complement proteins compared to primary infection . However , reinfection elicited neutralizing antibodies that protected rabbits from subsequent viral challenge . Our data from the rabbit model suggests that people exposed to MERS-CoV who fail to develop a neutralizing antibody response , or persons whose neutralizing antibody titers have waned , may be at risk for severe lung disease on re-exposure to MERS-CoV .
Since its discovery in 2012 , the Middle East respiratory syndrome coronavirus ( MERS-CoV ) has caused at least 2 , 040 human infections and 712 deaths worldwide [1 , 2] . Like other human coronaviruses ( 229E , OC43 , NL63 and HKU1 ) , MERS-CoV is associated with respiratory tract infection . However , unlike most other human coronaviruses , MERS-CoV has a zoonotic origin and can cause severe illness , resulting in acute respiratory distress syndrome . These characteristics are reminiscent of severe acute respiratory syndrome coronavirus ( SARS-CoV ) , which caused a large outbreak of human infections in 2003 [2] . Serological surveys of persons in the Arabian Peninsula have shown low or undetectable levels of preexisting antibody against MERS-CoV , although those in close contact with camels ( the reservoir host for MERS-CoV ) have higher rates of seropositivity than the general population [3–5] . Longitudinal studies have also indicated that serum antibody titers may wane over time , particularly following mild infections [6–8]; similar to what has been observed for other coronaviruses like SARS-CoV [9] . Since the discovery of MERS-CoV , only one autopsy report has been published and the course of MERS-CoV infection in humans is still not well understood [10] . This is particularly true for the mild or asymptomatic infections , which comprise a large number of MERS-CoV infections in healthy adults [11–14] . We wished to explore the immune response during non-lethal MERS-CoV infection , and to determine whether such infections would be protective . Several small animals , including ferrets , hamsters , and mice which are frequently used as animal models for human disease have proven resistant to infection with MERS-CoV [15–18] . The dipeptidyl peptidase 4 ( DPP4 ) protein , which is the cellular receptor for MERS-CoV in these animals differed from human DPP4 at key residues , and therefore did not bind to the MERS-CoV spike protein [18] . Several modified mouse models have been generated to overcome this receptor-mediated restriction including both transduced and transgenic animals expressing human DPP4 , and lethal infection models have been established [19–21] . Non-human primates have been successfully infected , with rhesus macaques displaying a mild , transient illness and marmosets demonstrating a more severe and sometimes lethal infection [22–25] , although there is some discrepancy in findings from marmosets [26] . Camels and alpacas have also been experimentally infected and exhibit transient viral replication in the upper respiratory tract [27 , 28] . However , the expense and care of camels and the ethical concerns surrounding the use of non-human primates limits their widespread utility for research studies . The New Zealand white rabbit supports productive replication of the MERS-CoV isolate EMC/2012 without associated clinical signs of disease [29] . We sought to characterize the role of antibodies in protection from reinfection following asymptomatic infection . We found that primary infection failed to induce neutralizing antibodies and reinfection was associated with increased pulmonary inflammation . Reinfection elicited neutralizing antibodies that protected rabbits from subsequent infection . Thus , whilst neutralizing antibodies are protective , they may not be elicited or may not last long after mild infection with MERS-CoV and infection in the presence of only non-neutralizing antibodies may be associated with enhanced pulmonary inflammation .
In order to study the initial disease progression and antibody response associated with MERS-CoV infection in rabbits , we infected nine New Zealand White rabbits with either a low dose ( 103 TCID50 ) or high dose ( 105 TCID50 ) of EMC/2012 ( Fig 1 ) . None of the rabbits displayed overt clinical signs of illness in the 14 days following infection . Viral RNA titers were measured by qRT-PCR using primer pairs targeting the nucleocapsid protein ( N3 ) and are reported as genome equivalents per gram of tissue [30] , as we found this method to be more sensitive and less variable than recovering infectious virus from infected rabbit tissues ( Fig 2A and 2B and S1 Fig ) as reported previously in the rabbit and other MERS-CoV animal models [22 , 29] . We observed a transient infection following inoculation , with detection of viral RNA largely limited to the respiratory tract . In the nasal turbinates , viral RNA was detected only sporadically , although the titers increased after day 1 post-infection . In general , the higher dose of virus resulted in greater mean genome equivalent titers than infection with the lower dose of MERS-CoV ( Fig 2A ) . A dose-response was observed in the lower respiratory tract; with the 105 TCID50 inoculum resulting in significantly higher titers than infection with 103 TCID50 of virus on days 1 and 3 post-infection ( Fig 2B ) ( p values of 0 . 03 and 0 . 0001 respectively ) . Following primary infection with 105 TCID50 , mild inflammation involving the perivascular , peribronchiolar , and alveolar interstitial regions was observed in the lungs at day 3 post-infection , with little to no cellular debris within airways ( Fig 3A , S1 and S3 Tables ) . The cellular infiltrate was largely composed of eosinophils and macrophages and fewer lymphocytes and plasma cells ( Fig 3A inset ) . Inflammation was not observed following infection with 103 TCID50 of virus or in media-only controls ( Fig 3B and 3C and S1 and S3 Tables ) based on blind scoring and digital quantitative analysis . Immunohistochemistry ( IHC ) revealed virus antigen following infection with the higher ( 105 TCID50 ) dose of virus ( Fig 3D and inset , S1 and S2 Tables ) , but not the lower dose or media-only inoculum ( Fig 3E and 3F and S1 and S2 Tables ) . In a separate experiment , rabbits were inoculated with 106 . 5 TCID50 of EMC/2012 ( Fig 2C and 2D ) . Viral RNA titers in the lung were sustained until day 5 , then dropped to almost baseline levels by day 10 , and were undetectable on day 28 post-infection . Serum was collected for detection of MERS-specific antibodies by both ELISA and microneutralization ( MN ) assays ( Table 1 ) . Eight weeks after inoculation , serum antibodies against the S protein were detected by IgG ELISA in all of the rabbits inoculated with 105 TCID50 [geometric mean titer ( GMT ) 1016] , but not in the group inoculated with 103 TCID50 . Antibodies against the nucleocapsid ( N ) protein were detected by IgG ELISA in two rabbits previously inoculated with 105 TCID50 and one rabbit with 103 TCID50 . However , neutralizing antibodies were not detected in rabbits inoculated with any dose ( 103 , 105 , or 106 . 5 ) of EMC/2012 ( Table 1 ) . These data indicate that there is a dose-response in MERS-CoV infected rabbits measured by viral RNA and antibody titers . The peak in viral titers occurs at day 3 post-infection , with higher titers observed following infection with 105 or 106 . 5 TCID50 of virus . Since the 105 and 106 . 5 TCID50 doses gave similar results , the 105 TCID50 dose was chosen for the remaining studies . Our findings in rabbits are reminiscent of a few reports of human cases of qRT-PCR confirmed infection with MERS-CoV that failed to elicit either a neutralizing antibody response , or any detectable antibody response against the virus [6 , 8 , 31] . In order to determine whether such patients would be susceptible to reinfection , we repeated the MERS-CoV infection in the previously infected rabbits . Eight weeks after primary infection , we challenged six rabbits that had previously received the high or low dose of MERS-CoV with 105 TCID50 of EMC/2012 ( Fig 1 ) . Additional naïve rabbits were inoculated for comparison . As in primary infection , clinical signs were not observed upon reinfection . Neither group of reinfected rabbits had viral RNA detected in the upper respiratory tract , although viral RNA was detected in the primary infection control group ( S2 Fig ) . However , all groups had evidence of pulmonary infection . The rabbits infected serially with 105 TCID50 of EMC/2012 ( 105//105 ) had lower viral RNA titers compared to both the 103//105 TCID50 and 105 TCID50 primary infection groups , with mean titers of 102 . 9 , 103 . 7 , and 104 TCID50 eq per gram of tissue respectively ( Fig 4A ) . Inflammatory changes were more severe upon reinfection compared to primary infection , with the greatest inflammation observed in the animals previously infected with the low dose of virus ( Fig 3A , 4B and 4C , and S1 and S3 Tables ) . These severely inflamed regions were characterized by an abundance of eosinophils , macrophages , lymphocytes and plasma cells which formed densely cellular collars of inflammatory cells around the affected perivascular and peribronchiolar regions . In addition , the cellular infiltrate expanded and obscured much of the adjacent alveolar interstitium . The alveolar interstitium also contained regions of proteinaceous fluid and diffuse type II pneumocyte hyperplasia ( Fig 4B inset ) . This inflammatory response was driven by reinfection , and was not residual inflammation from the primary infection . This was confirmed by including a group of previously infected rabbits that received diluent alone in the reinfection study ( Fig 4D , and S1 Table ) . The rabbits in the 103//105 TCID50 group had antigen levels comparable to primary infection , while the 105//105 TCID50 group had lower levels of antigen by IHC ( Fig 4E and 4F , S1 and S2 Tables ) . These data indicate that low titers of non-neutralizing antibodies do not protect rabbits from reinfection , and may instead result in enhanced inflammation . The S protein-specific IgG ELISA antibody titers were boosted following secondary infection and remained detectable for five weeks , with a GMT of 6451 for the 103//105 TCID50 group , and a GMT of 4064 for the 105//105 TCID50 group ( Table 1 ) . N protein-specific antibodies were found in all rabbits in the 103//105 TCID50 group and one of the 105//105 TCID50 group . Secondary infections in both groups resulted in the production of neutralizing antibodies , although the titer in one rabbit in the 103//105 TCID50 group dropped below the detection limit by five weeks post-infection ( Table 1 ) . Overall the 103//105 TCID50 group had lower neutralizing titers than the 105//105 TCID50 group , with GMTs of 27 and 73 respectively . To determine if neutralizing antibodies would protect from reinfection , three rabbits from each secondary infection group were re-challenged with 105 TCID50 EMC/2012 five weeks later ( Fig 1 ) . Clinical signs of illness were not observed in any of the rabbits . As was observed with the second infection , viral RNA was not detected in the upper respiratory tract samples from either group of rabbits on day 3 post-infection . In the lungs , a significant decrease in the amount of viral RNA was observed in both the 105//105//105 TCID50 group and the 103//105//105 TCID50 group compared to primary infection , with mean titers of 101 . 9 , 102 . 6 , and 104 . 5 TCID50 eq per gram of tissue respectively ( p values of 0 . 0006 and 0 . 003 ) ( Fig 5A ) . This decrease in viral load was also observed by IHC ( S1 and S2 Tables ) . Histologically , the lungs from both groups displayed mild inflammation and minimal antigen burden ( Fig 5B and 5C and S1 , S2 and S3 Tables ) . In these milder regions of peribronchiolar and perivascular inflammation , eosinophils and macrophages predominated ( Fig 5B inset ) . Thus , infection in the presence of neutralizing antibodies was associated with significant protection from viral infection and associated pathology in both the upper and lower respiratory tract of the rabbits . Moreover , the prechallenge serum neutralizing antibody titers inversely correlated with viral RNA titers following tertiary infection ( Fig 5A ) . To determine if non-neutralizing antibodies were responsible for the enhanced inflammation observed following reinfection , we performed a passive transfer ( PT ) experiment . Serum collected from rabbits four weeks following primary infection with 103 TCID50 of EMC/2012 was transferred either undiluted or at a 1:10 dilution in PBS to naïve rabbits that were challenged with 105 TCID50 of virus the following day . For comparison , a group of previously infected rabbits were reinfected . Although ELISA antibodies against the S protein were barely detectable in the serum ( Table 1 ) , after the serum was concentrated ten-fold prior to administration the ELISA titers ranged from 10 to 40 . Neutralizing activity was not detected , even after concentration of the serum . Thus , very low titers of non-neutralizing antibodies were present in the transferred serum . Passively transferred antibodies did not affect viral titers in the lower respiratory tract as determined by qRT-PCR ( Fig 6A ) . Rabbits that received the undiluted passively transferred ( PT ) serum exhibited immunopathology similar to that observed in previously infected rabbits based on blinded scoring ( Fig 6B and 6C and S1 Table ) . There was an increase in observed vascular congestion in this group of PT rabbits compared to the other groups . The rabbits that received serum antibodies at the lower dilution did not demonstrate enhanced inflammation ( Fig 6D , S1 Table ) . Overall , the pathology in the rabbits that were infected after PT of post-infection serum was milder than in other reinfection studies , possibly due to the shortened interval between primary infection and reinfection . Viral antigen levels appeared similar between all groups by IHC ( Fig 6E , 6F and 6G , S1 and S2 Tables ) . Non-neutralizing antibodies typically enhance inflammation and pathology during an immune response through interactions with Fc or complement receptors [32] . We first examined the possibility that the antibodies were causing enhanced inflammation due to an increase in viral uptake and replication in macrophages through interaction with their native cellular receptor or an Fc receptor , as happens in dengue [32–34] . Antibody-dependent enhancement ( ADE ) during infection has also been observed with other coronaviruses , such as feline infectious peritonitis virus ( FIPV ) [35] . However , it was not clear that such a mechanism was likely since the enhanced inflammation in secondary MERS-CoV infection in rabbits was not associated with an increased viral load by either qRT-PCR or IHC ( Fig 4A , S1 and S2 Tables ) . In order to examine the replication of MERS-CoV in macrophages , we differentiated THP-1 cells into macrophages and infected them with MERS-CoV in the presence or absence of rabbit sera and measured viral titers after 48 hours . Heat-inactivated sera from naïve rabbits ( week 0 ) , week 8 following primary infection ( only non-neutralizing antibodies present ) , and week 13 following secondary infection ( neutralizing antibodies present ) were tested at three dilutions . All dilutions displayed similar trends but only undiluted samples are shown in Fig 7 . Vero81 cells were used as a positive control ( Fig 7A ) and Raji cells were included as a negative control ( Fig 7B ) of infection . Compared to the level of viral replication observed in macrophages without serum , addition of rabbit sera produced no enhancement of viral replication . In fact , in the presence of rabbit sera , there was a significant decrease in viral replication in THP-1 cells compared to titers in the absence of rabbit serum ( Fig 7C ) , indicating that the non-neutralizing antibodies did not enhance MERS-CoV replication in these cells . The other possibility for ADE of inflammation is through interaction with complement receptors . We investigated the potential role of complement in the enhanced pulmonary inflammation by evaluating lung samples from primary and secondary MERS-CoV infections with an ELISA against rabbit complement protein C3a . Using this assay , we observed an increased amount of complement protein per gram of lung tissue in both secondary infection groups ( mean of 1084 ng/g for 103//105 and 939 ng/g for the 105//105 ) compared to primary infection ( mean value of 699 ng/g ) ( Fig 8A ) . This increase was significant for the 103//105 group compared to primary infection ( p = 0 . 02 ) . We further validated the association between complement and increased inflammation using an anti-complement ( C9 ) antibody . Immunofluorescence revealed complement recruitment through the deposition of virus antigen and C9 within the inflammatory milieu surrounding many vessels and airways in the lungs of the reinfected rabbits ( Fig 8B ) . This was in direct contrast to the primary infection group in which virus antigen was detected adjacent to small vessels and airways with minimal inflammation and no evidence of complement deposition ( Fig 8C ) . Staining for other complement targets ( C1q , C4b , C3a , and C3c ) was unsuccessful in the rabbit tissues . Since T cell responses could also be involved in enhanced inflammation , we stained lung tissues with an anti-CD3 antibody ( Fig 9 ) . Following reinfection , we observed a substantial increase in the numbers of CD3+ T cells in the lung compared to primary infection ( Fig 9A and 9B ) . These T cells were distributed in the same areas as virus antigen , largely in the areas immediately surrounding vessels and airways ( Fig 9C ) . Attempts to further characterize the CD3+ cells were unsuccessful .
While MERS-CoV is able to cause severe disease with a lethal outcome , many otherwise healthy individuals display a mild or asymptomatic disease course . Using the rabbit model , we examined the serum response following asymptomatic infection , and found that an antibody response lacking neutralizing activity was not protective against reinfection . Our study extends previously published information on the rabbit model [29] by examining a dose response and exploring the enhanced inflammation observed during reinfection . The inclusion of a low dose of virus in our studies revealed the potentially detrimental effects of non-neutralizing antibodies and further demonstrated the protective benefit of neutralizing antibodies [21 , 36 , 37] . To our knowledge , the rabbit infection model described here is the only model of MERS-CoV infection in which non-neutralizing antibodies are exclusively elicited following primary infection . Our observations in the rabbit are in general agreement with those reported by Haagmans et al [29] , with the exception that neutralizing antibodies were not detected following primary infection in our study . This is likely a consequence of different routes of inoculation . Haagmans and colleagues inoculated rabbits through both the intranasal and intratracheal routes , whereas in our study rabbits were inoculated through the intranasal route alone . In addition , the volume of the inoculum was not stated by Haagmans et al , although a recent publication indicates a volume up to 3ml may have been delivered intratracheally [38] . This differs significantly from the volume used in our studies ( 1ml ) and may affect viral load . Differences in inoculation routes and volumes have been shown to affect disease severity and immune response in other models [39 , 40] . The lack of neutralizing antibodies in the rabbits in our model allows us to examine a phenomenon that is otherwise only observed in humans . The potential clinical implication of our findings is a risk of severe pulmonary disease in persons who fail to develop a neutralizing antibody response following exposure to MERS-CoV or in persons in whom titers of neutralizing antibodies decay and are no longer detectable . In rabbits , the highest mean viral RNA levels were observed in the lower respiratory tract , as has been reported in human cases [10 , 41–43] . A direct correlation was observed between the distribution and amount of viral antigen and inflammation in the lungs of the rabbits . In a limited number of samples , aberrantly high levels of viral antigen were observed by both qualitative and quantitative measures that were not mirrored in the qRT-PCR results ( S1 and S2 Tables ) . We believe this occurred due to sampling of locations where the inoculum pooled after infection , particularly since these atypical values occurred most often in the caudal lobes . The cellular infiltrates observed in the rabbit lungs were largely composed of mixed populations of eosinophils and histiocytes . In mammalian species , the presence of eosinophilic inflammation is often associated with parasitic infections , hypersensitivity reactions , and less often , certain fungal infections [44] . There was no evidence of parasitic or fungal infection in the rabbits and the commonly observed features of hypersensitivity-related pneumonitis ( i . e . edema and bronchiolitis ) were absent [45] . Viral replication in the upper respiratory tract was detected only during primary infection , suggesting that immune responses prevented local replication of the virus during later challenge . While we only analyzed serum IgG antibodies by ELISA , mucosal IgA antibodies in the upper respiratory tract may also play a role in preventing reinfection . However , the serum IgG antibody lacked neutralizing activity and contributed to the enhanced pulmonary inflammation observed upon reinfection . Passive transfer of serum from previously infected rabbits to naïve animals followed by MERS-CoV challenge recreated the same histopathological depiction , though there was an increase in vascular congestion following passive transfer that was not observed in the previously infected rabbits . This is likely a consequence of transferring complete serum containing additional serum proteins , including those involved in the complement cascade into naïve rabbits instead of transferring purified MERS-CoV specific IgG antibodies . As we observed in these studies , the additional complement proteins could have led to increased inflammation and congestion in the lungs . Even in the presence of the enhanced inflammation following reinfection , the rabbits continued to lack any discernible clinical signs of infection . This occurrence could be explained by several factors . Rabbits are prey animals , which have evolved to mask signs of illness as a defense mechanism . Also , with the experiments being conducted in a high containment facility , we were limited in our ability to measure activity levels . Furthermore , the lesions in the lungs were typically multifocal and focally severe . The remaining lung tissue may have been sufficiently functional to limit clinical signs . The effect of non-neutralizing antibodies observed in this MERS-CoV study differs from those observed with FIPV and flaviviruses such as dengue , since the non-neutralizing antibodies did not enhance MERS-CoV replication . The enhanced inflammation observed in the rabbits after MERS-CoV reinfection appears to be mediated through interactions between non-neutralizing antibodies and complement proteins , resulting in activation of the complement cascade and formation of immune complexes . The increase of C3a and C9 proteins in the rabbit lungs following reinfection supports this possibility . Immune complexes have been implicated in the pathogenesis in other viral infections , including influenza and RSV [46 , 47] . While C9 and the formation of membrane attack complexes ( MACs ) are typically involved in response to bacterial pathogens , enveloped viruses are also susceptible to lysis by MACs [48 , 49] . Complement activation can also be responsible for an increase in the release of anaphylotoxins and more recruitment and activation of immune cells , leading to inflammation . The presence of CD3+ T cells in the same regions of the lung as MERS-CoV N antigen during reinfection is consistent with this scenario ( Fig 9 ) . Our studies demonstrated that MERS-CoV reinfection elicited neutralizing antibodies that protected rabbits from further viral challenge . Antibody-mediated protection has also been exhibited in rabbits and mice following prophylaxis with neutralizing monoclonal antibodies against the MERS-CoV spike protein [21 , 36 , 37] . These data support the induction of neutralizing antibodies as the primary goal for vaccines . However , the use of convalescent serum for treatment of MERS-CoV infected individuals has had limited , if any , benefit [50–52] . Also , past experience with SARS-CoV triggers a cautionary note . In mouse and hamster models , vaccine-induced neutralizing antibodies prevented or reduced replication of SARS-CoV [53 , 54] . In contrast , in ferrets and nonhuman primate models , SARS-CoV antibodies restricted replication of challenge virus but did not prevent pulmonary inflammation [55 , 56] . In addition , antibody-dependent enhancement and pulmonary immunopathology was seen following challenge with some vaccine strategies for SARS-CoV , including virus-like particles and inactivated vaccines [57 , 58] . There are some data to suggest that MERS-CoV vaccine approaches may also result in immunopathology; as eosinophilic infiltration with enhanced lung pathology was observed in vaccinated transgenic mice following MERS-CoV challenge [59] . Since the mice had neutralizing antibodies before challenge , and had significant reduction in viral titers following challenge compared to control mice , we believe the mechanisms behind these two phenomena are distinct , but still require consideration . These discrepant observations highlight the critical need for additional clinical data , and continued attention during the development and testing of coronavirus vaccines . Another approach to viewing our data is to consider primary infection in rabbits as a type of vaccination , resulting in an immune response without overt clinical symptoms . This “priming” infection produces an immune response that is inadequate for protection . The secondary infection then acts as a “booster” , activating the memory response elicited by the primary infection and inducing neutralizing antibodies . Either interpretation indicates that the production of neutralizing antibodies should be the goal of MERS-CoV vaccines . Additional vaccine doses may be needed if neutralizing antibody titers wane rapidly . Rare cases of qRT-PCR confirmed human MERS-CoV infections have been reported in which neutralizing or S protein ELISA antibody responses were not detected [6 , 8 , 31] , most often following mild or asymptomatic infection . The rabbit model , particularly with use of lower viral inoculum dose and volume , may recapitulate such cases . If neutralizing antibodies against MERS-CoV are not produced or wane over time , a mild or asymptomatic infection may prime individuals for more severe disease upon re-exposure . This possibility could occur after either infection or vaccination , and should be considered during the development of MERS-CoV vaccines .
Vero81 cells ( ATCC ) were grown and maintained in Opti-MEM media ( GIBCO ) with 5% FBS ( HyClone ) . Raji ( ATCC ) and THP-1 cells ( ATCC ) were maintained in RPMI-1640 media ( GIBCO ) with 10% FBS and 50μM β-mercaptoethanol ( Sigma ) . The virus HCoV-EMC/2012 was obtained from Erasmus Medical Center , Netherlands . Virus stocks were stored at -80°C . The titer of the stock virus was determined by serial dilution in Vero81 cells and calculated by the Reed and Muench method [60] . All experiments were performed in a biosafety level 3 ( BSL3 ) facility . Male New Zealand white rabbits ( Covance , Princeton , NJ ) between five to nine months of age were anesthetized with a combination of intramuscular dexmedetomidine and isoflurane inhalation . Animals were inoculated intranasally ( i . n . ) with virus diluted in 1ml of MERS-CoV in Leibovitz-15 ( L15 ) media ( GIBCO ) , or mock-infected with 1ml of media alone . Atipamezole was subsequently administered subcutaneously to reverse sedation . Rabbits were monitored daily for 14 days after infection for clinical signs of disease including temperature , weight , lethargy , ocular discharge , rhinitis , labored breathing , ruffled fur , inappetence , and diarrhea . Serum was collected via the ear vein prior to inoculation and at specified times following infection . Animals were euthanized by Beuthanasia D administration and tissues were collected for viral titration , histopathology , and immunohistochemistry ( IHC ) . For passive transfer ( PT ) studies , serum from rabbits infected 28 days prior was concentrated 10-fold using an Amicon Ultra-15 filter column and then transferred intravenously through the ear vein to naïve rabbits either undiluted or at a 1:10 dilution , one day prior to infection with 105 TCID50 of MERS-CoV in 1ml . All infections consisted of the EMC/2012 strain unless otherwise noted . All animal studies were conducted in ABSL3 laboratories at the National Institutes of Health ( NIH ) . All procedures were reviewed and approved by the NIAID DIR Animal Care and Use Committee . The animals were housed in rabbit/ferret bio-containment racks and maintained in accordance with the Animal Welfare Act , the Guide for the Care and Use of Laboratory Animals , and other Federal statutes and regulations relating to animals , in a fully AAALAC accredited facility . All procedures were performed utilizing appropriate anesthetics as listed in the NIAID DIR Animal Care and Use Committee approved animal study proposal LID 33E . Euthanasia methods were consistent with the AVMA Guidelines on Euthanasia and the endpoint criteria listed in the NIAID DIR Animal Care and Use Committee approved animal study proposal LID 33E . The NIAID DIR Animal Care and Use Program , as part of the NIH Intramural Research Program ( IRP ) , complies with all applicable provisions of the Animal Welfare Act ( http://www . aphis . usda . gov/animal_welfare/downloads/awa/awa . pdf ) and other Federal statutes and regulations relating to animals . The NIAID DIR Animal Care and Use Program is guided by the "U . S . Government Principles for the Utilization and Care of Vertebrate Animals Used in Testing , Research , and Training" ( http://oacu . od . nih . gov/regs/USGovtPrncpl . htm ) . The NIAID DIR Animal Care and Use Program acknowledges and accepts responsibility for the care and use of animals involved in activities covered by the NIH IRP’s PHS Assurance #A4149-01 , last issued 11/24/2014 . As partial fulfillment of this responsibility , the NIAID DIR Animal Care and Use Program ensures that all individuals involved in the care and use of laboratory animals understand their individual and collective responsibilities for compliance with that Assurance , as well as all other applicable laws and regulations pertaining to animal care and use . The NIAID DIR Animal Care and Use Program has established and will maintain a program for activities involving animals in accordance with the most recent ( 2011 , 8th edition ) of “The Guide for the Care and Use of Laboratory Animals” ( ILAR , NRC ) ( http://oacu . od . nih . gov/regs/guide/guide_2011 . pdf ) . The policies , procedures and guidelines for the NIH IRP are explicitly detailed in NIH Policy Manual 3040–2 , “Animal Care and Use in the Intramural Program” ( PM 3040–2 ) and the NIH Animal Research Advisory Committee Guidelines ( ARAC Guidelines ) . Those documents are posted on the NIH Office of Animal Care and Use public website at: http://oacu . od . nih . gov . Lungs and nasal turbinates collected for viral titration were stored at -80°C until processing . Tissues were weighed and homogenized in L15 media containing 1% antibiotic-antimycotic ( Invitrogen ) to a final 10% wt/vol . Homogenates were centrifuged for 10 minutes at 1500 rpm with a swinging bucket rotor ( Sorvall 75006445 ) . Viral RNA was then isolated from the homogenates using the QIAmp viral RNA mini kit ( Qiagen ) following manufacturer’s instructions . qRT-PCR reactions were amplified using 200ng of RNA per reaction with primer sets designed to detect MERS-CoV via the viral envelope ( UpE ) or nucleocapsid ( N2 and N3 ) protein with the SuperScript III Platinum One-Step qRT-PCR kit ( Life Technologies ) [30] . Results are displayed using N3 primers , the confirmatory primer set . A sample from a naïve rabbit was always run to verify no background was detected with the N3 primer set . A standard dilution set of a titered virus stock was run in parallel , and all samples were tested in duplicate . Titers are expressed as log10 TCID50 equivalents per gram of tissue . Lung tissue samples from all lobes were resected from formalin-fixed tissue . Tissue was embedded in paraffin , sectioned at 5-μm , and stained with hematoxylin and eosin ( Histoserv , Germantown , Maryland ) . Sections were examined by light microscopy ( LM ) or fluorescence microscopy ( FM ) , using an Olympus BX51 microscope , and photomicrographs were taken using an Olympus DP73 ( LM ) camera or DP80 camera ( FM ) . All histopathology scoring of tissues was blinded . Lung sections were baked at 60°C for 1 hour then paraffin was removed with xylene and the sample was rehydrated with alcohol-gradated washes . Sections were microwaved with Antigen Unmasking Solution ( Vector Laboratories ) , and then exposed to protein block ( Dako ) for 30 minutes . For immunohistochemistry ( IHC ) mouse anti-MERS nucleocapsid protein ( NP ) antibody ( Biorbyt ) was added at a dilution of 1:100 , followed by biotinylated horse anti-mouse immunoglobulin G ( IgG; Vector Laboratories ) at a dilution of 1:200 . Rat anti-CD3 ( AbD Serotec; 1:100 dilution ) and goat anti-DPP4/CD26 ( R&D Systems; 1:25 dilution ) antibodies were followed by a hydrogen peroxide blocking step for endogenous peroxidase activity , and then respective biotinylated goat anti-rat and horse anti-goat IgG antibodies ( Vector Laboratories ) at dilutions of 1:200 . Detection of MERS NP was completed with incubations of 30 minutes with Vectastain ABC-AP reagent ( Vector Laboratories ) and 25 minutes with Vulcan Fast Red ( Biocare ) . Detection of CD3 and DPP4 was completed with incubations of 30 minutes with Vectastain ABC RTU ( Vector Laboratories ) and 7 . 5 minutes with DAB . Immunofluorescence Antibody Assay ( IFA ) differentiated after the primary antibody incubations . Complement C9 antibody ( MyBioSource ) was added at a dilution of 1:50 , followed by goat anti-guinea pig IgG ( Vector Laboratories ) at 1:200 and streptavidin conjugated to AlexaFluor 594 ( Life Technologies ) at 1:500 . MERS NP ( same as IHC ) was detected with goat anti-mouse directly conjugated to AlexaFluor 488 at a 1:500 dilution . Slides were counterstained with hematoxylin ( IHC ) or DAPI ( IFA ) and evaluated by a veterinary pathologist . Image analysis was performed on MERS-CoV infected lung tissues to provide a quantitative analysis of the MERS virus antigen present in the lung and the associated inflammatory response . Following histological ( Hematoxylin and Eosin; H&E ) and immunohistochemical evaluation of the lung sections , tissue slides were digitized with a bright-field Leica Aperio AT2 slide scanner at 40x magnification ( S3 and S4 Figs ) . The images were evaluated using a web-based digital pathology information management system ( eslide manager ) used for both digital slide viewing and image analysis . The Aperio Positive Pixel Count ( PPC ) and Color Deconvolution V9 based algorithms were adjusted to recognize the inflammatory regions and the intensity of Vulcan Fast Red chromogen ( Biocare Medical ) , respectively . In the previously stained and scanned lung sections , analysis was based on the settings of Hue Value and Hue Width ( PPC ) or values of the red , blue , and green channels ( CD ) ; gating and selection of regions of interest prevented the incorporation of nonspecific staining in the results . After these macros were optimized to suit the desired application , the settings were saved and were used for the evaluation of all the slides . The channel parameters for the H&E and MERS-specific macros were as follows: MERS-CoV H&E ( PPC ) : Hue Value 0 . 647 and Hue Width 0 . 347 and MERS-CoV Fast Red ( CD ) : red component 0 . 561 , green component 0 . 679 , and blue component 0 . 185 . Neutralizing activity in rabbit sera were evaluated by a microneutralization ( MN ) assay . To determine the antibody titers , serial two-fold dilutions of sera were prepared . 100 TCID50 of virus was mixed with the sera in equal volume and incubated for one hour at room temperature , before the mixture was subsequently added in quadruplicate to Vero81 cell monolayer . The serum neutralization titer was determined as the reciprocal of the serum dilution that neutralized virus as evidenced by the absence of any cytopathic effect on day 4 and confirmed on day 6 . To quantify anti-S protein IgG antibodies from rabbit serum , 96-well plates were coated overnight with 100ng/well of recombinant MERS-CoV S protein ( Sino Biological ) in sodium bicarbonate buffer . Subsequently , the plates were blocked for 2 hours at room temperature with 10% FBS in PBS . Plates were washed and incubated for 2 hours with serial four-fold dilutions of heat-inactivated rabbit serum in duplicate . The plates were washed and further incubated at room temp with HRP conjugated goat anti-rabbit IgG ( Abcam ab6721 ) diluted 1:120 , 000 in PBS with 5% BSA and 0 . 05% Tween-20 . For detection , following additional washes , SureBlue TMB Microwell Peroxidase Substrate ( KPL ) was added to each well and TMB BlueSTOP solution ( KPL ) was added after 10 minutes . The optical density of each well was measured at 650 nm on a SpectraMax i3 plate reader ( Molecular Devices ) and an OD greater than two standard deviations above the mean of the background was considered positive . To examine anti-N protein IgG antibodies from rabbit serum , we utilized an ELISA protocol developed by the CDC [61] . Briefly , 96-well plates were coated overnight with purified MERS-CoV N antigen or irrelevant control antigen ( both obtained from Division of Viral Diseases , Centers for Disease Control and Prevention ) in PBS . Plates were then washed and serial four-fold dilutions of heat-inactivated rabbit serum were added for one hour at 37°C . After incubation the plates were washed further and incubated with HRP conjugated goat anti-rabbit IgG ( Abcam ab6721 ) diluted 1:120 , 000 in PBS with 5% BSA and 0 . 05% Tween-20 for one hour at 37°C . Following additional washes , positive sera were determined by the addition of ABTS Peroxidase substrate solution ( KPL ) that was incubated for 30 minutes at 37°C , followed by the addition of ABTS stop solution ( KPL ) . The optical density of each well was measured at 405 nm on a SpectraMax i3 plate reader ( Molecular Devices ) and an OD of 0 . 3 above the negative control was considered positive . To compare the amounts of C3a present in rabbit lungs following infection with MERS-CoV , we utilized the Rabbit Complement Fragment 3a ( C3a ) ELISA kit ( MBS703171 , MyBioSource ) , according to manufacturer’s instructions . Frozen rabbit lung samples were rinsed and homogenized to 10% w/v in PBS , and stored overnight at -20°C . Following two freeze-thaw cycles , the samples were centrifuges at 5000g for 5 minutes and the supernatants were assayed immediately , in triplicate . Undiluted samples were added to pre-coated plates for 2hrs at 37°C . The samples were removed , and the biotin-antibody was added for 1hr at 37°C . The plates were then washed 3 times before the addition of the HRP-avidin antibody for 1hr at 37°C . The plates were washed 5 times , before incubation with the TMB Substrate for 20min at 37°C . Stop solution was then added , and the OD was measured at 450nm within 5 minutes . Samples were quantitated based on a standard dilution series within the plate . To determine if antibodies resulted in increased viral replication in macrophages , we conducted an antibody-dependent enhancement ( ADE ) assay . THP-1 cells were differentiated into macrophages by addition of 20nM PMA into the RPMI media for 24 hours , followed by a week of culturing without PMA . The cells became adherent to the flask , and took on a macrophage-like appearance . The differentiation of THP-1 cells was confirmed by immunofluorescence with the loss of CD14 and increase of CD36 , CD68 , and CD71 on the cell surface compared to undifferentiated THP-1 cells , adapted from Genin et al [62] . For a positive control , we infected Vero81 cells . As a negative control , we included Raji cells , which lack both DPP4 and Fc receptors . Heat-inactivated rabbit sera at three dilutions ( undiluted , 1:10 , and 1:100 ) were incubated with EMC/2012 at an MOI of 1 for 1 hour at 37°C before addition onto each cell type in duplicate in 96-well plates for 2 hours at 37°C . Cells were then washed and incubated for 48 hours before supernatants were collected for viral titration . Mean viral titers are displayed with the standard error of the mean . Statistical significance was determined using one-way ANOVA with multiple comparisons tests in GraphPad Prism v7 .
|
New Zealand white rabbits display an increase in lung inflammation following reinfection with MERS-CoV that is associated with non-neutralizing antibodies and complement proteins . The development of neutralizing antibodies resulted in protection from infection . These findings may have implications for individuals that fail to develop a neutralizing antibody response , or for those whose response wanes over time , upon re-exposure to MERS-CoV .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"complement",
"system",
"medicine",
"and",
"health",
"sciences",
"coronaviruses",
"immune",
"physiology",
"enzyme-linked",
"immunoassays",
"pathology",
"and",
"laboratory",
"medicine",
"respiratory",
"infections",
"pathogens",
"immunology",
"microbiology",
"rabbits",
"vertebrates",
"animals",
"mammals",
"pulmonology",
"viruses",
"animal",
"models",
"rna",
"viruses",
"signs",
"and",
"symptoms",
"experimental",
"organism",
"systems",
"antibodies",
"immunologic",
"techniques",
"research",
"and",
"analysis",
"methods",
"immune",
"system",
"proteins",
"inflammation",
"proteins",
"medical",
"microbiology",
"microbial",
"pathogens",
"immunoassays",
"immune",
"response",
"immune",
"system",
"biochemistry",
"immunohistochemistry",
"techniques",
"diagnostic",
"medicine",
"viral",
"pathogens",
"physiology",
"biology",
"and",
"life",
"sciences",
"amniotes",
"histochemistry",
"and",
"cytochemistry",
"techniques",
"organisms"
] |
2017
|
Enhanced inflammation in New Zealand white rabbits when MERS-CoV reinfection occurs in the absence of neutralizing antibody
|
Although bacteria are unicellular organisms , they have the ability to act in concert by synthesizing and detecting small diffusing autoinducer molecules . The phenomenon , known as quorum sensing , has mainly been proposed to serve as a means for cell-density measurement . Here , we use a cell-based model of growing bacterial microcolonies to investigate a quorum-sensing mechanism at a single cell level . We show that the model indeed predicts a density-dependent behavior , highly dependent on local cell-clustering and the geometry of the space where the colony is evolving . We analyze the molecular network with two positive feedback loops to find the multistability regions and show how the quorum-sensing mechanism depends on different model parameters . Specifically , we show that the switching capability of the network leads to more constraints on parameters in a natural environment where the bacteria themselves produce autoinducer than compared to situations where autoinducer is introduced externally . The cell-based model also allows us to investigate mixed populations , where non-producing cheater cells are shown to have a fitness advantage , but still cannot completely outcompete producer cells . Simulations , therefore , are able to predict the relative fitness of cheater cells from experiments and can also display and account for the paradoxical phenomenon seen in experiments; even though the cheater cells have a fitness advantage in each of the investigated groups , the overall effect is an increase in the fraction of producer cells . The cell-based type of model presented here together with high-resolution experiments will play an integral role in a more explicit and precise comparison of models and experiments , addressing quorum sensing at a cellular resolution .
Bacteria have evolved signaling networks enabling them to sense the environment by producing , exporting and importing small signaling molecules called autoinducers . By using autoinducers that can rapidly diffuse across cell populations and accumulate over time , bacterial cells can receive information about the cellular density in the surrounding environment . The information can then be used to generate decentralized population-wide responses at high enough cell densities . This phenomenon , known as quorum sensing ( QS ) , has been shown to be important for several biological mechanisms since the initial discovery of it as a regulator of bioluminescence [1]–[3] . In particular , it appears to be a key regulator of several bacterial phenotypes with medical implications , e . g . virulence factor production , biofilm development , and synthesis of antibiotics [4]–[6] . Typically , quorum-sensing Gram-negative bacteria use largely homologous quorum-sensing networks [3] , wherein the autoinducers are acylated homoserine lactones ( AHL ) , detected and regulated via the genetic circuits similar to the LuxIR circuit in Vibrio fischeri ( Figure 1A ) . The lux operon in V . fischeri is positively regulated by AHL , and apart from controlling bioluminescence , it upregulates the expression of the AHL-synthase LuxI . This creates a positive feedback loop that increases AHL production in an AHL-sensitive fashion . LuxR is an AHL-dependent luxI activator , whose dimerized complex with AHL leads to transcriptional activation of the operon [7] , [8] . LuxR has also been implicated in regulation of its own expression [9]–[11] , providing an additional positive feedback loop in the system . The lux operon circuit may be regarded as the central network for controlling QS behavior , but other regulatory mechanisms have also been identified ( see e . g . [12] , [13] ) . Studying QS in detail at a population level introduces some interesting complications . The internal concentration of the autoinducer is dependent not only on its production and degradation , but also on the permeability of the bacterial cell wall as well as on the diffusive properties of the surrounding medium . While the response switch from a low lux gene expression state ( off state ) to its high expression state ( on state ) is easily predictable in experiments where the exogenous autoinducer concentration is controlled , the cell response in the presence of autoinducer auto-regulation is more complex to analyze and understand . For instance , waves of QS signaling might develop or be arrested , depending on the mutual location of signaling cells , as the probability for a cell to be induced might depend on the transport properties of the medium and the signaling levels of the neighboring induced cells . The intracellular switch of the QS network is dependent on the autoinducer concentration just outside the cell , and since this concentration increases with the number of nearby cells even if they are in the basal “off” state , the QS can be switched at high densities of bacteria . However , the autoinducer levels are highly dependent not only on the population size , but also on the degree of local cell clustering and on the geometry of the environment in which the bacteria are growing . Since these parameters are not controllable by the individual bacteria , there is an ongoing discussion as to whether the main benefit derived by cells in QS is from measuring cell density ( or reaching a “quorum” ) , the diffusion of autoinducer away from the cell ( diffusion sensing , DS ) or the potential efficiency of a process metabolically more expensive than secretion of AHL ( efficiency sensing , ES ) [14] , [15] . QS can be beneficial from a population perspective , but since there is a cost associated with ensuring a new beneficial trait for the colony , it is exploitable by the so-called cheater cells , e . g . , those that do not contribute to the production of autoinducer or expression of the QS-regulated operon , but still take advantage of whatever benefit the QS response provides to the colony [15]–[20] . This has recently been highlighted in experiments with mixed populations [21] , [22] . These experiments have measured the relative fitness of cheater cells , depending on the initial ratio of producer and cheater cells within the colony [21] . In particular , an example of Simpson's paradox was seen [22] , wherein QS signal producing cells taken together have a net advantage if cell populations form groups with different initial ratios of producing and cheater cells , even though the producing cells are at a disadvantage in each of the individual groups . Several mathematical models have been used to describe the molecular network central for quorum sensing [23]–[26] . The models have all used networks with single or double positive feedback loops , and assumed different regulatory mechanisms of luxI via the AHL-LuxR complex . Despite the differences , the models converge in their predictions of a bistable switch-like behavior dependent on the external concentration of the autoinducer . Although the models have provided information on how the intracellular QS-network behaves , the effect at the population level have thus far been excluded in all computational investigations . To be able to investigate the behavior of quorum sensing in a bacterial colony where the autoinducer is produced within the colony , we introduce a model that explicitly includes growing bacteria interacting with each other and the surrounding environment via both molecular and mechanical interactions . The model assumes two positive feedback mechanisms where a dimerized LuxR-AHL complex activates both LuxI and LuxR production similar to recently published models [25] , [26] ( Figure 1A ) . We use a combination of analytical and numerical investigations of the model to explore how for example colony size , local clustering , and confinement , affects the behavior both on the single-cell level as well as on the colony level . In mixed population simulations we investigate the competition between autoinducer producing cells and non-producing cheater cells .
First we analyzed the single-cell system described in Equations 1–4 without transport , i . e . and are set to zero . At the steady state , all derivatives are equal to zero , which gives a set of algebraic equations , which in turn can be simplified into a single equation ( Equation 6 in Methods ) , which can have one or multiple positive real roots . Equation 6 was solved numerically to create bifurcation diagrams in the model parameters ( Figure 1B and C ) . It is clear that within a certain parameter region the system has multiple stable solutions , but this region can be complex with several surrounding monostable regions . Adding intercellular transport and external diffusion is expected to affect the parameters in Equation 6 , so that the system trajectory would be able to move into and out of the multistability region ( s ) ( cf . Figure 1B and C ) . Several single-cell QS models have predicted an -dependent switch-like response of the QS network [23]–[26] . To address the effect of communicating AHL with the cell environment , we first assumed a constant and added transport terms to see how this would affect the equilibrium behavior . This generated two important differences as compared to the non-transport analysis above . The transport out of the cells ( term in Equation 1 ) has the same form as the degradation ( ) , so an increase in outwards transport moves the state of the system towards a monostable “off” state in our equilibrium analysis ( upwards in in Figure 1C ) . The transport into the cells ( ) gives an dependent constant contribution and will hence effectively increase the constant which will move the state towards a monostable “on” state ( right in Figure 1C ) . Hence , the addition of transport terms affects the control parameter values and results in changes with opposite effects . At low extracellular AHL concentrations the outflux can dominate the influx and thus drive the bacteria towards an “off” state , whereas high extracellular AHL concentrations are expected to drive the bacteria towards the “on” state . Note that the analysis above was only for the equilibrium behavior and to investigate the dependence on the external autoinducer concentration in a dynamically growing cell-based model , we next performed simulations wherein first slowly increased , and then decreased ( Video S1 ) . As expected , the colony displayed QS response hysteresis ( Figure 1D ) . For the parameters used here the transition between states was fairly smooth , but for other parameter values the transition can be steeper and can even be irreversible ( Figure S1 ) . The requirement to have the switching capability in the QS network due to changes in external does not put severe constraints on the model . As long as an “off” state is available at low concentrations , a sufficient increase in will always lead to a switch to an “on” state , due to the dependent increase of the constant term in Equation 1 ( , cf . Figure 1C and Equation 8 in Methods ) . In nature , however , the QS switching is more restricted since it is the bacteria themselves that produce the autoinducers and there is an upper limit of how high the concentrations of can reach within the colony . Furthermore , the switch threshold needs to be reached while the bacteria are still in the “off” state . The production of AHL cannot be so high as to allow a single-cell to switch by itself , but it must be high enough , so that at high enough densities the colony is able to reach the threshold in . A simplified equilibrium analysis of Equations 1–5 including a single external compartment , but multiple cells , leads to a single change from the non-AHL-transport analysis above . The dependent term is changed: , where is the number of bacteria and is the diffusion out from the extracellular milieu ( see Methods ) . As discussed above , an increase of the parameter moves the state of the bacteria towards a stable “off” state , upwards in Figure 1C , and this simplified model shows that adding diffusive interaction with the extracellular domain can only drive the bacteria towards that monostable “off” state . The change due to the addition of transport is bounded ( ) , wherein the lower bound is for . Thus the effect of increasing the population size , , corresponds to decreasing or going downwards in Figure 1C . However , since the contribution is bounded from below , the system can never move beyond , or below , the initial state . Hence , this simplified equilibrium analysis predicts that , in order to have a QS response of the colony , the parameters must be chosen such that a single bacterium without AHL transport is in an “on” state , but close enough to the multistable region to allow inclusion of the transport terms to “move” the single bacterium into its “off” state . The analysis presented is of course for a very simplified description of the QS , and to be able to investigate how QS works in a more realistic non-equilibrium environment , we used a cell-based model and a spatially meshed extracellular domain with dynamically diffusing AHL . We simulated the system of growing communicating bacteria , starting with a single bacterium that grows , divides and communicates with the environment ( see Methods ) . The overall simulation domain was assumed to be a thin rectangular layer , of the same thickness as the bacteria , and with on the boundary . This assumption , in addition to simplifying visualization and analysis of the results , corresponds to the experimental design used to analyze bacterial colony growth in microfluidic devices , thus providing a potential model validation platform [27] . The colony displayed a quorum-sensing behavior with a clear unanimous switch in and at a specific population size ( Figure 2A , see also Figure S2 and Video S2 ) . When the number of cells was small the colony was in an “off” state . At a threshold population size , cells in the spatial center of the colony started to switch on , leading to a short time period of an inhomogeneous colony with cells both in “on” and “off” states , but the switching propagated quickly and soon virtually all cells of the colony were switched on . A reason for the homogeneity of responses of all the cells in the colony is the positive feedbacks ensuring that the production of AHL is much higher in cells that are turned on compared to the constitutive basal production . As soon as a few cells turn the signaling on , the amount of AHL in the environment can quickly rise , driving the fast propagation of the switching throughout the colony . However , the low constant production of AHL is necessary for the initiation of the switching behavior . To investigate the system dependence on the model parameters , we performed a parameter scan and studied how parameter variation affected the colony behavior ( Figures S3 , S4 , S5 ) . In Figure 2B we show the effect of varying the three transport parameters ( , , and ) . We observed that the colony response moves into and out of the bistability region at different population sizes . Specifically , we found that , at low diffusion rates , the system was inclined to switch , whereas at higher diffusion rates the colony was no longer able to accumulate sufficient amount of AHL to make the switch possible , gray line in Figure 2B . This is in accordance with the simplified equilibrium analysis above , wherein changing affected via . Hence , at low , the system is essentially in the situation without AHL transport , whereas at high , we get which might be enough of a change in the effective AHL removal rate to move the system into the “off” state . From the same simplified model it is also clear that and should effectively change in opposite directions , which is also exactly what one observes in Figure 2B . Thus far we have shown that the switching mechanism of QS in individual cells is dependent on the extracellular AHL concentration , and that for a bacterial colony this concentration depends on the net loss of local . Our simplified analysis showed that this loss can be approximated by a change in , given by addition of the term , which explicitly shows that this depends on the outflux ( or loss ) in the exterior ( ) and the density of bacteria ( ) , with the individual bacteria thus not being able to distinguish whether or is changed in the environment . The simulations of the colony growth ( Figure 2 ) also showed that being in the center of a dense population facilitates the QS switch . Taken together , these results demonstrate that the model confirms that bacteria cannot measure cell density , exterior loss of autoinducer , and spatial clustering independently , in agreement with prior qualitative arguments [15] . It has been shown that bacteria often actively seek out small cavities and populate them to very high densities [28] , [29] . To see how local density and confinement might affect colony behavior , we compared dense cell population simulations with simulations of a sparsely populated colony . The populations were simulated with open boundaries as before ( Figure 3A ) . We also considered colonies confined in a small cavity with a single small outlet ( Figure 3A ) . In the simulations , we fixed the population size at different values and examined the resulting QS . The system switched once the population reached a certain number of cells , and the system switch occurred at lower cell number in the dense population than in the sparse population ( Figure 3B , see also Figure S6 ) . This result demonstrated that although QS is generally a population-size effect , it can be facilitated by local clustering of bacteria [15] . We also observed that confinement of the sparse population makes its switching behavior similar to that of the dense population . Thus , not only the local density and the number of bacteria matters for the response of the colony but also the geometry of the surrounding environment . In light of the results in Figure 3B the strategy of populating cavities makes sense as a way of facilitating the onset of quorum sensing . However , the geometry of the cavity may also affect the ability of the colony to perform the switch in concert , e . g . by controlling the escape of AHL . To address this possibility more directly , we performed simulations of colony growth and QS in a cavity geometry similar to previously used microfluidic chambers [27] , but with variable number of outlets . Simulations were initiated with a single bacterium and simulations were run until the expanding colony completely filled up the cavity ( Figure 4A and Videos S3 , S4 , S5 , S6 ) . At sufficiently high values of the population only partially switched states ( Figure 4A ) . Typically , it is only at the regions furthest away from the exits that the colony was able to accumulate sufficient levels of to undergo the switch . Figure 4B shows the fraction of cells in the “on” state as a function of time , with the clear result of an organized population-dependent behavior . At first no cells are in the on state , but at a colony size determined by the number of outlets , parts of the population make a sudden sharp switch and reaches a new stable configuration ( cf . Videos S3 , S4 , S5 , S6 ) . The bacteria furthest away from the exits are those that initiate the switching behavior . In Figure 4C the and concentrations of individual bacteria are plotted as a function of the spatial position along the horizontal axis for Chamber 2 . At positions far away from the exits the bacteria are homogeneously in the “on” state , while closer to the exits the population is less homogeneous due to the loss of AHL at the exits ( the upper leg marked with in Figure 4C ) . Note also that the signaling in the chamber legs between two exits is tightly concentrated around ( e . g . in Figure 4C ) . The system can show multi-stable responses and the cells in these legs are clearly at the stable fixed point wherein the production has switched on while the production has not . Taken together our simulations show a complex behavior with the switching of each bacterium being dependent on its location within the chamber , and with local subpopulations with high signaling homogeneity created . In natural habitats , bacteria live in environments with a mixture of different bacterial strains . This property can affect the QS behavior and lead to a problem of emergence of cheater cells that can exploit the “common good” produced by the QS population . The phenomenon was recently studied in controlled environments for bacteria [21] , [22] . These cheater cells do not produce the autoinducer ( or other QS resulting common good molecules ) themselves but do take advantage of the metabolically expensive QS signaling by the rest of the population . By not participating in the generation of QS response , cheater cells can instead use metabolic energy to more rapidly grow and divide . We considered this situation by modeling cheater cells as the other bona fide signaling cells , but with no production of ( ) . Furthermore , we assumed that once the normal cells switch on and thus increase their QS response , their growth rate slows down ( see Methods ) . Data from simulations in a confining chamber starting from different initial states are presented in Figure 5A where we tracked the population dynamics in the mixed colonies . Initially the growth rates of the producer and cheater sub-populations were equivalent , but once some of the producer cells switched states , the cheater cell population rapidly started to dominate the chamber ( cf . Video S7 ) . Note that although the fraction of producing cells that turned on was quite small ( about 10% , dashed-dotted line in Figure 5A ) , this was sufficient to break the symmetry and give the cheater cells a clear advantage . The domination of cheater cells leads to a dilution of producing cells which lowers the AHL concentration in the chamber . This resulted in a decrease in the number of producer cells that were switched on and thus diminished the advantage of the cheater cells . In the end the relative cell numbers of the two sub-populations can stabilize . The simulations in the other chambers displayed similar behaviors ( see Figure S7 ) . The dynamics of the colonies ( Figure 5A ) clearly showed that whether or not the cheater cells were at an advantage , depended on the composition of the mixture of cheater cells and normal cells [21] . To investigate this further we performed simulations wherein the initial colony consisted of different ratios of cheater and producer cells . In these simulation we added the assumption that the producing cells could provide the population with some advantage or “common good” , a property beneficial for the survival and growth of the population as a whole . In the model this was simplified by assuming an autoinducer dependent growth rate ( see Methods ) . In Figure 5B the resulting relative fitness of the cheater cells is displayed , indicating decreasing advantage for increasing initial ratios of cheater cells , as seen in experiments [21] . The model predicts that the advantage of cheater cells is directly related to the number of producing cells that are in the “on” state , which in turn is dependent on the number and location of the cheater cells . This leads to an effective negative feedback , so that the producer cells are not completely overtaken by the cheater cells in any of the cases in Figure 5B . In fact , the addition of the AHL-dependent growth does not alter the relative fitness behavior in the individual simulations ( data not shown ) but actually leads to a total increase of producer cells if all initial colony configurations are summed up ( Figure 5C ) . Although cheater cells always have a local advantage and never grow slower than the producing cells , the colonies with more producer cells will grow faster and this is sufficient for generating more producer cells in total . This has recently been reported for synthetic bacteria strains and is referred to as the Simpson's paradox [22] . The simulations with mixed populations show that cheater cells may have a local advantage , but a negative feedback via the colony growth and dilution of producing cells leads to a situation where this advantage is only transient .
Quorum sensing is a key example of the ability of unicellular bacteria to act not only as individual cells but also as an ensemble , resembling in many respects a multicellular organism . This collective cell behavior phenomenon is important for various biological behaviors , with considerable implications for the physiology and pathology of plants and animals [13] . Hence it merits further understanding both for a better appreciation of the fundamental properties of cell-cell communication and for its applications . With the increasing amount of quantitative data for the molecular networks at the center of the cellular QS signaling , the use of mathematical models has emerged as an important tool for understanding how the molecular network structure with its multiple feedbacks can explain the complex behavior of the population . Previous models have mainly discussed the intracellular network with the underlying QS switch , and have treated the extracellular environment as a boundary condition [23]–[26] . An exception is the static model briefly described in Hense et al . 2007 [15] . Recent development of microscopy techniques together with the increased use of microfluidic devices have increased the ability to study cell colony behaviors at a cellular resolution [30] . Here we have presented a model explicitly taking into account individual growing bacteria as well as the transport and geometry of the extracellular milieu . This resulted in a model framework with the results directly comparable with data from cell-based experiments in microfluidic devices and other experimental settings , and allowed for an explicit investigation of how population-level behavior emerges from single-cell mechanisms . In this report , we presented simulations investigating cell-to-cell variations in homogeneous populations as well as the behavior of mixed populations . An equilibrium analysis of the model was used to find the parameter values capable of population-size dependent QS switching and the analysis highlighted the differences between a situation where autoinducer levels are tuned extracellularly and when bacteria themselves are the only source of the autoinducer . In the former case , we showed that QS switching was not very constrained . However , in the latter case , the effect of adding the autoinducer transport boiled down to variation of a single parameter of the model: the effective degradation of the autoinducer . The variation of the effective degradation was shown to be dependent on the transport parameters characterizing the autoinducer and the cell medium , and on the number of bacteria present , and was shown to be bounded by the rate of autoinducer transport out of the cells . Hence , the ability of QS switching is only ensured if this bounded parameter can change so that the systems can visit both “on” and “off” states . A clear prediction from this analysis is that if autoinducer membrane transport is blocked , the cells would have be to be in an “on” state . Simulations of growing and proliferating bacteria showed a population-size dependent switching behavior , wherein although it is the bacteria in the center of the colony that initially switch on , the whole colony quickly follows creating a very homogeneous behavior . This is mainly due to the strong positive feedback in the signaling system , ensuring that the autoinducer production greatly increases in the cells that are switched on . A scanning of the model parameters orders of magnitudes around their initial values showed that the main QS feature , the population switching , is very robust , while the actual population size where the switch happens is quite dependent on parameter values . We further found , as expected , that the switching of the population is driven by the external autoinducer concentration . This is dependent on the population size , but also on how much autoinducer is lost from the colony , which depends on the local density ( clustering ) and the confinement of the external geometry; parameters that to a large extent are beyond the control of single cells . We explored these parameters explicitly in our model simulations showing that growing dense populations in small confined cavities facilitates population switching , a potentially common strategy [28] , [29] . This relates to the discussion of the evolutionary fitness advantage provided by a collective cell population behavior , with the quorum sensing , diffusion sensing and efficiency sensing have been suggested as different explanations [15] . Our model suggests that cells can sense different aspects of their environment through determination of the value of a single , albeit complex parameter ( ) , comprising all these different possibilities . Additionally , the model suggests that a possible evolutionarily selectable strategy of populating small cavities as a means to control diffusion , local density , and confinement in order to facilitate the onset of quorum sensing . Bacteria live in environments where different biological organisms compete . It has been noted that a QS behavior can be exploited by strains of cheater cells that do not participate in some aspects of QS , but still take advantage of the benefits this provides . The corresponding advantages of this behavior for cheater cells have recently been investigated in controlled experiments [21] , [22] . The cell-based approach allowed us to investigate competition between autoinducer producing and non-producing cells by adding a growth reduction for producing cells that are in their “on” state . We showed that the cheater cells did have an advantage as soon as producing cells switched on . This advantage , however , led to a dilution of producing cells , and hence the amount of autoinducer per cell , within the mixed population as the cheater cells increased their relative number . The decrease in autoinducer further led to producing cells switching “off” , which diminished the cheater cell advantage . Hence , the growth dynamics in these mixed populations creates a feedback that disallows a cheater strain to fully overtake a population . If we assumed also that the growth was dependent on the production of autoinducer or the corresponding beneficial population trait ( e . g . , the ability to cause the host to provide nutrients ) , we could observe situations where populations initiated with different ratios of cheater cells generated an overall advantage for producing cells , although in each individual local sub-population , cheater cells were never at an disadvantage . The phenomenon is known in statistics as the Simpson's paradox , and was recently demonstrated for synthetic bacterial strains [22] . The number of molecules , including members of the transcription machinery present in bacteria can be very low . Hence it is expected that effective transcription and reaction rates might be noisy , and segregation of the transcription factor molecules into the daughter cells at cell division can be inhomogeneous [31] . Interestingly , a test with complete random placement of all molecular species at division had very minor effects for the cell population ( data not shown ) . This shows a model robustness of the population behavior to molecular fluctuations in individual cells , but it also points out a limitation of our deterministic approach . In the deterministic model , a switch from a low stable state to a high stable state does not spontaneously happen in the bistable region . Hence , to get a switch in the simulations a change of condition ( e . g . increasing the number of cells ) will need to move the system into the monostable high region of the state space . A fluctuation in concentrations at division will then quickly move back to the only stable state . In a stochastic model , on the other hand , it could be enough to be in the bistable region where switching between high and low states could be initiated by fluctuations in concentrations . Given the number of bacteria , external compartments , and reactions in our simulations , a complete stochastic treatment may be out of reach , but an interesting future improvement would be to add stochasticity to the model , for example via adding noise terms to the ODEs . Recent experimental developments have changed our ability to quantify cell states , from the population averages to the dynamics of single cells . The presented work is important since it represents the same development for the mathematical models used to analyze cell-based behavior . The combination of high-resolution experiments where colonies are grown in regulated environments , and models where single cells are growing to form colonies will help understanding of how population dependent behaviors , such as quorum sensing , can be derived from single cell molecular networks .
Following earlier efforts [27] , [32] each cell is modeled as an individual object , described as two semi-spheres attached at opposite sides of a cylinder . The dynamics of the bacteria is governed by a potential , where the different contributions describe cell-cell interactions , cell-wall interactions and the internal potential respectively . We further assume that the dynamics of the colonies is dominated by viscous friction so the equations of motion for a given cell is described bywhere and are the two coordinates , chosen as the centers of the two semi-spheres , is the friction coefficient and denotes the derivative with respect to and respectively . For the friction coefficient , we assume a generalization of Stokes' formula [33]where is the distance between the sphere-centers , is the radius of the sphere , is the unit direction of the main axis of the bacterium and is the unit direction of its velocity . For the cell-cell interaction and the cell-wall interaction we use an excluded volume like potential where the potential is given bywhere denotes the set of neighbors to cell and is the linear overlap between a cell and a cell [34] . The interactions with the chamber walls are modeled in the same way , but with the only difference that the walls are assumed to be static . The internal potential is a spring potential which is introduced to allow the coordinates to be treated as two separate degrees of freedom . where is a constant and is the rest length . The cells grow exponentially along the symmetry axis according to , As the cells grow the intracellular molecular concentrations will decrease because of dilution . This dilution corresponds to an extra degradation term in Equations 1–4 with degradation constant . Once a cell reaches a certain threshold length , it divides into two cells of almost the same size . At each cell division we introduce some randomness in order to break the axial symmetry of the system , giving two daughter cells with slightly different sizes and imperfect alignment [27] . The cell-surrounding medium is modeled explicitly by dividing the space into small elements . The autoinducer molecule , , can penetrate the cell walls of the bacteria , which is modeled as a flux of given bywhere is the concentration of the element enclosing the center-of-mass of the bacterium . The contributions to the derivatives are given bywhere and are the volumes of the bacterium and the element respectively and is the surface area of the bacterium . The diffusion in the extracellular medium domain is modeled via Fick's law with a finite-difference version of the normal diffusion equation , the derivative of an element is given bywhere denotes the neighbors to element , is the area between elements and and is the distance between the two elements . The quorum-response of the colony typically leads to the production of some “common good” or trait that is beneficial to the population as a whole , In the simulations leading up to Figure 5B–C we simplify this somewhat by having a direct autoinducer dependence in the growth function , where we set slightly below the peak value of , in Figure 5B–C we use . We use the same value to define if a cell is “on” or “off” , thus cells with are considered to be in their “on” state . In order to model the cost of autoinducer production , we multiply the growth-rate of all the “on” cells of the system by a factor , . In Figure 5A we use and in Figure 5B–C we use . The relative fitness measure of Figure 5B , is defined aswhere is the initial fraction of cheater cells , is the fraction for producer cells , and and are the fractions at the end of the simulation . Numerical simulations where done using an in-house developed C++ software package specifically developed to handle proliferating cells and background compartments . The differential equations are numerically solved using a fourth-order Runge-Kutta solver [35] . The software is available upon request . In order to obtain the equilibrium behavior we set Equations 1–4 to zero . This leads to two coupled equations , where , , , , , and , which can be combined into ( 6 ) where , , , , . Equation 6 is obtained by setting , where , and the equation is solved numerically by finding the roots to ( 7 ) We took advantage of the fact that and that as by bracketing the solutions starting by choosing two small regions , one around and one around a sufficiently big value of . We extended these regions until had different signs at each endpoint . This provided us with two regions where it was known that Equation 7 had solutions which could be found by a simple bisection search . By comparing the two solutions we knew whether Equation 7 had one or several solutions , see Figure 1B Equation 6 had grouped parameters to parameterize the equilibrium solutions using only four parameters . We also used the original parameters from the model , and described the equilibrium solutions to Equations 1–4 with a similar equation ( 8 ) where and again is defined as . Equation 8 was solved numerically in the same way as discussed for Equation 6 above , to generate the bifurcation diagram in Figure 1C . To address the effect of the autoinducer transport into and out of the extracellular environment and to examine the effect of multiple bacteria in the system we considered two simplified cases: ( i ) assuming a constant external concentration , leading to a change given by and in Equation 8 and ( ii ) a simplified description using a single background compartment and assuming identical bacteria , with a constant transport of out of the compartment leading toThis leads to the equilibrium conditionThe change compared to the single-cell analysis is thus given bywhere ( The lower bound comes from and the upper from ) . At low the transport will lead to a movement towards the monostable off region , i . e . for a situation with few cells the transport can lead to that a cell ( which without transport would have been in its on state ) is off . The effect of increasing the population size , , will have the opposite effect , moving it back towards the monostable on region . However , since the contribution tends to zero in the limit of big population sizes , we can never move beyond the single-cell case with no transport . This means that we must choose our single cell parameters in the monostable high region if we want a quorum-sensing response of the system .
|
Unicellular organisms have the ability to communicate with each other via signaling molecules , leading to correlated behaviors resembling that of higher organisms . This process , called quorum sensing , allows the cells to monitor the population size or density in a decentralized fashion and perform a common task when these parameters exceed predefined threshold values . The quorum sensing mechanism has been implicated in diverse functions such as producing bioluminescence , virulence factors , and initiating biofilm formation . Complex emergent behaviors , such as quorum sensing , can be hard to analyze and understand without the assistance of mathematical and computational models . Here , we present a cell-based model of proliferating bacterial microcolonies and investigate how population-level responses can emerge from the signaling and mechanical properties of individual cells . We study both signaling variations within homogeneous ( homotypic ) bacterial populations as well as signaling and competition in mixed heterotypic populations . We investigate in particular how population size , local cell density , and spatial confinement affect colony growth and predict strategies for facilitating quorum sensing . We also show that the interplay between “honest” quorum sensing signal producing bacteria and non-producing “cheaters” can lead to emergent feedback regulation via differentiated growth that provides only a transient benefit for cheating cells .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"biophysics/theory",
"and",
"simulation",
"computational",
"biology/systems",
"biology",
"cell",
"biology/cell",
"signaling"
] |
2010
|
A Cell-Based Model for Quorum Sensing in Heterogeneous Bacterial Colonies
|
The globally distributed ectoparasite Varroa destructor is a vector for viral pathogens of the Western honeybee ( Apis mellifera ) , in particular the Iflavirus Deformed Wing Virus ( DWV ) . In the absence of Varroa low levels DWV occur , generally causing asymptomatic infections . Conversely , Varroa-infested colonies show markedly elevated virus levels , increased overwintering colony losses , with impairment of pupal development and symptomatic workers . To determine whether changes in the virus population were due Varroa amplifying and introducing virulent virus strains and/or suppressing the host immune responses , we exposed Varroa-naïve larvae to oral and Varroa-transmitted DWV . We monitored virus levels and diversity in developing pupae and associated Varroa , the resulting RNAi response and transcriptome changes in the host . Exposed pupae were stratified by Varroa association ( presence/absence ) and virus levels ( low/high ) into three groups . Varroa-free pupae all exhibited low levels of a highly diverse DWV population , with those exposed per os ( group NV ) exhibiting changes in the population composition . Varroa-associated pupae exhibited either low levels of a diverse DWV population ( group VL ) or high levels of a near-clonal virulent variant of DWV ( group VH ) . These groups and unexposed controls ( C ) could be also discriminated by principal component analysis of the transcriptome changes observed , which included several genes involved in development and the immune response . All Varroa tested contained a diverse replicating DWV population implying the virulent variant present in group VH , and predominating in RNA-seq analysis of temporally and geographically separate Varroa-infested colonies , was selected upon transmission from Varroa , a conclusion supported by direct injection of pupae in vitro with mixed virus populations . Identification of a virulent variant of DWV , the role of Varroa in its transmission and the resulting host transcriptome changes furthers our understanding of this important viral pathogen of honeybees .
Host-pathogen interactions can be broadly divided into asymptomatic or symptomatic infections [1] . In the former , the absence of symptomatic disease is typically due to restricted pathogen replication , which reduces the opportunities for horizontal transmission within its host population . Conversely , prolonged survival of the infected host increases the likelihood of vertical transmission of the pathogen [2] . In contrast , symptomatic infections are typically characterized by high levels of pathogen replication , with consequent enhanced virulence , thereby maximizing horizontal transmission [1]–[4] . The ‘lifestyle choice’ of asymptomatic or symptomatic infection is determined by multiple factors including the duration of host-pathogen co-evolution , host physiology and anti-pathogen responses , routes of transmission and environmental factors . Evolutionary changes in pathogen virulence may be triggered by changes in pathogen-host assemblages [5] . In the case of multi-host pathogens with interspecies transmission , a pathogen's virulence may increase following introduction of a second host , when the constraint on pathogen virulence in a given host is removed [6] . The European honeybee ( Apis mellifera ) is the predominant managed pollinating insect and delivers economically important pollination services for agriculture which are estimated to add ∼$40bn globally to crop value/annum [7] . Factors that influence colony health and viability are therefore important for colony survival and pollination performance . In addition to the bacterial foulbroods , the most important diseases of A . mellifera are caused by a range of viruses many of which are vectored by the ectoparasitic mite Varroa destructor when feeding on honeybee haemolymph . Varroa is believed to have expanded its host range from Apis cerana to A . mellifera during the first half of the 20th century and subsequently spread to all beekeeping regions of the world with the exception of Australia [8]–[11] . Deformed wing virus ( DWV ) , a picorna-like single-stranded , positive-sense , RNA virus [12] , [13] , is present in the majority of honeybee colonies [10] . DWV is closely related to Varroa destructor virus type 1 ( VDV-1 ) [14] . Their recombinants [15] , [16] and Kakugo virus ( KV ) [17] , which together exhibit at least 84% nucleotide identity , can be considered as strains of the same virus ( henceforth we use the term DWV to refer to this related group of viruses ) . In the absence of Varroa , DWV generally causes asymptomatic infection and is present at low levels in honeybees . In contrast , in Varroa-infested colonies , mite-exposed pupae can exhibit very high DWV levels which may result in impaired development of the teneral adult honeybee and increased mortality in all honeybees in these colonies , including asymptomatic adults [10] , [13] . The mechanisms underlying the transition of DWV from a relatively benign virus to a major honeybee pathogen in the presence of Varroa remain unclear . Two possibilities , not mutually exclusive , have been proposed: suppression of honeybee antivirus defences by Varroa mites which allows the virus to proliferate [18] , [19] , and a Varroa-driven selection of particular DWV genotypes , potentially due to replication in the mite [15] , [20] . Previous studies using functional or gene expression analysis have produced contradictory conclusions on the impact of Varroa on the immune responses of honeybees . Initial reports indicated that Varroa-exposed honeybees were immuno-compromised [18] , [19] , although later transcriptome analysis found little or no effect on genes implicated in insect immunity [21] , [22] . Additional studies have shown down-regulation of a honeybee NF-κB transcription factor [23] . Recent reports have implicated the Drosophila Toll , Imd and Jak-Stat signalling pathways in controlling RNA virus infection [24] and RNA interference ( RNAi ) , which has long been considered the major antiviral mechanism in insects [25] , has recently been associated with controlling the persistence of RNA virus infections in Drosophila [26] . It was therefore possible that high levels of DWV in Varroa-exposed honeybees could be the result of a suppression of these antivirus responses and so warranted further analysis . We have previously demonstrated that Varroa infestation is associated with the accumulation in mite-exposed pupae of a particular subset of DWV-like viruses [15] . These recombinant forms ( RF ) are predominantly comprised of genomes with structural and non-structural coding regions that most closely align with VDV-1 and DWV respectively . The organisation of these recombinants suggests that , as with other picorna-like viruses , DWV likely has a modular genome , with a 5′ untranslated region ( 5′-UTR ) driving translation of the structural or capsid ( CP ) and non-structural ( NS ) ‘modules’ [15] . We hypothesised that such recombinants were transmitted more efficiently between Varroa and honeybees , resulting in their amplification to the markedly elevated levels observed in Varroa-parasitized pupae ( about 1000 times higher than in unexposed pupae ) . In recent complementary studies , changes in the composition of the DWV population over a large temporal and spatial scale following Varroa infestation were reported for honeybees colonies following accidental introduction of Varroa into the Hawaiian islands [20] . The introduction of Varroa was associated with a marked restriction in DWV diversity measured in the pooled honeybee samples collected from the Varroa-infested colonies , although the precise identity of the dominant virus was not determined [20] . In the present study we devised a novel experimental system to specifically test two hypotheses on the role of Varroa in the development of high-level DWV infection in the honeybee , namely that the mite ( i ) amplifies and transmits virulent genotypes of DWV , and ( ii ) suppresses antiviral responses , including immune signalling pathways and RNA interference . The experimental procedure included exposure of Varroa-naïve honeybees to mites and their associated DWV payload together with the per os in-hive horizontal transmission . The use of Varroa-naïve honeybees from a Varroa-free region allowed us to monitor changes in DWV diversity and loads , as well as potential antivirus responses in the honeybee responses , following exposure to the viral genotypes associated with Varroa infestation . Importantly , we analysed immune responses and viral load/diversity in individual mite-exposed and –unexposed pupae , rather than in pooled samples . This allowed us to stratify individual responses into four distinct experimental groups , characterised by Varroa exposure and viral load , that clearly correlated with characteristic changes in the transcriptome and virus population diversity . In addition , we recapitulated the exposure of Varroa-free honeybees to DWV by direct injection and analysed virus diversity in bees of a colony with long-established Varroa infestation . Our results indicate that a virulent recombinant form of DWV , while transmissible orally , only replicates to high levels when directly inoculated into honeybee haemolymph – by Varroa or experimental injection . This results in massive reduction of DWV diversity in bees with high virus levels , both in the Varroa-exposed pupae and newly emerged bees with symptomatic deformed wing disease . Significantly , the same virulent recombinant form of DWV reached the highest levels in mite-exposed pupae and in adult bees exhibiting characteristic deformed wing symptoms . Although exposure to Varroa resulted in changes in expression of a number of immune-related genes , the roles of which should be further explored , we demonstrate that it is the route of virus acquisition that is responsible for the amplification of this virulent form of DWV in a Varroa-infested colony .
Worker honeybee larvae from a Varroa-free colony ( sourced from a region with no historic contacts with or presence of Varroa ) were moved in a frame transfer experiment to a Varroa-infested colony . The larvae were subsequently exposed through feeding to DWV strains circulating in the infested colony from day 4 until the cells were capped at day 9 ( all times relative to egg laying; Figure 1 , Treatment 1 ) . Varroa mites enter brood cells immediately prior to capping . Therefore , pupae located within brood cells that contain Varroa mites are also subjected to the mite feeding on haemolymph during pupal development ( Figure 1 , Treatment 2 ) until sampling on day 15 ( the purple-eye stage ) , six days after cell capping . Feeding of the mites ( adult females ) on pupae was confirmed by the presence of at least one protonymph in the capped cell [27] . We assessed the total levels of DWV viruses in 80 individual pupae by qRT-PCR using a primer pair for a conserved polymerase-coding region , designed to detect all known DWV strains , including DWV , VDV-1 and KV ( Table S1 ) . The real-time PCR Ct values showed a clear bimodal distribution , with distinct low- and high-levels of DWV ( p<10−16; Figure 1 , Figure S1 ) . Low DWV levels were observed in all ( n = 23 ) sampled pupae maintained in the Varroa-free colony ( group C , “Control” ) , in all 19 sampled pupae transferred to the Varroa-infested colony that were not capped with a Varroa mite and therefore subjected only to Treatment 1 ( oral DWV infection; group NV , “No Varroa” ) , and in 10 of 33 pupae upon which Varroa mites had fed , Treatment 2 ( group VL , “Varroa Low” ) . In contrast , high levels of DWV-like viruses were detected in the remaining 23 of 33 Varroa-associated pupae , which experienced both Treatment 1 and Treatment 2 ( group VH , “Varroa High” ) . The Ct ranges for the VH group lay entirely below the VL range , indicating significantly higher virus levels in VH ( Figure S1A ) whereas the Ct values in groups C , NV and VL were statistically indistinguishable ( ANOVA , p = 0 . 5261 ) . We have previously reported similar proportions of Varroa-associated pupae with low and high levels of DWV in an independent ( temporally and geographically ) study [15] . These results indicate that direct Varroa exposure does not inevitably lead to high , presumed pathogenic , DWV levels , as reported previously [15] , [28] , [29] , at least when age-matched , synchronously exposed pupae are analysed individually . The difference in DWV levels between pupae in the VL and VH groups could not be explained by different mite loads - both contained an average of 1 . 375 adult female Varroa mites per cell ( data not shown ) . These two distinct classes of Varroa-exposed pupae , and their associated mites , were included as separate groups in subsequent analyses to investigate host or parasite determinants that influenced the outcome of exposure . We sampled eight honeybee pupae selected at random from each of the four groups ( C , NV , VH and VL; Figure 1 ) for further analysis . With the exception of the siRNA responses ( for which pooled samples of each of the four groups were used ) , subsequent analysis of transcriptional responses ( microarray transcriptional profiling ) and virus diversity ( qRT-PCR , cloning and sequencing ) were conducted individually on each of the eight pupae from the four response groups . We used a two-colour dye-balanced loop design microarray [30] , [31] to determine the genome-wide transcriptional profile using RNA extracted from the 32 samples defined above , ( 8 pupae from each experimental group ) . The oligonucleotide expression array contained probes to all protein-coding transcripts of A . mellifera [32] , as well as probes to all known viral and fungal pathogens of honeybees , including distinct DWV and VDV-1 probes . After array normalization , differentially expressed ( DE ) genes were determined for each contrast between experimental groups ( Figure 2A , listed in Table S2 with commonalities between contrasts shown in Table S3 ) . Microarray results were validated by qRT-PCR using oligonucleotide primers to a set of honeybee DE genes and the constitutively expressed ribosomal protein 49 ( Rp49 ) gene ( GB10903; Table S1 ) , showing strong positive correlations between the processed microarray signals and normalized Ct values ( Pearson correlation coefficients between 0 . 504 and 0 . 873 ) . Additionally , there was a strong positive correlation between the DWV microarray signal and qRT-PCR Ct values for DWV-like viruses using generic DWV primers ( Table S1 , Primers 59 and 60 ) , Pearson correlation coefficient 0 . 797 . Other than DWV-like viruses , no other honeybee pathogens were detected . There were high levels of commonality and additivity for DE genes in the contrasts considered ( Figure 2A , Figure S2 A , C ) . For example , the C to VH contrast ( in which ∼10% of genes were DE ) can be decomposed into two sub-contrasts by exposure regime , i . e . split C to VH at oral exposure ( NV ) or at mite feeding ( VL ) . Similarly the C to VL contrast can be split at NV . These decompositions exhibit high orthogonality ( Figure 2 A , B; Figure S2 B , C ) . This suggests that expression of essentially different sets of genes are influenced following oral exposure to DWV , Varroa feeding , and the markedly elevated levels of DWV in Varroa-exposed pupae . To explore this further we conducted principal component analysis ( PCA ) . Distinct clustering by experimental group was observed when two independent sets of DE genes with the lowest p-values were analysed i . e . those from the DE genes pooled from all contrasts ( Figure 2C ) , or DE genes in each of six contrasts ( Figure S3 ) . Consequently , PCA strongly suggests that the experimental groups exhibit characteristic gene expression signatures reflecting their fate after exposure in a Varroa-infested colony . To obtain insight into the functional consequences of DE gene expression we carried out Gene Ontology ( GO ) analysis , focusing on the GO Biological Process ( BP ) [33] . A number of overrepresented GO BP terms related to cell division were associated with DE genes in the C to NV contrast , while those related to regulation of various cellular processes were associated with the DE genes in the NV to VL contrast ( Figure 2A , Table S4 ) . Notably , no overrepresented GO BP terms were associated with the genes DE following increase of DWV levels ( VL to VH contrast ) . We then looked in detail at the expression patterns of likely immune-related genes as it had previously been reported that Varroa and/or viruses could influence honeybee immunity [18] , [19] , [23] . The list of 381 putative honeybee immune-related genes included those previously published [34] , [35] together with honeybee homologs of the Drosophila genes associated with the GO term “Immune system process” ( GO: 0002376 ) . The C to VH and C to VL contrasts exhibited the highest number of DE immune-related genes ( n = 42 and n = 26 respectively , 22 of the latter also being in the C to VH contrast ) , whereas oral exposure ( C to NV ) resulted in 18 DE immune-related genes ( Table 1 , Figure S4 ) . Independent confirmation of DE of immune-related genes was obtained by qRT-PCR analysis of persephone protease ( GB14044 ) , Tollo ( GB10640 ) , and Vago ( GB10896 ) with Pearson correlation coefficients of 0 . 598 , 0 . 504 and 0 . 692 respectively . Although no significant changes in expression of genes associated with the RNAi response ( e . g . Argonaute , Dicer ) [25] were observed in the microarray analysis , there could be post-transcriptional effects on RNAi generation . We therefore analysed the DWV-related RNAi population and compared it with the levels and identity of virus in pupae from the four experimental groups ( Figure 1 ) . Small RNA fractions ( 15 to 40 nt ) were isolated from total RNA samples , pooled according to the experimental groups and used as templates for Illumina high-throughput sequencing . One library was generated for the group C honeybees and two libraries for each of the other groups . These libraries , each containing 11 to 35 million reads , were aligned to the reference viral sequences ( DWV and VDV-1 , GenBank accession numbers GU109335 and AY251269 respectively; Figure 3 ) , as well as to the honeybee miRNA sequences [36] , using Bowtie [37] . All RNA libraries analysed contained similar proportions of host-encoded miRNA reads , 12 to 18% of the total ( Table S5 ) , indicating both successful isolation of small RNA libraries and broad equivalence of the pooled sample sets . DWV- and VDV-1-specific siRNAs of both polarities were present in all treatment groups . DWV- and VDV-1-specific siRNAs could originate from either DWV or VDV-1 , or from the previously reported [15] recombinants between these parental viruses ( Figure 3 ) . Approximately 50% of all viral reads were 22 nt in length and 25% were 21 nt , with three to four times the number of sense orientation reads to antisense , irrespective of the read length ( Table S5 ) . To exclude variation due to the efficiency of library preparation , we normalised the siRNA number to the total number of honeybee miRNA reads in a library . The normalised loads of DWV/VDV-1-specific siRNA reads were similarly low in group C and the two NV group libraries ( 0 . 341 , 0 . 377 and 0 . 397 siRNA per 1000 miRNAs respectively ) , ∼5 times higher in the two VL group libraries ( 1 . 926 and 2 . 066 siRNAs per 1000 miRNAs ) which exhibited similar viral loads to groups C and NV ( Figure 4A , see below ) , but markedly higher in the VH group samples ( 285 and 287 siRNA per 1000 miRNAs; Table S5 ) . The profiles of the DWV-and VDV-1-specific siRNA coverage of the DWV and VDV-1 reference genomes ( Figure 3 ) were most similar between groups VL and VH ( Pearson correlation 0 . 955 to 0 . 963 for DWV , 0 . 945 to 0 . 962 for VDV-1 , Table S6 ) . The profiles for groups C and NV were more distinct from each other , and to VH or VL ( Pearson correlation 0 . 593 to 0 . 786 for DWV , 0 . 399 to 0 . 726 for VDV-1; Table S6 ) . We and others have previously reported changes in virus diversity at the population level [20] and the predominance of particular virus recombinants ( i . e . a reduction in diversity ) in honeybee pupae exhibiting high viral loads [15] . To quantify both viral load and diversity in individual honeybee pupae and their associated Varroa mites we used generic NS qRT-PCR primers or primer pairs specific for DWV or VDV-1 CP or NS coding regions ( Table S1 ) . We quantified the total virus levels ( Figure 4A ) and the levels of the DWV-type and VDV-1-type CP and NS regions ( Figure 4B ) in each of the 32 pupae as well as in each of 15 Varroa mite samples co-isolated with the VH and VL group pupae . As already indicated ( Figure 1 ) , the qRT-PCR Ct values used to separate the VH from the VL , NV and C experimental groups , indicated significant differences in viral loads in representative pupae ( Figure 4A ) , with the VH group exhibiting at least 3 log10 higher levels of DWV-like viruses per pupa . When analysed using specific CP or NS primer pairs , the most pronounced difference was the increase in the number of genomes with the VDV-1 CP and the DWV NS sequences in the VH group pupae compared to the other treatment groups ( Figure 4B ) . In comparison with the control group C , the VH group exhibited a 6 , 000-fold increase in the VDV-1 CP region and a 26 , 000-fold increase in the DWV NS coding region . When compared with the NV group , the VH group showed lower relative increases ( 312-fold for VDV-1 CP and 2500-fold for DWV NS , P<0 . 0001 in both cases ) because significant amplification of viruses bearing VDV-1 CP- ( by 27-fold [P = 0 . 0217] ) and DWV NS-regions ( 10-fold [P = 0 . 0314] ) also occurred in the NV group relative to the control group C ( Figure 4B ) . The dramatic rise of the recombinant genome ( s ) containing VDV-1 CP and DWV NS in the VH group was also accompanied by a statistically significant 30-fold decrease ( P = 0 . 0151 ) of the DWV-type CP and 26-fold decrease ( P = 0 . 0477 ) of the VDV-1 NS compared with the NV group . The levels of VDV-1 CP and DWV NS coding regions showed strong positive correlation ( r = 0 . 9691 ) suggesting that this particular recombinant was preferentially acquired or amplified in the VH group pupae . The Varroa-exposed VL group also potentially acquired DWV from the mite as well as during larval feeding . It was therefore interesting to note that , when compared to the NV group , there was a statistically significant 54-fold decrease of DWV CP- ( P = 0 . 0057 ) and a 36-fold decrease of VDV-1 NS-regions ( P = 0 . 0151 ) in the VL group . Since both the VL and VH group pupae were mite-exposed but contained distinct levels and populations of DWV-like viruses we also characterised the viruses , and evidence of their replication , in the associated mites to determine if there was a correlation between high levels of virus in the honeybee and replication in the mite , as previously reported [38] . DWV- or VDV-1 specific qRT-PCR analysis demonstrated only a weak correlation with virus levels in the corresponding honeybee pupae ( Figure S5 ) . The highest correlations were found for the total load of DWV-like viruses determined using universal NS primers across the VL and VH groups ( r = 0 . 567 ) . Notably , we found that correlation between the levels of VDV-1 CP- and DWV NS-regions ( sequences present in the predominant virus population in Varroa-infested VH group pupae; Figure S5 ) in the Varroa mites and the bee pupae were lower , r = 0 . 403 and r = 0 . 465 , respectively . We went on to investigate whether we could distinguish between the mite-associated viruses in the VL and VH groups on the basis of their ability to replicate ( as determined by negative strand synthesis ) in the ectoparasite . Negative strand RNA was generally low but detectable in 10/15 mites analysed , with no significant difference between the DWV or VDV-1 CP levels ( Figure S6 ) . Together , these observations suggest that the low levels of DWV-like viruses in the VL group pupae cannot be explained by corresponding low levels of the virus in the mite and , similarly , that higher levels of the recombinant virus genomes in the mite-exposed honeybees ( VH group ) could not be attributed to either the preferential replication or absolute levels of these viruses in the associated mites . The dominance of recombinant viruses bearing VDV-1 CP and DWV NS coding regions in the VH group was strongly suggested by qRT-PCR ( Figure 4B ) . Since recent studies have demonstrated that mite infestation is associated with a marked reduction in virus diversity at the regional scale [20] , we extended our analysis to determine DWV-like virus diversity in individual pupae of the four exposure groups and , where appropriate , the co-isolated mites . In parallel , we also sampled random purple-eye stage pupae from the Varroa-infested colony to determine the pre-existing virus population at frame transfer . Nested PCR using generic ( outer ) and four specific ( inner ) primer pairs ( Table S1 ) – for each possible combination of CP and NS region – was used to amplify a 1 . 3 kb fragment spanning a central region of the virus genome ( corresponding to nucleotides 4926–6255 of the DWV genome; GenBank accession No . AJ489744 ) containing both CP and NS coding regions . We noted that no recombinants bearing a DWV CP region and VDV-1 NS region were detected in any of the experimental groups . For each of the eight pupae from the four exposure groups ( C , NV , VL , VH ) , and pupae-associated individual mites from the VL and VH groups , PCR fragments were cloned and 8–18 individual clones sequenced . In total , 93 individual sequences were obtained of the 1330 nt . region and aligned with 12 DWV-like sequences ( DWV , VDV-1 , KV and recombinants thereof; see Materials and Methods ) to generate a robust phylogenetic tree ( Figure 5 ) due to the 22 . 71% sequence divergence in the region analysed . The resulting dendrogram contained six distinct clusters , one each for non-recombinant DWV- or VDV-1-like sequences , together with four different VDV-1/DWV recombinant forms ( designated RF1–RF4; Figure 5 ) . Individual sequences obtained from pupae in exposure groups C , NV , VL and the Varroa-infested colony were present in all the major clusters indicating that these contain a significant diversity of viruses . In striking contrast , viral sequences from the VH experimental group exhibited almost no sequence divergence ( 0 . 15% at the nucleotide level ) , and consequently all clustered within a single clade ( designated VDV-1/DWV RF4 in Figure 5 ) . Therefore , the reduction in viral diversity ( as previously determined by high resolution melting analysis ) associated with the introduction of Varroa observed at the scale of tens of colonies exposed to the mite over several years [20] is reflected at the level of individual honeybee pupae following exposure to Varroa for 6 days . One interpretation of the near-clonality of viral sequences in the VH group was that these were the only ones carried , and hence transmitted , by the mite . However , with the exception of the non-recombinant DWV cluster , which was not detected in the mite , the 32 viral sequences obtained from Varroa were widely distributed within the dendrogram ( open symbols in Figure 5 ) . These results imply that , with the possible exception of non-recombinant DWV , Varroa is capable of acquiring and maintaining a diversity of DWV-like viruses , but that – either during or following transmission to naïve pupae – only a subset of these ( RF4 in Figure 5 ) are amplified to the very high levels observed in the VH group . Since the obvious difference between the horizontal transmission of DWV per os ( larval feeding ) and by Varroa is that the latter involves direct inoculation of virus to the haemolymph in pupae we investigated the recapitulation of this process by direct injection of pupae in vitro . We directly injected white eye pupae ( day 12–13 of development ) maintained in vitro ( as described in [39] ) with virus particles purified from groups C , NV and VH pupae as described previously [15] . As before , we determined the proportion of the DWV- and VDV-1-type CP coding regions in the inocula and injected pupae ( following incubation to the purple-eye stage for 3 days ) by qRT-PCR using strain-specific primers to the CP and universal primers to the NS region . Virus preparations from groups NV and VH contained higher and broadly similar levels of VDV-1-like CP coding regions . The amount of DWV-like CP coding regions was much higher in the virus preparation from the group C pupae ( where it accounted for ∼12% of the population ) than from either the NV or VH group pupae ( Figure 6A ) . Pupae inoculated with buffer alone exhibited no significant increased accumulation of DWV-like viruses when compared with untreated pupae ( Figure 6A ) . In striking contrast , irrespective of the source of viral inocula , pupae directly injected with virus preparations exhibited high virus levels characterised by markedly amplified VDV-1-like CP coding regions when compared to DWV-like CP sequences ( Figure 6A ) . Directly injected pupae were therefore similar , in both DWV-like virus levels and identity , to those previously observed in the VH experimental group ( Figure 4A and Figure 1 ) . We additionally conducted next generation sequencing ( Illumina paired-end reads ) to comprehensively characterise the group C inocula and the viruses present in pupae injected with the group C virus . The composition of the inoculum , as determined by qRT-PCR and subsequent MosaicSolver [40] analysis of the NGS reads , were in close agreement and consisted of 12 . 5% DWV ( in excellent agreement with the qRT-PCR-determined level , see above ) , 42% VDV-1 with the remainder being VDV-1 CP-encoding recombinants with a DWV-like NS region ( Figure 6B ) . Three days after injection , the pupae inoculated with group C virus exhibited a marked reduction in the DWV content ( from >12% to 1% ) and a concomitant increase in recombinant forms of the virus ( 70% of the total ) that were characterised by the presence of VDV-1 CP coding region and DWV-like NS regions ( Figure 6B ) . These results further support our previous conclusion that DWV-like viruses bearing VDV-1 CP coding regions , and particularly recombinant forms with DWV-derived NS coding regions [15] , have a selective advantage in Varroa-infested honeybee colonies , and additionally indicate that this advantage is manifest after transmission of the virus by direct inoculation and is not dependent upon Varroa per se . The sequence analysis of DWV in Varroa-exposed pupae ( Figure 5 ) in the frame-transfer study supported the presence of a single , near-clonal , recombinant form of the virus in VH group honeybees . To formally exclude a role for PCR-biased amplification in this result and to extend our analysis to investigate virus diversity in independent samples ( geographically and temporally ) , including asymptomatic and symptomatic newly emerged workers , we investigated virus diversity using next generation sequencing ( NGS ) . We sampled individual adult nurse worker bees , both asymptomatic and exhibiting the obvious wing deformities and abdominal stunting characteristic of DWV disease , from a naturally Varroa infested colony . We additionally investigated virus diversity in purple-eye stage pupae to which we had injected ( at the white-eyed stage 3 days previously ) virus purified from pupae from the same colony a month earlier . Analysis was conducted on individual pupae using a high-throughput RNA-seq approach [41] with an mRNA protocol which allowed unbiased detection and quantification of all poly ( A ) containing RNA , this would include both host mRNA and the polyadenylated DWV-like genomic RNA [12] . The NGS reads were aligned to reference DWV and VDV-1 sequences ( GenBank Accession numbers GU109335 and AY251269 respectively ) , and the pileup profiles were analysed . The proportions of DWV and VDV-1 reads in the libraries ( each containing about 10 million reads ) showed a bimodal distribution and were either very high ( from 7 . 41% to 83 . 87% , Figure 7A horizontal axis ) for injected pupae and symptomatic nurse bees , or about a thousand fold lower ( 0 . 04% to 0 . 11% ) for Varroa-naïve control pupae , for pupae inoculated with buffer alone and for asymptomatic nurse bees from the Varroa-infested colony ( Figure 7A , Table S7 ) . The remaining reads were of honeybee mRNAs . Distribution of the reads with similarity to DWV and VDV-1 suggested that all samples , irrespective of viral load , contained recombinant viruses with the CP derived form VDV-1 and NS region derived from DWV , as described above and in previous studies [15] , [16] . To assess virus diversity , we calculated Shannon's diversity index [42] for the aligned NGS reads from each experimental pupa or adult bee . Despite the ubiquitous presence of recombinant DWV-like genomes ( all consisting of a VDV-1 capsid and DWV non-structural coding regions ) there was a striking reduction of virus diversity in the bees and pupae exhibiting high virus loads ( Figure 7A , B , C ) . Average Shannon's diversity index for the NS and CP regions of the viral genomic RNA were significantly higher in the samples tested with low virus levels compared to those with high virus levels ( 0 . 1% level Fishers LSD test ) . At the same time , we observed no significant differences in Shannon's diversity index for NS and CP regions at the 5% level ( Fisher LSD test ) within the low virus group which consisted of Varroa-naïve control pupae , pupae injected with buffer alone , and asymptomatic nurse bees from a Varroa-infested colony ( Figure S7 A , B ) . For the samples tested with high virus levels ( pupae injected with virus in vitro and symptomatic nurse bees from a Varroa-infested colony ) , no differences were observed at the 5% level ( Fisher LSD test ) for the NS region . Indeed , combined low virus level and high virus level groups showed significant differences in Shannon's diversity index values for the CP and NS regions even at the 0 . 1% level ( Figure 7B , C ) . For comparison we determined Shannon's diversity index for a sample prepared by in vitro transcription of two full-length DWV cDNA clones , GenBank accession numbers HM067437 and HM067438 [15] , mixed , post transcription , at a known ratio and used as a template for NGS . We additionally used this control sample to determine the component of the observed diversity that was attributable to NGS sequencing errors which we quantified at about 0 . 5% , similar to that previously reported [43] . We calculated the threshold Shannon's diversity , a 95% confidence limit for clonal input RNA library ( shown as dashed line in Figure 7A ) using the approach described in Text S1 and ( Wood et al . , unpublished data ) . Remarkably , while the diversity index of all samples with low DWV levels ( control and buffer-injected pupae , and asymptomatic nurse honeybees ) were well above this threshold , diversity values of samples with high DWV levels ( virus-injected pupae and symptomatic nurse honeybees ) were either very close or below this clonality threshold . Similar results were obtained when diversity was estimated using multiple sampling as described in Material and Methods ( Figure S7 ) . In this case the clonality threshold value ( the range shown with the dotted lines in Figure S7 ) was also almost indistinguishable from the diversity present in symptomatic nurse bees from the Varroa-infested colony indicating that the diversity in these honeybee samples was close to the limit of detection using NGS analysis . This reinforces the near-clonal nature of the virus population in Varroa-exposed symptomatic nurse bees and is in good agreement with the sequence analysis of VH group pupae following PCR amplification of the central region of the virus genome ( Figure 5 ) . To further explore the near-clonal nature of the virus population in symptomatic nurse bees from a Varroa-infested colony we used a pair of flanking primers to the DWV-like genomic RNA to amplify and clone full-length viral cDNAs from these samples ( GenBank accession number KJ437447 ) . The central 1330 nt . region of this clone was identical to that previously characterised from VH group pupae ( Figure 5 ) despite being sampled from a separate colony in a different apiary over two years later . The consensus viral sequences , which were assembled from the NGS libraries from symptomatic honeybees with high DWV levels , showed highest overall identity with the full-length clone KJ437447 ( specifically 99 . 15% [SD = 0 . 31%] nucleotide and 99 . 78% [SD = 0 . 09%] amino acid identity ) and the 1330 nt . sequences from VH group pupae , e . g . JX661656 ( 98 . 84% [SD = 0 . 57%] nucleotide and 100 . 00% amino acid identity; Table S7 ) . In respect to the samples with low DWV levels , we found that Shannon's diversity index for the NV group sequences ( Varroa-free orally infected pupae; Figure 5 ) was 0 . 04172 . This value was very close to the Shannon's index values in the same genomic region for the pupae exhibiting low virus levels from the Varroa-infested hive , 0 . 03623 , SD = 0 . 00026 , for control ( i . e . not injected ) pupae ( 0 . 03929 , SD = 0 . 00097 ) and for pupae injected with buffer alone ( 0 . 03929 , SD = 0 . 00097 ) ; Figure S7 C ) .
Previous analyses of transcriptome or immune response changes in Varroa-exposed honeybees have produced contradictory results , perhaps due to the analysis of pooled individuals and/or pupae of different ages . These confounding influences may have obscured the marked changes in gene expression that we observed in response to either mite or viral pathogens , as emphasised by the transcriptome differences observed in the Varroa-associated pupae in groups VL and VH , which respectively exhibited 493 ( ∼5% of transcriptome ) and 951 ( ∼9% ) significantly differentially expressed ( DE ) genes when compared with the control group C ( Figure 2A ) . By stratifying Varroa-exposed pupae by viral load we can provisionally define transcriptome changes resulting from mite-associated activities such as wounding , feeding and exposure to salivary peptides ( the 444 genes shared by VL and VH groups ) and those triggered by the high viral load ( >3 log10 higher in VH than VL; Figure 2A ) . We acknowledge that the C to VL contrast may include genes involved in suppressing high levels of mite-transmitted DWV accumulation , an interpretation that warrants further study . The NV , VL and VH pupal groups also acquired DWV during larval feeding in the Varroa-infested colony , which on account of the preferential amplification of particular recombinant forms of DWV ( Figure 4B , discussed further below and [15] ) contains a distinct virus population , the composition of which differs from historically Varroa-free control colonies . Transcriptome comparison between the C and NV groups showed significant changes in a large number of genes ( 416 , ∼4% of transcriptome ) , many of which were also altered in the C to VL ( 220 ) and C to VH ( 385 ) contrasts ( Figure 2A , B ) . These may reflect per os exposure , and the resulting responses to the particular virus population circulating in the Varroa-infested hive , which resulted in changes of DWV strain composition in NV compared to C ( Figure 4B ) , together with changes resulting from environmental differences ( such as circulating pheromone ) between the originating and test colonies which would be common to all three exposure groups . The set of 59 genes DE in the contrast NV to VL ( exposure to Varroa feeding at the pupal stage which did not result in elevated viral loads ) , was largely different from the DE genes in contrast C to NV with only one gene shared . At the same time , the NV to VL set showed high commonality with the DE genes in the contrasts C to VH , C to VL , and NV to VH ( 34 , 51 , and 27 genes respectively; Figure 2A , B ) . Observed DE gene commonality in the contrasts was consistent with an orthogonal expression pattern ( Figure 2B ) following treatments ( Figure 1 ) , with different sets of genes DE in response to per os infection ( C to NV ) , exposure to mite feeding ( NV to VL ) , and high DWV ( VL to VH ) ( Figure 2B ) . Gene Ontology ( GO ) analysis gave additional insights into the transcriptional responses in honeybees following experimental treatments ( Figure 1 ) . It has been demonstrated that genes associated with the same GO terms are likely to have the same transcription factors binding to their promoter regions , which may result in co-regulated expression of these gene sets [47] . Therefore , statistically significant overrepresentation of GO BP terms associated with the DE genes may suggest coordinated and distinct honeybee responses to per os exposure ( C to NV contrast ) and to Varroa and/or Varroa-injected virus ( NV to VL contrast ) ( Figure 4A , Figure S1 ) . Such coordinated transcriptional responses may include genes involved in suppression of virus replication in the NV and VL group pupae ( Figure S1 ) , which will require further analysis . In contrast to this situation , the transcriptional changes specifically associated with the increased virus levels in group VH ( VL to VH contrast ) had no significantly over- or under-represented GO BP terms ( Figure 2A , Table S4 ) . This suggests that honeybees did not respond in a coordinated manner to the increased virus load , and that presumably unrestricted DWV replication caused dysregulation of transcription and/or mRNA stability in the honeybee similar to that previously reported in picornavirus infection of mammalian cells [48]–[50] . We analysed changes in expression of known and presumed immune-related genes ( Table 1 , Figure S4 ) defined previously [34] , [35] and by gene ontology ( GO ) terms associated with Drosophila homologs [33] . In particular , a number of proposed components of the Toll signalling pathway were affected , although the lack of activation of the antimicrobial peptide genes suggested that no activation of the Toll and Imd pathways had occurred [34] , [35] , [51] . In contrast to both the Varroa-exposed groups ( VL and VH ) the NV group was the only group in which there were more up- than down-regulated immune-related genes when compared with the control ( Table 1 , Figure S4 ) . The majority of the changes seen in the C to NV contrast were also seen in the groups that acquired DWV both orally and via Varroa ( C to VH , C to VL contrasts ) , implying that Varroa exposure may exert a dominant immunosuppressive influence over any up-regulation observed following oral exposure alone . Significantly enhanced expression of the honeybee orthologue of Vago ( GB10896; Table 1 ) , a secreted protein upregulated in Drosophila and Aedes following detection of viral dsRNA by Dicer during virus infection [52] , [53] , was observed in all groups exposed to oral DWV in the Varroa-infested colony ( NV , VL and VH ) when compared with group C . Varroa exposure ( VL or VH groups ) resulted in down-regulation of several putative components of the honeybee Toll signalling pathway [51] , including two Toll receptor orthologs ( GB10640 , GB17781 ) , CLIP-domain protease spirit ( GB14044 ) and the Toll receptor ligand spatzle ( GB15688 ) . In addition , spatzle was down-regulated when the VH group was compared against the other experimental groups , suggesting down-regulation of this gene may be a response to the elevated virus levels in the group VH , rather than the presence of Varroa per se . Toll signaling pathways are implicated in antiviral resistance to the RNA virus Drosophila virus X [54] , possibly controlling proliferation of haemocytes which , because of their involvement in phagocytosis , play a central role in insect immunity [54] , [55] . We also observed down-regulation of a Tetraspanin 68C ( Tsp68C ) ortholog ( GB16002 , GB13670 ) , a cell surface membrane scaffolding protein previously implicated in receptor modulation during hemopoiesis [56] , an ortholog of pannier ( GB19895 ) , a GATA transcription factor required for differentiation of plasmatocytes ( which resemble the mammalian macrophage lineage [57] ) , and a serrate ortholog ( GB15155 ) , a membrane ligand for the Notch receptor implicated in differentiation of haemocyte-related crystal cell precursors which function in pathogen defence via melanisation [58] . These transcriptome changes may help explain functional studies in which salivary secretions from Varroa mites damage moth caterpillar haemocytes [59] and suggest that Varroa-mediated depletion of haemocytes , a key component of the immune response of insects [60]–[62] , may contribute to enhanced susceptibility to DWV and other viruses . Interestingly , we also observed suppression of the Friends-of-GATA transcription factor U-shaped ( ush ) ortholog ( GB16457 ) , in the C to VL and NV to VL contrasts . Drosophila ush is reported to antagonise crystal cell development [63] , [64] , implying that the low level of virus accumulation in the VL group may be due to elevated numbers of crystal cells resulting from ush down-regulation . Although by definition descriptive , transcriptome analysis of pupae stratified according to Varroa and virus-exposure , also provides direct insights into possible pathogenic mechanisms . In the contrast C to VH we observed differential expression of orthologs of five Drosophila homeobox genes ( summarised in Table S8 ) encoding transcription factors which are involved in insect development [65] . Most of these DE genes are reported to be expressed at early pupal stages and involved in abdomen ( Abdominal B ) , appendage ( apterous ) or brain development ( extradenticle ) . This may explain previously reported developmental abnormalities in the honeybee that are associated with high DWV levels at the pupal stage [13] and warrant further investigation to potentially determine the molecular mechanism underlying DWV pathogenesis . Notwithstanding the absence of significant changes in gene expression of key components , such as Dicer and Argonaute , of the RNAi response – the major antiviral mechanism in insects [25] – we explored the relationship between DWV-like virus levels and the corresponding siRNA populations . In particular , we sought to investigate if high levels of DWV in VH honeybees was associated with the limited accumulation of virus-derived siRNA , implying the virus may express an siRNA suppressor as , for example , demonstrated in Alphavirus infection of mosquitos [66] . Although DWV- and VDV-1 specific siRNAs were recently detected in adult honeybees [67] , [68] , these studies could not show if RNAi is involved in suppression of the virus , because viral genomic RNA levels were not quantified . Analysis of siRNAs in the honeybees of the frame transfer experiment showed that the predominant DWV- and VDV-1-specific siRNAs were 22 nt in length with genome sense strand-specific siRNAs present at a 3–4 fold excess over antisense . This was consistent both with the presence of replicating DWV-like viruses in all experimental groups and with the known activity of Dicer in other insects [25] and strongly suggests normal functioning of Dicer in honeybees [69] . As insects do not amplify siRNA populations [25] , it was unsurprising that virus-specific siRNA levels were broadly proportional to the level of viral genomic RNA determined by qRT-PCR; the C and NV groups exhibited ∼103 times less viral genomic RNA than the VH group and normalised siRNA levels differed by ∼770 times ( Table S5 ) . The exception to this was the siRNA response in the VL group which was ∼5 times higher proportionally than the level of VL virus genomic RNA ( Table S5 ) . The relationship between the levels and compositions of the viral genomic RNA and virus-derived siRNA may be altered by differences in targeting of the individual components of DWV-like virus population by the honeybee RNAi machinery , as observed during West Nile virus infection of mosquitos [70] . Although the presence of virus-specific siRNAs does not necessarily correlate with effective silencing – viruses may encode late-acting suppressors such as the Argonaute-inhibiting 1A protein of cricket paralysis virus [71] – the robust siRNA response in the VL group may contribute to suppression of DWV replication and the differences between this response and that observed in the VH group may be a fruitful area for further analysis . The introduction of the parasitic Varroa mite elevates the level of DWV-like viruses [20] , amplifies particular strains that that are best defined as recombinant forms ( RF ) bearing the capsid determinants of VDV-1 and non-structural genome region from DWV [15] , [16] and dramatically reduces the diversity of DWV-like viruses in a population [20] . Using complementary approaches including strain-specific qRT-PCR and sequencing together with next generation sequencing of the virus genome and host siRNA response to infection , we analysed individual pupae exposed to DWV during larval feeding and following mite exposure , and recapitulated horizontal transmission of virus by Varroa using direct injection . The C , NV and VL exposure groups all carried low viral loads and exhibited high virus diversity ( Figure 4B , Figure 5 ) . However , the virus populations carried were distinct , with the NV and VL experimental groups containing a diverse range of recombinant forms of DWV-like viruses bearing the capsid coding region of VDV-1 and the non-structural coding regions of DWV . In contrast , the VH group exhibited very high levels of a specific near-clonal ( 0 . 15% divergence in the regions sequenced ) recombinant form of DWV ( labelled RF4 in Figure 5 ) . Due to the subsequent identification of the same near-clonal virulent virus in temporally and spatially distinct samples ( see below ) we henceforth designate this virus DWVV to discriminate it from other circulating recombinants forms . This suggests that the changes reported in virus levels and diversity at a regional scale [20] reflect events occurring within a few days ( uncapped to the purple-eye stage ) in individual mite-exposed pupae . Nearly identical , clustering tightly within the DWVV clade , were also detected in pupae from the C , NV and VL groups ( Figure 5 ) . Since these groups have significantly lower viral loads it implies that the high viral loads seen in the VH group cannot be solely attributed to their infection with a particular recombinant form of the virus . We reasoned that there were two possibilities that might account for the marked amplification of DWVV in the VH group pupae . Firstly , the mite may have delivered a high dose of one specific recombinant form , perhaps reflecting its preferential replication in the ectoparasite . Secondly , we considered that DWVV might have a growth advantage when inoculated into haemolymph by Varroa ( potentially in addition to the preferential amplification in the mite ) . To distinguish between these possibilities we sequenced qRT-PCR amplified viral RNA from mites co-isolated from capped cells containing group VL and VH pupae . We also investigated the consequences of Varroa-independent mechanical virus transmission by direct injection of mixed virus preparations to Varroa-naïve pupae and subsequent monitoring of virus levels and diversity . Individual Varroa mites contained a diversity of DWV-like sequences that were well distributed throughout the phylogenetic tree of virus sequences from pupae ( square symbols , Figure 5 ) . Using VDV-1- and/or DWV-specific primer pairs spanning the central 1 . 3 kb region of the virus genome mites were detected containing VDV-1 and all four distinct RFs identified in the four experimental groups in the frame transfer study . Only non-recombinant DWV was absent from the 32 mite-associated viruses sequenced . We also detected negative strand sequences of both DWV and VDV-1 CP regions in the majority of the 15 mites tested ( Figure S6 ) , implying that virus replication does occur in the mite . Although we did not detect DWV CP among the central 1 . 3 kb region sequences amplified from the mites ( Figure 5 ) , this could be a consequence of limited experimental sampling and the higher levels of VDV-1 CP in the population , a conclusion supported by analysis of the negative strands present ( Figure S6 ) . Although further studies will be required to determine whether sampling stochasticity accounts for the apparent absence of non-recombinant DWV in Varroa , together these results suggest that – at least at the level of the entire mite – there is no selection , either by absolute presence or replication capability , for the DWVV RF that accumulates to high levels after mite exposure in VH group pupae . Since the diversity of virus present in Varroa indicates that the near-clonal virus population in the VH group is not due to the mite delivering either a restricted virus type or to elevated levels of DWVV in the mite we went on to inoculate pupae with a mixed virus population prepared from group C pupae and characterised the resulting virus population after three days . Recipient purple-eye stage pupae contained high virus loads which , compared with the inocula , had markedly reduced levels of DWV-like virus and elevated levels of a VDV-1/DWV recombinant ( Figure 6 ) . Although , the resulting virus diversity was not as limited as seen in the naturally infected VH group , we attribye this to the restricted incubation time between inoculation and sampling ( 3 days vs . 6 days ) , in part imposed by experimental limitations of working with late-stage larvae and early-stage pupae which are vulnerable to handling damage . Despite these limitations , these results clearly demonstrate that direct inoculation of a mixed virus preparation , recapitulating virus inoculation by the mite , results in a marked reduction in virus diversity . We additionally demonstrated , by RNA-seq analysis of temporally and geographically independent symptomatic nurse bees and similarly independent pupae directly injected with virus preparations , that essentially the same near-clonal virus ( DWVV ) was present as previously identified in the VH group pupae . In parallel , control asymptomatic nurse bees or mock-injected pupae exhibited high diversity and low levels of virus ( Figure 7 , Figure S7 , Table S7 ) , as previously seen in the C and NV groups during the frame transfer study . The unselective RNA-seq methodology excludes the possibility that virus clonality at high virus loads was a consequence of PCR bias . The remarkable restriction in virus diversity in both injected pupae or symptomatic nurse bees exhibiting high viral loads was in good agreement with that seen in group VH pupae ( Figure 5 ) determined following qRT-PCR amplification ( 0 . 15% diversity ) . We propose that the strikingly elevated levels and associated restricted diversity of DWVV ( RF4-type; Figure 5 ) in both the Varroa-exposed VH group pupae and characteristically DWV symptomatic nurse bees is because this virus has a preferential advantage when delivered directly to haemolymph of developing pupae . There remains the possibility that DWVV alone replicates to elevated levels in the salivary glands of Varroa and is the only DWV-like virus transmitted during feeding . However , we do not favour this hypothesis as we would expect it to result in DWVV being the predominant virus detected when whole-mite RNA samples were analysed . Furthermore , we also present evidence ( Figure 7A , Table S7 ) that DWVV predominates when a mixed virus population is directly inoculated . Nevertheless , it remains an intriguing avenue for further study . Assuming this is not the explanation , the molecular mechanisms underpinning the advantage of the near-clonal DWVV , be it evasion of the host antiviral response , specific tissue tropism or some other aspect of the virus-host interaction , will require further studies . This will necessitate immunohistological analysis of orally infected or injected pupae , the development of a reverse genetic system to identify determinants of DWV tropism , and the analysis of the contribution of immune-related ( and other ) host genes using RNAi-based strategies [72] , [73] . Without proper management Varroa has a devastating effect on honeybee colony viability and consequent honey production and pollination services . We show here that the markedly elevated levels of DWV-like viruses in Varroa-exposed honeybee pupae are likely attributable to the direct inoculation of a specific virus , DWVV , by Varroa to haemolymph . Repeated cycles of Varroa-replication within an infested colony would preferentially amplify DWVV , potentially resulting in it becoming the predominant virus present , transferred both by Varroa and per os . Further studies will be required to determine whether such a virus , if sufficient were ingested , would also cause symptomatic infection . Oral susceptibility to a virulent form of DWV may also explain reported cases of deformed wing disease symptoms seen in Varroa-free colonies in Hawaii [20] and Scotland ( Andrew Abrahams , pers . comm . ) , but may also reflect genetic variation and the presence of particularly susceptible pupae in the colony . Our study demonstrates that a proportion of Varroa-exposed pupae ( the VL group ) do not exhibit elevated levels of the near-clonal DWVV recombinant ( Figure 5 ) . Further in vitro studies will be required to determine whether these are naturally resistant – and therefore form the basis for genome wide association studies of the genetic determinants of virus resistance – or if they reflect the stochastic nature of the transmission event from the mite .
This study was based around an experiment in which a brood frame containing newly hatched larvae from a Varroa-free colony was introduced into a Varroa-infested colony . The larvae were left to develop within the Varroa-infested colony , and pupae were collected 11 days later from capped brood cells at the purple eye stage and analysed using a range of molecular methods . The Varroa-free honeybee ( Apis mellifera ) colony with a naturally-mated one-year-old queen was imported from Colonsay , Scotland , an island with no historic reports of Varroa incidence and no imports of honeybees from Varroa-infested areas . This allowed us to exclude the presence of DWV strains associated with Varroa mite infestation . As a source of Varroa mites and the mite-associated DWV strains , we selected a Warwickshire honeybee colony , heavily infested with Varroa and having high DWV levels in honeybees and mites . The Varroa-free and Varroa-infested colonies were contained in separate mesh flight cages ( dimensions: 6 meters long , 2 . 5 meters wide , 2 meters high ) and maintained on an artificial diet of sugar syrup and pollen . The pollen was imported from Varroa-free Australia to exclude possible contamination with Varroa-associated viruses through foraged food and was pre-screened by PCR before use for DWV-like viruses . In order to minimise possible effects on honeybee gene expression due to the differences in nutrition , both the control Varroa-free and the Varroa-infested colonies were maintained in flight cages in the same apiary ( at the University of Warwick , UK ) and were fed on the artificial diet for two months before the start of the frame transfer experiments . Neither colony was treated with miticides . The experimental infestation , summarised in Figure 1 , was conducted on 4th–15th August 2011 . As stated previously , it involved the transfer of a brood frame , which contained newly hatched honeybee worker larvae ( on day 4 of development ) , from the Varroa-free to the Varroa-infested colony . As a result , the transferred larvae were exposed to Varroa-selected DWV-like viruses in brood food delivered by the nurse honeybees of the Varroa-infested colony for five days before brood cell capping on day nine of development ( Figure 1 , Treatment 1 , Oral DWV infection ) . Honeybee larvae were left to develop in the capped cells for six days and then were sampled as pupae on day 15 , when they had reached the purple-eye stage of development [74] . A proportion of these brood cells were naturally infested with Varroa and hence contained pupae that were subjected to mite feeding ( Figure 1 , Treatment 2 , Mite feeding ) . We sampled Varroa-infested pupae and the mites associated with individual pupae , with mite feeding confirmed by the presence of the mother mite and at least one protonymph [27] . Control pupae at the same developmental stage were sampled from the Varroa-free hive at the same time . A colony from a separate apiary in Warwickshire that exhibited Varroa mite infestation for over a year was sampled in August 2013 to assess the virus populations in colonies with established Varroa infestation . The pupae and the Varroa mites associated with each infested pupa were individually snap frozen in liquid nitrogen immediately after being removed from brood cells and stored at −80°C prior to total RNA extraction . For total RNA extraction , whole individual honeybee pupae were ground to fine powder in liquid nitrogen , and half of the powder used for RNA extraction , carried out using 1 mL of Trizol Reagent ( Invitrogen ) according to the manufacturer's instructions . Total RNA extraction from Varroa mites was carried out using RNeasy spin columns ( Qiagen RNeasy Plant Mini kit ) . Virus purification from honeybee material and extraction of the viral genomic RNA from virus particles were carried out as described previously [15] . For genome-wide analysis of the honeybee transcriptome total RNA preparations from eight individual honeybees from each of the four experimental groups ( 32 honeybees in total ) were purified further using RNeasy Plant Mini kit spin columns ( Qiagen ) . RNA concentration , purity and integrity were assessed using a 2100 Bioanalyzer and an RNA 6000 LabChip ( Agilent Technologies ) . The probe preparation , hybridization and scanning were carried out according to the Agilent instructions , essentially as in [75] . Total RNA extracts from an individual honeybee were used to produce Cy3- and Cy5-labelled aRNA samples using a Low Input RNA fluorescent linear amplification kit ( Agilent Technologies ) , according to the manufacturer's instructions . The Cy3-and Cy5-labelled samples were used in a two-colour dye-balanced loop design [30] , [31] for a genome-wide analysis of the honeybee transcriptome with the custom expression oligonucleotide microarray . Four slides , each with eight two-channel arrays were employed , allowing two replicates per sample , one green and one red . Different treatment groups were allocated to the green and red channels in each array; the loop design ensured that each sample was indirectly compared with all other samples . The array , in 60K ×8 format , included 60 nt oligonucleotides specific to 10 , 157 transcripts of the Apis mellifera Official Gene Set 1 , OGS1 [32]; the array also contained probes to all the honeybee RNA viruses known to date . Each probe was replicated five times to enable robust statistical analysis . Sequences of the probes of the honeybee whole genome expression microarray , a 60-mer oligonucleotidide array based on the Apis mellifera transcriptome ( OGS1 ) and Apis mellifera fungal and viral pathogens ( Agilent ID: 027104 , SurePrint G3 Custom GE 8×60K ) , are available in the ArrayExpress database ( www . ebi . ac . uk/arrayexpress ) under accession number A-MEXP-2251 . Following hybridisation , the microarrays were scanned using Agilent Technologies G2565CA Scanner and the fluorescence intensity data were processed using feature extraction software ( Agilent Technologies ) . Cy3 and Cy5 fluorescence intensities for each spot were measured as values of green and red pixels respectively . The details of the array experiment design , sample description , and microarray data are available in the ArrayExpress database ( www . ebi . ac . uk/arrayexpress ) [76] under accession number E-MTAB-1285 . One array failed ( assigned to VL green and NV red ) leaving 62 channels for final analysis . For additional confirmation we conducted qRT-PCR analysis with primers specific to Paenibacillus larvae ssp and Melissococcus plutonius ( Table S1 ) , the causal agents of American foulbrood and European foulbrood , respectively , which showed that the samples were free of detectable levels of these bacterial pathogens . The unprocessed intensity scanning values were both within-array and between-array normalized using the linear model based Limma R package [77] . Differentially expressed ( DE ) genes in all six possible contrasts were found using Limma ( via function “lmscFit” incorporating intraspot correlation ) and also the R GaGa package for gamma-gamma Bayesian hierarchical modeling [78]–[80] . A gene was considered as differentially expressed ( DE ) in a given contrast ( using a t-statistic moderated across genes ) when the average expression exceeded 6 . 0 , the fold change exceeded 14% , the Limma analysis p-value adjusted for multiple genes was less than 0 . 05 and the posterior probability determined by GaGa was above 0 . 6 . Microarray results were validated by qRT-PCR using a set of primers for certain honeybee genes and DWV ( Table S1 ) . For Gene Ontology ( GO ) analysis a three-stage process was used . Genes in the latest A . mellifera genome annotation , Amel_4 . 0 ( http://hymenopteragenome . org/beebase ) , corresponding to genes in A . mellifera OGS1 were found using protein blast . GO terms associated with Amel_4 . 0 genes were then obtained using Blast2GO [81] with the SwissProt database option . Finally , over- and under-represented GO terms in the sets of DE honeybee OGS1 genes in each contrast were obtained with BiNGO , using a hypergeometric test , a Benjamini and Hochberg FDR correction and a significance level of 0 . 05 [82] . For Principal Component Analysis ( PCA ) , the significant DE genes in all six contrasts were pooled and ranked by their adjusted p-value . The 60 with the lowest adjusted p-value were selected , all of which appeared in the contrast C to VH; the other contrasts' contributions were 35 ( C to VL ) , 21 ( NV to VH ) , 19 ( C to NV ) , 4 ( NV to VL ) and 11 ( VL to VH ) genes . Principal components of the expression profiles across the 62 microarray channels were found and ( the first two ) plotted using the princomp and biplot functions in R [83] . For high throughput sequencing of small RNA , we pooled equal amounts of the Trizol-extracted total RNA from individual honeybees and isolated the 15 to 40 nt RNA fraction , which was separated using denaturing polyacrylamide gel . The RNA pools were ligated to the oligonucleotide adapters , reverse-transcribed and amplified using the TruSeq Small RNA Sample Prep Kit ( Illumina small RNA kit ) . The libraries were sequenced using the Illumina HighSeq 2000 platform , producing 15–25 million reads per libraries ( GATC-Biotech , Germany ) . The small RNA NGS sequencing data are available in the ArrayExpress database ( www . ebi . ac . uk/arrayexpress ) [76] under accession number E-MTAB-1671 . The reads were cropped to remove adapter sequences and aligned to reference viral and miRNA sequences using Bowtie [37] . Samtools mpileup was used to produce the siRNA and miRNA coverage profiles . Real-time reverse transcription PCR was carried out essentially as in [15] . In brief , RNA extracts were treated with DNAse , then purified DNA-free total RNA preparations were used as a template to produce cDNA using random primer and Superscript III reverse transcriptase ( Invitrogen ) . The cDNA samples produced were used for real-time PCR quantification of the DWV or host transcripts using SYBR green mix ( Agilent Technologies ) . Oligonucleotide primers are summarized in Table S1 . For strand-specific quantification of viral RNA of DWV and VDV-1 types reverse transcription was carried out at 50°C using Superscript III reverse transcriptase ( Invitrogen ) and the tagged primers designed to anneal to the negative strands RNA of DWV or VDV-1 , primers 389 and 391 respectively ( Table S1 ) . The qPCR step was carried out using corresponding DWV or VDV-1 specific primers in negative polarity ( Table S1 , Primers 1384 or 1382 ) and primer 388 identical to the sequence of the tag ( Table S1 , primer 388 ) . Amplification of the cDNA fragments corresponding to the central region of DWV genomic RNA was carried out by nested PCR using GoTaq PCR mix ( Promega ) and primers 155 and 156 ( Table S1 ) using the cDNA extracted from the honeybees and the mites , pooled according to their treatment groups . The outside PCR primers were designed to amplify all known DWV-like sequences . For each first round reaction we carried four second round amplification reactions using VDV-1- or DWV-specific primers , 151–154 ( Table S1 ) , which allowed distinction of VDV-1-type and DWV-type CP and NS regions , thereby enabling amplification of all potential combinations , even those present at very low levels . The PCR fragments were cloned into pGemT-Easy ( Promega ) and sequenced using the Sanger dideoxy method . GenBank accession numbers for the reported sequences are JX661628–JX661712 and KC249926–KC249933 . The full-length cDNA of DWV , GenBank accession number KJ437447 , was amplified by RT-PCR using primers specific to the published termini of DWV and VDV-1 RNA and cloned into the pCR-TOPO-XL vector ( Invitrogen ) as described in [15] . The sequences were aligned using CLUSTAL X [84] , and phylogenetic analysis of the sequences was carried out using the PHYLIP package [85] . For the next generation sequencing of RNA , a series of overlapping cDNA fragments were produced using viral RNA or total RNA preparations using the set of primers designed to the sequences of the genomic RNA conserved among DWV , VDV-1 and KV ( Table S1 ) . The fragments were pooled and libraries of paired-end reads ( 101 nt . ) , about 5 million per sample , were generated using an Illumina HiSeq 2000 ( GATC-Biotech ) . The virus genomic RNA NGS sequencing data are available in the ArrayExpress database ( www . ebi . ac . uk/arrayexpress ) [76] under accession number E-MTAB-1675 . The next generation sequencing of the poly ( A ) RNA fraction ( RNA-seq ) of the total RNA preparations isolated from the honeybees was carried out using Illumina HiSeq 2000 ( GATC-Biotech ) protocol , with about 10 million 101 nucleotide-long reads generated for each sample . The RNA-seq sequencing data are available in the EBI Sequence Read Archive [86] under accession number PRJEB5249 . This RNA-seq dataset was used to calculate Shannon's diversity index values of DWV populations using the following procedure . First we selected the reads aligning to the reference DWV and VDV-1 sequences ( GenBank Accession numbers GU109335 and AY251269 respectively ) from the original RNA-seq libraries using Bowtie . To take into account the effect of difference in coverage of low virus levels and high virus level RNA-seq libraries we used two approaches , ( i ) correction for NGS error for complete libraries ( Text S1 and Wood et al . , unpublished data ) and ( ii ) multiple sampling . For the latter we produced five samples of 3285 reads ( the lowest number of the viral reads among the libraries ) , which were aligned using Bowtie to the reference DWV and VDV-1 sequences , and the NGS nucleotide pileups were then generated for each nucleotide position of the reference sequences using samtools . Shannon's diversity index of the aligned nucleotides was calculated for each position in the reference sequence . Then , the average Shannon's index values were calculated for the selected regions in the reference genomes for each sample . The averages values of and standard deviation of five samples were used in the statistical analysis .
|
Honeybees are the most important managed pollinating insect , contributing billions of dollars to annual global agricultural production . Over the last century a parasitic mite , Varroa , has spread worldwide , with significant impacts on honeybee colony health as a consequence of its transmission of a cocktail of viruses while feeding on honeybee ‘blood’ . The most important virus for colony health is deformed wing virus ( DWV ) , high levels of which cause developmental deformities and premature ageing resulting in high overwintering colony losses . In experiments on individual Varroa-exposed pupae we demonstrate that a single type of virulent DWV is amplified 1 , 000–10 , 000 times in the recipient pupae , despite the mite containing a high diversity of replicating DWV strains . We could recapitulate this by direct injection of pupae with mixed virus populations , showing the virulent strain is advantaged by the route of transmission . In parallel , we detected changes in the immune response and developmental gene expression of the honeybee and propose that these contribute to the characteristic pathogenesis of DWV . Identification of a virulent strain of DWV has implications for therapeutic or prophylactic interventions to improve honeybee colony health , as well as contributing to our understanding of the biology of this important honeybee viral pathogen .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"epizootics",
"sequencing",
"techniques",
"genome",
"evolution",
"parasite",
"evolution",
"rna",
"analysis",
"nucleic",
"acid",
"sequencing",
"parasitology",
"phylogenetic",
"analysis",
"genome",
"analysis",
"molecular",
"biology",
"techniques",
"veterinary",
"science",
"sequence",
"analysis",
"veterinary",
"diseases",
"rna",
"sequence",
"analysis",
"veterinary",
"parasitology",
"parasitism",
"molecular",
"evolution",
"molecular",
"biology",
"molecular",
"biology",
"assays",
"and",
"analysis",
"techniques",
"community",
"ecology",
"veterinary",
"pathology",
"ecology",
"high",
"throughput",
"sequencing",
"veterinary",
"virology",
"trophic",
"interactions",
"nucleic",
"acid",
"analysis",
"transcriptome",
"analysis",
"biology",
"and",
"life",
"sciences",
"species",
"interactions",
"computational",
"biology",
"evolutionary",
"biology"
] |
2014
|
A Virulent Strain of Deformed Wing Virus (DWV) of Honeybees (Apis mellifera) Prevails after Varroa destructor-Mediated, or In Vitro, Transmission
|
In many human fungal pathogens , genes required for disease remain largely unannotated , limiting the impact of virulence gene discovery efforts . We tested the utility of a cross-species genetic interaction profiling approach to obtain clues to the molecular function of unannotated pathogenicity factors in the human pathogen Cryptococcus neoformans . This approach involves expression of C . neoformans genes of interest in each member of the Saccharomyces cerevisiae gene deletion library , quantification of their impact on growth , and calculation of the cross-species genetic interaction profiles . To develop functional predictions , we computed and analyzed the correlations of these profiles with existing genetic interaction profiles of S . cerevisiae deletion mutants . For C . neoformans LIV7 , which has no S . cerevisiae ortholog , this profiling approach predicted an unanticipated role in the Golgi apparatus . Validation studies in C . neoformans demonstrated that Liv7 is a functional Golgi factor where it promotes the suppression of the exposure of a specific immunostimulatory molecule , mannose , on the cell surface , thereby inhibiting phagocytosis . The genetic interaction profile of another pathogenicity gene that lacks an S . cerevisiae ortholog , LIV6 , strongly predicted a role in endosome function . This prediction was also supported by studies of the corresponding C . neoformans null mutant . Our results demonstrate the utility of quantitative cross-species genetic interaction profiling for the functional annotation of fungal pathogenicity proteins of unknown function including , surprisingly , those that are not conserved in sequence across fungi .
Diseases produced by fungal infections are challenging to diagnose and treat , making these infections a major cause of morbidity and mortality worldwide [1] , [2] . Genetics and genomics have led to the identification of numerous pathogen genes required for replication in the mammalian host [3]–[7] . Unfortunately , many , if not most , identified virulence genes lack in vitro phenotypes that could explain their effects in the host [3]–[8] , and the predicted protein sequences often offer few clues to function . Thus , our power to identify pathogen genes required for disease far outstrips our ability to understand their molecular function in the host . Historically , the expression of human genes in the model yeasts Saccharomyces cerevisiae and Schizosaccharomyces pombe has been used as a tool to identify specific genes and to determine their cellular function [9]–[14] . In a classic example , complementation of a fission yeast cdc2 mutant was used to identify human Cdk1 [11] . More recently , a number of groups have combined the expression of foreign genes with high-throughout tools available in S . cerevisiae to identify suppressor genes to obtain insights into the function of human proteins , ranging from those involved in neurodegeneration to cancer [9] , [11] , [14] . Likewise , expression of viral and bacterial proteins in yeast , coupled with subsequent genetic analysis , has proven to be informative . For example , the genes responsible for biosynthesis of the eEF2 modification diphthamide were identified in selections for resistance to the F2 fragment of diphtheria toxin [15] . Identification of S . cerevisiae gene deletion mutants hypersensitive to the expression of the Shigella virulence factor OspF , a type III secretion substrate , coupled with transcriptional profiling experiments , led to the identification of the cell wall integrity MAP kinase pathway as a likely target of OspF in yeast [10] , [13] . Importantly , the latter study took advantage of phenotypic information for yeast deletion mutants available at that time to obtain clues to gene function [10] . The construction of a library of all nonessential gene deletions for S . cerevisiae [8] together with the development of genetic selections led to the development of the synthetic genetic array ( SGA ) method for quantitatively measuring genetic interactions on a genome scale [16] , [17] . This approach has facilitated the systematic annotation of gene function in S . cerevisiae [18] , [19] . Genetic interaction , or epistasis , measures the degree to which two genes affect each other [16] , and is measured by comparing the phenotype of a double mutant to that of the two corresponding single mutants . Genes that act in the same pathway display similar patterns of genetic interactions with other genes [16]–[19] . Recently , the large-scale application of these methods led to production of a remarkable genome-scale genetic interaction map based on the analysis of ∼5 . 4 million gene pairs . Such a comprehensive genetic interaction dataset has only been described to date for the model yeast S . cerevisiae [19] . Below we test the utility a cross-species genetic interaction approach for fungal pathogen gene annotation that combines expression of pathogen genes in S . cerevisiae with genetic interaction profiling . We used genes from the human pathogen Cryptococcus neoformans , an opportunistic basidiomycete fungal pathogen that is very distantly related to the model yeasts S . cerevisiae and S . pombe . C . neoformans is the most common cause of fungal meningitis in humans , and among the most important causes of morbidity and mortality in AIDS patients , leading to ∼1 million infections and ∼600 , 000 deaths annually in sub-Saharan African alone [1] . Our laboratory previously generated a library of 1201 gene deletion strains and used a signature-tagged mutagenesis approach to identify genes required for pathogen fitness during experimental infection of mice [5] . In addition to identifying new genes required for the synthesis of known virulence factors , these studies identified several dozen genes required for virulence whose mutation failed to yield in vitro phenotypes that could explain its role in the host . As a proof-of-principle , we expressed six C . neoformans genes of interest in each member of the S . cerevisiae deletion library and quantified their impact on fitness , thereby producing cross-species genetic interaction profiles . We exploited their similarities to existing S . cerevisiae knockout genetic profiles to predict possible functions for each C . neoformans protein . For two of these C . neoformans proteins , Liv6 and Liv7 , we describe validation experiments that support the functional assignment . For Liv7 , additional experiments connect its newly identified function to the evasion of phagocytosis , an important virulence trait . The cross-species genetic interaction profiling approach described here offers a generalizable avenue toward the functional annotation of pathogenicity factors of fungal agents of infectious disease .
We sought to develop a generic approach for developing testable hypotheses for the function of novel C . neoformans virulence genes by taking advantage of the methods and datasets that exist in S . cerevisiae . We created S . cerevisiae strains that each expressed a C . neoformans gene of interest ( described further below ) . We crossed these to the S . cerevisiae gene deletion library using automated SGA methods and measured fitness of the progeny strains using high-throughput colony imaging methods [16] , [17] ( Figure 1 ) . Measurements ( n = 8 per double mutant ) were converted into significance scores ( S-scores ) [20] ( See Methods ) . We refer to these data as a “cross-species genetic interaction profile” which is the set of quantitative genetic interactions between strains expressing a particular C . neoformans gene and each S . cerevisiae deletion mutant . We calculated correlations between these cross-species profiles and the available genetic interaction profiles of deletion mutants in S . cerevisiae [21] . We reasoned that the expression of a C . neoformans gene could , in some cases , produce dominant-negative effect and produce genetic interaction profiles that correlate positively with those of S . cerevisiae gene deletions that function in the homologous pathway . Alternatively , the expression C . neoformans gene might have a dominant-positive effect , producing a profile that anti-correlates with those of S . cerevisiae deletions mutants in the same pathway . Scenarios on which both behaviors occurred could also be imagined . We further expected that the expression of some , but not all , of C . neoformans genes would produce profiles that would allow us to develop experimentally testable hypothesis for gene function . We focused on six C . neoformans genes ( Figure 2 and Table S1 ) , four of which ( LIV5 , LIV6 , LIV7 , and LIV13 ) our previously work identified as necessary for growth in a murine infection model [5] . Two others , BLP1 and MEP1 , are targets of Gat201 [22] , a master transcriptional regulator of virulence [5] , [22] . Blp1 is important for C . neoformans to evade phagocytosis by macrophages . Four of these genes ( LIV6 , LIV7 , MEP1 , and BLP1 ) lack S . cerevisiae orthologs . Several contain conserved domains identified by BLAST [23] , but the function of these domains are poorly understood ( Figure 2 and Table S1 ) . The application of PHYRE , a threading-based structure prediction algorithm , provided information for only Liv6 , which it predicts to be structurally related to a lectin [24] . We generated cross-species genetic interaction profiles using S . cerevisiae strains carrying two control constructs and six different bait constructs: pGPD ( promoter-only control ) , pGPD-GFP ( nonspecific protein control ) , pGPD-LIV5 , pGPD-LIV6 , pGPD-LIV7 , pGPD-LIV13 , pGPD-MEP1 , and pGPD-BLP1 . To ensure reasonable expression levels ( see Materials and Methods ) used the strong GPD1 promoter because the C . neoformans genome is GC-rich [25] compared to S . cerevisiae [26] , which is anticipated to inhibit protein translational efficiency due to differences in codon usage and an increased propensity to form inhibitor RNA structures [27] . We calculated Pearson correlations ( correlation score ) to compare cross-species genetic interaction profiles with the previously described genetic interaction profiles of produced by crosses of 1712×3885 S . cerevisiae gene deletions [19] . To avoid potentially spurious correlations , we filtered out correlations with S . cerevisiae deletions whose profiles yielded significant correlations with either of the two control baits . Significance testing revealed that correlations with a value of greater than 0 . 08 are highly significant ( P<0 . 001 , two-tailed test , Bonferroni-corrected for multiple hypothesis testing ) . Quantile-quantile plots of the correlations with S . cerevisiae deletions versus standard normal quantiles revealed outliers on one or both tails for all baits ( Figure S1 ) . We focused on correlations that departed from the mean by at least three standard deviations ( |Z|>3 ) . This conservative strategy yielded from 2–15 hits , depending on the bait ( Table 1 ) . The profile of LIV7 displays the largest number hits , and their identities strongly points to a role in Golgi transport , a prediction whose validation via experiments in C . neoformans is described below . The LIV6 profile correlates positively and negatively with two S . cerevisiae genetic profiles , those of deletions in SYN8 and ECM21 , respectively . Strikingly , both genes play a role in endosome transport and/or fusion [28] , [29] , predicting a role for Liv6 in these processes in C . neoformans . Support for this prediction via experiments in C . neoformans is also described in this paper . Several other profile hits were obtained , but have yet to validated . The Liv5 profile correlates with those of deletions affecting the cell cycle and autophagy [30]–[32] and the Liv13 profile negatively correlates with the genetic profiles of knockouts involved in alleviating protein folding stress [33]–[35] . The profile of the MEP1 metalloprotease correlates with that of a knockout in a S . cerevisiae metalloprotease of a different family , YBR075W [23] , as well as proteins involved in nucleotide and RNA metabolism . Finally , the genetic interaction profile of the S . cerevisiae strain expressing Blp1 positively correlates with that of the deletion of an S . cerevisiae gene , ETR1 , that has a role in fatty acid synthesis . This observation may be related to the Rare Lipoprotein A ( RlpA ) domain prediction for Blp1 ( Table S1 ) . Liv7 ( Figure 2 ) is a 330-residue protein that contains a DUF3752 domain , which is annotated as a conserved domain of unknown function [36] . The profile of the S . cerevisiae strain expressing LIV7 displays the strongest three positive correlations with the published genetic interaction profiles of S . cerevisiae gene deletions trs33Δ , tlg2Δ , and vps51Δ ( Figure 3A ) . Strikingly , all three of these genes function in transport events that involve the Golgi apparatus ( Figure 3B ) . Trs33 is one of two nonessential subunits of the TRAPP complex , an essential vesicle tethering complex involved in ER-to-Golgi transport [37] . Vps51 is a member of the GARP complex , another vesicle tethering complex that promotes endosome-to-Golgi transport and retrograde transport within the Golgi [38] . Tlg2 is a t-SNARE that is important vesicle fusion within the Golgi [39] . These data make a strong prediction that the function of the unannotated Liv7 protein is in transport events involving the Golgi apparatus . Below we describe experiments in C . neoformans that support this prediction and additional follow-up experiments that led us to find that the Liv7 protein is required for the suppression mannose exposure on the cell surface and the suppression of mannose-dependent phagocytosis by mammalian macrophages . Given that Trs33 is a nonessential subunit of TRAPP , we anticipated that if Liv7 functions to promote TRAPP function in C . neoformans , that its gene deletion mutation should display a synthetic lethal or synthetic sick phenotype with a deletion of TRS33 in C . neoformans . We tested this prediction by creating single and double targeted knockouts of LIV7 and an ortholog of TRS33 we identified in the C . neoformans genome . We found that wild-type , liv7Δ , and trs33Δ strains all grow at approximately the same rate , with a doubling time of two hours ( Figure 3C . In contrast , the liv7Δ trs33Δ double mutant cells display a severe growth defect , having a doubling time of four hours ( Figure 3C ) . These data demonstrate that LIV7 and TRS33 interact genetically in C . neoformans , as inferred from analysis of the cross-species genetic interaction profiles described above . We also constructed a deletion in the gene coding for a member of the C . neoformans GARP complex , Vps52 ( we were unable to delete the C . neoformans VPS51 gene ) , and found that it displayed a growth phenotype similar to that of the liv7Δ trs33Δ double mutant . We next tested the hypothesis that LIV7 functions in the ER-Golgi system by using a chemical biology approach that takes advantage of the small molecule Brefeldin A ( BFA ) . BFA is a fungal secondary metabolite that inhibits eukaryotic Sec7-family guanine nucleotide exchange factors that are involved in vesicle transport and themselves localize to the membranes of the ER and Golgi apparatus [40]–[42] . BFA blocks anterograde transport from the ER to the Golgi , fusion of ER and Golgi compartments , and loss of Golgi apparatus itself [40] , [41] . We grew strains with and without a growth-inhibitory , sublethal concentration ( 40 µg/ml ) of BFA ( Figure 3D ) . Wild-type , liv7Δ , and trs33Δ show identical responses to BFA: a sharp increase in doubling time from two hours to over 12 hours ( p≤0 . 01 ) ( Figure 3D ) . liv7Δ trs33Δ mutants , which already exhibit slow growth ( p≤0 . 01 ) , do not show any further increase in their four-hour doubling time . The resistance to BFA exhibited by liv7Δ trs33Δ double mutants demonstrates that either Liv7 or Trs33 function is required for BFA to inhibit cell growth ( Figure 3D ) . These data could be explained if Liv7 and Trs33 have a severe defect in the assembly and/or function of the Golgi apparatus ( which we show to be the case below ) . In this scenario , the growth rate of such cells would thus not be affected by BFA since the have greatly reduced the target organelle most strongly affected by the drug . A more formal statement of such a model would be that in the absence of BFA , Liv7 and Trs33 redundantly promote growth ( via a role in Golgi biogenesis ) , but in the presence of the drug , cells convert to a state in which either Liv7 or Trs33 inhibits growth ( Figure 3E ) . This genetic behavior is analogous to that of the S . cerevisiae MAP kinase Kss1 , which is converted from an inhibitor of filamentous growth to an activator via phosphorylation by the upstream MAP kinase Ste7 [43] . The vps52Δ mutant also displays resistance to BFA ( Figure 3D ) . To further test the hypothesis that Liv7 functions in the Golgi , we examined the colocalization of an mCherry-tagged version of Liv7 with compartment markers . The levels of Liv7 protein are low and we could not detect it by Western or microscopy under yeast culture conditions ( data not shown ) . However , under the same tissue culture conditions we use to study pathogen phagocytosis ( DMEM , 5% CO2 , without shaking ) , we observed a punctate Liv7-mCherry signal that was well above background signal observed in an untagged control strain ( Figure 4A–4C ) . To label the ER and Golgi , we briefly incubated cells with a fluorescent derivative of Brefeldin A ( fBFA ) [44] at sub-inhibitory concentrations ( 0 . 5 µg/ml for 40 min , 80-fold less than the minimal inhibitory concentration ) . To confirm that the compound was labeling the anticipated compartments , we stained a C . neoformans strain with fBFA carrying a mCherry-tagged version of the conserved Erd2 protein , which is found in both ER and Golgi compartments [45] and found that the fBFA signal colocalizes with the Erd2 signal ( Figure S2 ) . Importantly , the Liv7-mCherry colocalizes with the fBFA signal . The respective puncta co-localize in almost 100% cells that display signals for both fluorophores ( Figure 4A–4C ) . As a control , we stained mitochondria with MitoTracker did not observe co-localization with Liv7-mCherry signal ( Figure S3 ) . To test whether mutations LIV7 and TRS33 impact the formation of the ER and Golgi we stained single and double mutants with fBFA . Wild-type , liv7Δ , and trs33Δ strains showed similar cytoplasmic punctate staining ( Figure 4D–4G ) . However , liv7Δ trs33Δ mutants did not exhibit detectable fBFA staining ( Figure 4G ) , consistent with a severe defect in organelle formation . These data show that Liv7 is important in promoting organelle formation in cells lacking Trs33 . Together with the impact of the mutants on BFA sensitivity and the colocalization of Liv7 with fBFA , these observations provide strong evidence for a role for Liv7 in Golgi function . Key functions of the Golgi include the sorting and modification of proteins and the biosynthesis of polysaccharides . The cell surface of microbes often contain pathogen-associated molecular patterns ( PAMPs ) , molecular signatures that are recognized by the mammalian immune system [46] . Previous studies of the human fungal pathogen Candida albicans has shown that there are mechanisms by which this pathogen masks PAMPs to order to avoid recognition by neutrophils [47] . To test whether LIV7 or TRS33 are involved in PAMP exposure , we examined the cell surface exposure of two well-established fungal PAMPs , mannose and β-glucan . These experiments were performed in tissue culture conditions , which modestly induces production of the C . neoformans polysaccharide capsule . In addition , we stained cells for the glucuronoxylomannan ( GXM ) component of the capsule and as well as the cell wall polysaccharide chitin . We used an antibody to detect glucuronoxylomannan ( GXM ) component of C . neoformans polysaccharide capsule ( Figure 5A ) , the lectin CBP to detect chitin ( Figure 4A ) , an antibody to detect β-glucan ( Figure 5B ) , and the lectin concanavalin A ( conA ) to detect exposure of mannose ( Figure 5C ) . Wild-type , liv7Δ , and trs33Δ all showed similar PAMP exposure , with modest staining of β-glucan and mannose under tissue culture growth conditions ( Figure 5A–5C ) . We also observed modest staining using reagents that detect chitin and GXM ( Figure 5A–5C ) . In contrast , we observed strikingly different results in liv7Δ trs33Δ double mutant cells and in the vps52Δ mutant . Most remarkably , we observed a dramatic increase in mannose exposure in these mutants as measured by conA staining ( Figure 5C ) . In contrast , GXM or β-glucan staining is virtually eliminated ( Figure 5A , 5B ) . The chitin signal is reduced in intensity and localizes to a focus at the cell pole . The increase in conA signal cannot be explained by the lack of capsular GXM in the double mutant , as GXM- and capsule-deficient mutant strains cap10Δ [48] and cap60Δ [48] , [49] do not exhibit this phenotype ( Figure 5C ) . These data suggest that LIV7 and TRS33 act redundantly in the transport of molecules required to suppress the exposure of mannose on the cell surface and that the integrity of the GARP complex is also required for this process . Mannose and mannoproteins ( mannan ) are highly immunogenic [50] , and , consequently , masking their exposure would be expected to be critical for pathogen evasion of the host immune system . It is well-established that C . neoformans evades phagocytosis by macrophages ( anti-phagocytosis ) , the first line of host immune defense , and that this attribute is important for mammalian infection [5] , [51] . In prior work , we demonstrated that C . neoformans evades phagocytosis by at least two pathways , one requiring capsule production and a second that is independent of capsule production and programmed by the transcriptional regulators Gat201 and Gat204 [5] , [22] . Strikingly , mutations that abrogate capsule formation and mutations in the capsule-independent pathway do not result in detectable exposure of mannose or β-glucan on the cell surface , suggesting that these pathways do not act by masking these known PAMPs , even though their exposure would be anticipated to activate phagocytic receptors on macrophages . Since we observed a dramatic increase in mannose exposure in the liv7Δ trs33Δ double mutant , we anticipated that it would display high levels of phagocytosis . To test this , we cultured wild-type , liv7Δ , trs33Δ , and liv7Δ trs33Δ C . neoformans cells with RAW 264 . 7 cells , a murine macrophage cell line . To test the potential impact of opsonization , C . neoformans strains were treated or not with fetal bovine serum prior to incubation ( Figure 6A ) . Wild-type C . neoformans displays a low level of phagocytosis ( 4% macrophages with associated C . neoformans cells ) that increased ( ∼17% ) upon opsonization ( p≤5×10−4 ) . As anticipated from their mannose exposure , liv7Δ trs33Δ mutants and vps52Δ mutants show high levels ( ∼80% ) of phagocytosis regardless of opsonization ( p≤2×10−3 ) . The lack of anti-phagocytosis activity by liv7Δ trs33Δ cells and vps52Δ cells is not solely due to lack of GXM , as GXM mutants cap10Δ and cap60Δ show increased association with macrophages ( p≤3×10−3 ) but not to the same extent as liv7Δ trs33Δ cells , and are still sensitive to opsonization ( p≤5×10−3 ) . Surprisingly , even though there was no gross increase in mannose exposure in the liv7Δ single mutant , it displays a small ( 11% ) but reproducible increase in phagocytosis without opsonization ( p≤5×10−5 ) and no further increase with opsonization . In contrast , the trs33Δ mutant does not show this phenotype . The single liv7Δ and trs33Δ mutants show distinct phagocytosis phenotypes yet the mannose exposure ( as determined by conA staining ) of both mutants is not distinguishable from wild-type . We hypothesized that liv7Δ cells might exhibit an increase in mannose or mannan on their surface not present in trs33Δ cells that is too subtle to detect by microscopy-based lectin staining assays . A functional prediction of this hypothesis is that the increase in phagocytosis of the liv7Δ mutant should be specifically blocked by an excess of free mannose . We performed phagocytosis assays using unopsonized C . neoformans cells and added either soluble mannose ( to block recognition of mannose and mannans by macrophage mannose-recognition receptors ) or laminarin ( a control oligosaccharide that blocks recognition of beta-glucan ) [52] . Strikingly mannose , but not laminarin , blocks the increased phagocytosis of liv7Δ mutants ( p≤10−3 ) ( Figure 6B ) . Mannose addition also partially rescues the anti-phagocytosis defect of liv7Δ trs33Δ cells ( p≤10−3 ) . Importantly , this treatment did not impact phagocytic index of gat204Δ cells ( Figure 6B ) , a mutant we described previously that produces similar increase in phagocytosis , supporting the view that Liv7 and Gat204 function via distinct mechanisms [22] . The genetic interaction profile produced by the expression of LIV6 in S . cerevisiae shows positive and negative correlations with the corresponding profiles of the S . cerevisiae syn8Δ and ecm21Δ deletion mutants , respectively ( Table 1 and Figure 7A ) . These genes act in endosome transport and/or fusion [28] , [29] , a process that mediates transport from either the plasma membrane or the late Golgi to the vacuole [53] . These correlations predict that Liv6 participates in endosomal functions in C . neoformans . We first tested this prediction by assessing the impact of Liv6 on vacuole number . S . cerevisiae genes involved sorting to the vacuole include those that function in endosome biology and often impact vacuole number and morphology [54]–[56] . Vacuoles can be detected by staining with LysoTracker Green ( Invitrogen ) , a dye that is taken up by the cell during endocytosis and fluoresces in acidified compartments , including endosomal vesicles , and typically strains the outer rims of vacuoles . Wild-type C . neoformans cells grown in yeast culture conditions and strained with LysoTracker show efficient uptake , many internal vesicles , and rim-stained vacuoles ( Figure 7B ) . This pattern is remarkably similar to those reported for S . cerevisiae stained with LysoTracker or FM4-64 [56] , [57] , an older vital stain used to study protein sorting to the vacuole [57] . Strikingly , liv6Δ cells consistently exhibit a greater number of vacuoles than wild-type cells ( p<0 . 005 ) ( Figure 7C , 7D ) . Notably , the S . cerevisiae gene SYN8 , whose deletion mutant's genetic interaction profile displays a positive correlation with the profile produced by LIV6 expression ( Table 1 ) , has been reported to function with another SNARE to promote normal vacuolar morphology [28] . The increase in vacuolar number seen in liv6Δ cells is highly specific , as knockout mutant in any of the bait genes did not exhibit a change in vacuole number ( Figure 7D ) . We next exploited the aminoglycoside antibiotic neomycin , which interferes with eukaryotic endosomal activity by binding phosphytidylinositol phosphates [58] , [59] necessary for endosome function [60] . As a consequence , loss-of-function mutations in S . cerevisiae genes involved in endosome function [61] , [62] are sensitive to neomycin [61] . Supporting a role for Liv6 in endosome function , we found that C . neoformans liv6Δ knockout mutants are sensitive to this drug ( Figure 7E ) . Cells lacking LIV7 display a subtle reproducible neomycin resistance which could be due to altered cell permeability , a characteristic of neomycin-resistant S . cerevisiae strains [63] , [64] . Knockout mutants in the other bait genes do not display a change in sensitivity to this compound . liv6Δ cells do not exhibit a growth defect on fluconazole , suggesting that their growth defect is specific to neomycin . Together , the changes in vacuole number and sensitivity to neomycin in produced by the liv6Δ mutation support the prediction from cross-species genetic profiles of a role for Liv6 in the endovacuolar system of C . neoformans .
Genetic approaches to understanding mechanisms of virulence in human fungal pathogens can efficiently identify genes necessary for pathogens to cause disease . However , a key roadblock to progress is the lack of tools that can help define the function of a gene product when its predicted sequence offers few clues to its biochemical function , a common occurrence . We described here a case study of a cross-species genetic interaction profiling approach to develop testable hypotheses for the function of fungal virulence factors of unknown function . Notably , this proof-of-principle study shows that the approach can provide information on fungal pathogenicity factors that lack S . cerevisiae orthologs . Although many studies have used S . cerevisiae to investigate the function of foreign genes [9]–[11] , [13] , [14] , the cross-species genetic interaction profile used here represents an application of quantitative genetic profiling of foreign proteins in S . cerevisiae coupled with comparison to recently described genetic map of S . cerevisiae [19] to the problem of annotation of fungal virulence factors . Because S . cerevisiae is a fungus , we anticipate that this approach may be particularly useful for fungal genes but that the method may also find utility in the study of bacterial and viral proteins that impact conserved intracellular processes in eukaryotic host cells . Our approach involves expression in S . cerevisiae of cDNAs encoding Cryptococcus neoformans virulence factors identified in systematic genetic screens; the generation of genetic profiles by assessing the effect of C . neoformans gene expression in the context of each nonessential S . cerevisiae deletion mutants; and , correlation analysis with the existing database of genetic interactions to develop testable functional hypotheses . As mentioned above , one mechanism whereby expression of a C . neoformans gene could produce impact S . cerevisiae would be “dominant-negative” effect thereby inhibiting the activity of an S . cerevisiae pathway . Our results with LIV7 in both S . cerevisiae and C . neoformans are consistent with this scenario . The expression of LIV7 in S . cerevisiae produces a profile that correlates with that of the S . cerevisiae trs33Δ deletion mutant , but in C . neoformans , the liv7Δ mutation produces a synthetic phenotype with the trs33Δ mutation . Alternatively , expression of a C . neoformans gene product could act in a “dominant-active” fashion to increase the activity of a pathway which might result in a negative correlation with the profile of a gene deletion in the corresponding pathway . With Liv6 , we observed both positive and negative correlations that led us to test a role in endosome function . Although we have focused on the extensive deletion mutant genetic interaction dataset [19] , comparisons of the cross-species profiles generated here with genetic interaction profiles produced using chemicals [65] , [66] and/or overexpressed genes [67] , [68] will likely be equally useful as these approaches are applied on a larger scale . Thus , the analysis of correlations between cross-species genetic interaction profiles and existing “within-species” genetic interaction profiles offers a tool for generating testable predictions for pathways in which foreign genes operate . The genetic profiling studies and validation experiments described in this paper provide new information on two C . neoformans pathogenicity factors identified previously , Liv7 and Liv6 . These proteins lack orthologs in S . cerevisiae and lack orthologs of known function in other species . Our studies of Liv7 suggest it functions in Golgi transport in a process that suppresses the exposure of the PAMP mannose on the cell surface ( Figure 6C ) . The increased phagocytosis phenotype of the liv7Δ single mutant and its specific suppression by soluble mannose appears specific to liv7Δ cells and is specific to mannose versus other carbohydrates ( Figure 6B ) . The anti-phagocytic properties of C . neoformans are critical for mammalian infection [5] , [22] , [69] and the capsule is important for the anti-phagocytosis activity of opsonized C . neoformans cells [51] . Our previous work identified a capsule-independent pathway necessary for anti-phagocytosis under unopsonized conditions [22] . The suppression cell surface exposure of PAMP mannose appears to represent a third anti-phagocytosis pathway ( Figure 6C ) since mannose does not rescue the anti-phagocytic defect of gat204Δ cells ( Figure 6B ) , which are defective in the capsule-independent anti-phagocytosis pathway [22] . This argument is supported by the observation that cap10Δ and cap60Δ cells , which lack GXM [48] , [49] , do not exhibit increased conA staining ( Figure 5C ) . We suggest that Liv7 is important for mammalian infection [5] because it inhibits macrophage recognition of mannose-containing patterns on the C . neoformans cell surface ( Figure 6C ) . Although our studies of Liv6 point to a role in endosome biology that impacts neomycin resistance and vacuole number , understanding how this function relates to its role in pathogen fitness in the host will require further investigation . One possibility is that Liv6 is involved in the endocytic uptake of limiting factors required for proliferation from the host milieu . One anticipates that functional annotation of fungal virulence factors identified genetically will continue to be a major challenge for the future . The approach described here represents one generic tool that could be applied to this problem on a larger scale . We expect that a substantial number of virulence genes of unknown function in fungal pathogens will impinge on conserved cellular processes and that their genetic profiling in S . cerevisiae could therefore yield testable functional predictions in a significant number of cases . The cross-species interaction profiling could also be useful for studying genes from highly virulent pathogens that are difficult to work with due to the requirement for extensive containment .
We inserted the GPD1 promoter region , our C . neoformans cDNA of interest , and a NAT resistance marker into the multicloning site of pRS316 . For recombination into S . cerevisiae , we cut with a restriction enzyme that cleaved within the URA3 locus , the transformed the linearized vector into S . cerevisiae using standard lithium acetate-based transformation techniques . We verified expression of C . neoformans genes by extracting total RNA from log-phase S . cerevisiae cultures grown at 30°C in YNB , selecting for mRNA , and making cDNA as previously described [70] . Expression of C . neoformans genes was verified by qPCR performed as previously described [70] . We expressed each C . neoformans gene under the GPD1 promoter and we measured RNA by qRT-PCR ( Figure S4 ) . We then measured the levels of BUD1 mRNA in the same RNA preparation . BUD1 is a small GTPase expressed at low levels [71]–[73] along with its two co-regulators BUD2 and BUD5 [74] , [75] . We used published data on the molecules of BUD1 RNA per cell averaged with co-regulators BUD2 and BUD5 [71]–[73] to estimate the number of RNA molecules per cell for C . neoformans genes from the ratios in Figure S4 , then calculated its rank position compared to other S . cerevisiae genes . BLP1 and LIV6 were in the lowest 10% of genes with detectable RNA ( ∼5090 of ∼6580 genes had detectable RNA [71]–[73] ) . LIV5 and LIV7 were in the 10–20th percentile , as were the BUD genes . MEP1 was the best expressed of the C . neoformans genes , ( ∼35th percentile ) . LIV13 was expressed based on the increase in LIV13 primer products with and without RT ( data not shown ) but not compared to BUD1 . We performed SGA screens as described in Tong et al [16] , [17] using a RoToR colony pinning robot ( Singer Instruments ) . All screen plates were scanned on a flatbed scanner with autofocus . We extracted colony size data using the publicly available ScreenMill software [76] . We then adjusted the raw colony size data to control for plate position , edge effects , and slow growth of knockout mutants using the S-score method developed by Collins et al [20] . The final S-scores , one for each double mutant strain , indicate the strength of the genetic interaction ( absolute value ) and whether the interaction is synthetically sick ( negative numbers ) or buffered ( positive numbers ) [20] . We then adjusted S-scores so that they were on a scale between −1 . 0 and 1 . 0 and calculated the Pearson correlation between S-scores and ε-scores from Constanzo et al [19] . We calculated p-values of the Pearson correlations by calculating the Z-score of the Pearson correlation for each interaction , then using the Z-score to determine the p-value of each interaction . C . neoformans was routinely grown in yeast culture conditions in either YPAD ( 1% yeast extract , 2% peptone , 2% glucose , 0 . 015% L-tryptophan , 0 . 004% adenine ) or yeast nitrogen base ( YNB ) ( Difco ) . Strain construction and genetic manipulation was previously described [5] . Whenever more than one knockout mutant for a single gene is shown , mutants were made by independent transformations . Growth curves ( Figure 3A ) were performed in YNB at 30°C by taking measurements of OD600 every two hours for 10 hrs . The growth curve was repeated three times and representative data are shown . When C . neoformans cells were grown in tissue culture conditions , they were first grown overnight to saturation in YNB , then washed once in 1× PBS and resuspended at a density of 1 OD600/ml ( ∼1 . 7×107 cells/ml ) in DMEM , then incubated for the specified amount of time in 5% CO2 at 37°C . Samples were grown in overnight in YNB at 30°C , then subcultured to OD600 = 0 . 2 . BFA or DMSO ( -BFA control ) was added to each culture and the OD600 taken every hour for 10 hr . Doubling time was calculated over the interval from 4–8 hr . The treatment curve was repeated three times and the data shown are averages of the three experiments . Samples were grown overnight in YNB at 30°C , then washed 3× in 1× PBS and resuspended at 1 OD/ml in DMEM , then incubated 16 hr under tissue culture conditions ( 5% CO2 , 37°C ) . Samples were then either imaged ( unstained samples ) or fBFA ( Life Technologies ) was added to the medium to a final concentration of 0 . 5 µg/ml . fBFA samples were incubated 40 min , washed 1× in PBS , then imaged immediately . C . neoformans cells were grown overnight under yeast culture conditions ( yeast nitrogen base ( YNB ) , 30°C with rotation ) , then subcultured to OD600 of ∼0 . 2 and grown to midlog phase . LysoTracker Green was added to a final concentration of 500 nM and incubated for five minutes with shaking at 30°C . Cells were then harvested and immediately imaged . Strains were grown under tissue culture conditions for 12 hr . MitoTracker Green ( Invitrogen ) was added to a final concentration of 1 µM ( from 1 mM stock in DMSO ) , incubated 30 min at 37°C , then imaged . Samples were grown overnight in YNB at 30°C , then washed three times in 1× PBS and resuspended at 1 OD/ml in DMEM , then incubated 16 hr under tissue culture conditions ( 5% CO2 , 37°C ) . Samples were then fixed for 15 min in 4% paraformaldehyde , washed three times in 1× PBS , and then used for staining . To stain with concanavalin A ( conA ) staining for mannose residues , cells were incubated 5 min in 50 µg/ml Alexa Fluor 594 ( Invitrogen ) , washed once in 1× PBS , then imaged . Samples for staining for chitin and GXM were incubated with αGXM antibody mAb 339 ( 1∶1000 ) as previously described [5] and fluorescein-conjugated chitin binding domain ( New England Biolabs ) ( 1∶500 ) for 4 hr , then washed twice in 1× PBS and incubated with TRITC-conjugated donkey anti-mouse secondary antibody ( Jackson ImmunoResearch ) and fluorescein-conjugated chitin binding domain ( 1∶500 ) for 1 hr . Samples were then washed once and imaged using an Axiovert 200 M ( Zeiss ) microscope running Axiovision software . β-glucan staining was performed using the same procedure as GXM staining but with anti-β-glucan antibody ( 1∶1000 ) ( Biosupplies Australia ) . Phagocytosis assays were performed as previously described [5] , [22] . RAW264 . 7 macrophages ( 2×104 cells/well ) were seeded into 96-well tissue-culture treated plates in DMEM medium and allowed to adhere overnight . C . neoformans cells grown in YPAD medium were washed three times with PBS , then resuspended to a density of 107 cells/ml in PBS . 200 µl fresh DMEM was added to RAW264 . 7 cells . 5 µl C . neoformans culture ( 5×104 cells ) were then added to each well for a multiplicity of infection of two yeast to one macrophage . Following 24 hr co-incubation , the macrophages were washed three times with PBS to remove unphagocytosed yeast and then fixed with 1% formaldehyde/PBS prior to visualization on an inverted light microscope . Percentage of yeast cell-associated macrophages was determined by counting the number of macrophages with yeast internalized or associated with their cell surface , divided by the number of macrophages counted . At least 200 macrophages were assayed per well , and each strain was tested in triplicate . If performing phagocytosis experiments under opsonizing conditions , C . neoformans cells were grown overnight in YNB , washed three times in 1× PBS , resuspended to a density of 107 cells/ml in either fetal bovine serum ( opsonized samples; FBS ) or 1× PBS ( unopsonized samples ) , incubated for 30 min at 30°C on a shaking platform , washed once in 1× PBS , then resuspended at 107 cells/ml in 1× PBS and used to infect macrophages as above .
|
HIV/AIDS patients , cancer chemotherapy patients , and organ transplant recipients are highly susceptible to infection by opportunistic fungal pathogens , organisms common in the environment that are harmless to normal individuals . Understanding how these pathogens cause disease requires the identification of genes required for virulence and the determination of their molecular function . Our work addresses the latter problem using the yeast Cryptococcus neoformans , which is estimated to cause 600 , 000 deaths annually worldwide in the HIV/AIDS population . We describe a method for determining gene function in which C . neoformans genes are expressed in deletion mutants of all nonessential genes of the well-studied model yeast S . cerevisiae . By examining the impact on growth ( enhancement or suppression ) we generated “cross-species” genetic interaction profiles . We compared these profiles to the published genetic interaction profiles of S . cerevisiae deletion mutants to identify those with correlated patterns of genetic interactions . We hypothesized that the known functions of S . cerevisiae genes with correlated profiles could predict the function of the pathogen gene . Indeed , experimental tests in C . neoformans for two pathogenicity genes of previously unknown function found the functional predictions obtained from genetic interaction profiles to be accurate , demonstrating the utility of the cross-species approach .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"genetics",
"biology",
"genomics",
"microbiology",
"genetics",
"and",
"genomics"
] |
2012
|
Approaching the Functional Annotation of Fungal Virulence Factors Using Cross-Species Genetic Interaction Profiling
|
During embryonic development , signalling molecules known as morphogens act in a concentration-dependent manner to provide positional information to responding tissues . In the early zebrafish embryo , graded signalling by members of the nodal family induces the formation of mesoderm and endoderm , thereby patterning the embryo into three germ layers . Nodal signalling has also been implicated in the establishment of the dorso-ventral axis of the embryo . Although one can infer the existence of nodal gradients by comparing gene expression patterns in wild-type embryos and embryos in which nodal signalling is diminished or augmented , real understanding can only come from directly observing the gradients . One approach is to determine local ligand concentrations in the embryo , but this is technically challenging , and the presence of inhibitors might cause the effective concentration of a ligand to differ from its actual concentration . We have therefore taken two approaches to visualise a direct response to nodal signalling . In the first , we have used transgenic embryos to study the nuclear accumulation of a Smad2-Venus fusion protein , and in the second we have used bimolecular fluorescence complementation to visualise the formation of a complex between Smad2 and Smad4 . This has allowed us to visualise , in living embryos , the formation of a graded distribution of nodal signalling activity . We have quantified the formation of the gradient in time and space , and our results not only confirm that nodal signalling patterns the embryo into three germ layers , but also shed light on its role in patterning the dorso-ventral axis and highlight unexpected complexities of mesodermal patterning .
During embryonic development , secreted molecules known as morphogens generate a concentration gradient of positional information that instructs developing tissues to adopt particular cell fates [1] . One example of this phenomenon is the patterning of the early zebrafish embryo by the nodal/TGF-β signalling pathway [2] . Thus , the zebrafish nodal ligands squint ( sqt ) and cyclops ( cyc ) are expressed in the most marginal cells of the developing zebrafish embryo and homozygous mutations in both sqt and cyc result in embryos that lack all endoderm and mesoderm , apart from a few somites in the tail [3] . Similar phenotypes are achieved through the loss of nodal signalling by mutation of both maternal and zygotic one-eyed-pinhead ( MZoep ) or by misexpression of the nodal antagonist lefty [4 , 5] . Misexpression of sqt and cyc in the zebrafish animal pole indicate that cyc acts only over short distances , whereas sqt functions as a morphogen and exerts its effects over long distances to induce target gene expression [2] . High levels of nodal signalling activate goosecoid ( gsc ) expression , whereas lower levels activate no-tail ( ntl ) . Thus , gsc is expressed in cells near a source of sqt , and ntl is expressed in cells further away . The correct regulation of ntl is essential for patterning of the zebrafish embryo , because homozygous mutations in ntl disrupt mesoderm and notochord formation [6] . The same tissues are disrupted in Xenopus embryos lacking Brachyury ( Xbra ) [7] , and ectopic expression of Xbra in isolated animal regions converts ectodermal cells into a mesodermal fate [8] . Consistent with the requirement of nodal signalling for mesoderm formation , ntl expression fails to initiate in embryos with diminished nodal signalling [3–5] . Together , these experiments suggest that nodal family members form a gradient that induces target gene expression and specifies mesoderm and endoderm . Activation of the nodal signalling pathway within a cell results in the phosphorylation of Smad2 , which then interacts with Smad4 [9] . The resulting Smad2/4 complex translocates to the nucleus where it activates the transcription of target genes . To visualise the formation of a nodal gradient , we have first made use of transgenic embryos expressing a Smad2-Venus fusion protein under the control of a ubiquitous promoter: nodal signalling causes such constructs to enter the cell nucleus [10] . In addition , however , we have exploited the greater signal-to-noise ratio afforded by the technique of bimolecular fluorescence complementation ( BiFC ) [11] . In this approach , the N- and C-terminal halves of a fluorescent protein are brought into proximity by interactions between the two unrelated proteins to which they are fused . They can then assemble into a functional fluorescent protein that can be detected by conventional microscopy . This approach has previously been used to visualise , in a quantitative manner , interactions between Smad2 and Smad4 in the Xenopus embryo [12] . Yolk autofluorescence in Xenopus prevented a proper study of endogenous signalling events [12] , but in this Research Article , we show that the technique is effective in the zebrafish embryo . Our data allow us to follow in space and time the formation of a gradient of nodal signalling activity within the developing zebrafish embryo . The results illustrate the dynamics of gradient formation , and in contrast to previous studies [13] , clearly demonstrate a role for nodal signalling in dorso-ventral patterning , explaining why target genes such as gsc are only expressed in dorsal marginal cells . Our data also highlight the complexities of ntl regulation and of the formation of the border between dorsal mesendoderm and the neural plate .
In a preliminary attempt to investigate nodal signalling levels during zebrafish development , we generated transgenic embryos that express a Smad2-Venus fusion protein under the control of a ubiquitous promoter . During zebrafish development , marginal cells are thought to receive the highest levels of endogenous nodal signalling while cells at the animal pole experience low levels , if any [14] . As predicted , nuclei of animal pole cells of transgenic embryos were only weakly fluorescent ( Figure 1A ) , but expression of a constitutively active version of the TGF-β receptor Taram-A-D ( Taram-D* ) [15] , caused strong nuclear fluorescence ( Figure 1B ) . We note that in unstimulated cells , Smad2-Venus appeared to be concentrated at the centrosomes , and fluorescence accumulates in the nucleus shortly before nuclear envelope breakdown , only to disperse about a minute later , and then weakly associate with the mitotic apparatus ( Video S1 ) . In an effort to improve the signal-to-noise ratio in such experiments , we turned to BiFC . When zebrafish embryos were injected with the N- and C- terminal halves of a modified form [12] of the fluorescent protein Venus [16] , no fluorescence was observed , demonstrating that these fragments are suitable for BiFC experiments in this species ( Figure 1C , C' ) . We therefore created fusions of the N- and C- terminal halves of Venus with the N termini of zebrafish Smad2 and Smad4 , respectively , to create VNSmad2 and VCSmad4 . When these constructs were expressed in the zebrafish embryo , ntl expression was unaffected in 90% of cases ( n = 124 ) , and in the remaining embryos , expression was normal in the marginal zone with weak ectopic expression in animal pole cells ( unpublished data ) . These experiments demonstrate that our Smad BiFC constructs are suitable reagents for the analysis of endogenous nodal signalling . Consistent with the experiments described above , we observed no nuclear BiFC fluorescence in animal pole cells of embryos injected with VNSmad2 and VCSmad4 ( Figure 1D ) . As in the Xenopus embryo [12] , however , and in contrast to the behaviour of Smad2-Venus , intense fluorescence appeared to be associated with chromosomes during cell division ( Figure 1D , arrows , and Video S2 ) . This is discussed below . When embryos received injections of both Smad BiFC constructs and the constitutively active version of Taram-A-D [15] , strong nuclear fluorescence was observed in animal pole cells ( Figure 1E ) . Similar results were observed when embryos were co-injected with RNA encoding our Smad2/4 BiFC constructs and the TGF-β ligand sqt ( unpublished data ) . Activation of the TGF-β signalling pathway results in the phosphorylation of receptor-regulated Smads in their C-terminal SXS motifs [9] . Deletion of the SXS phosphorylation site in the VNSmad2 construct ( VNS2ΔSXS ) abolished TGF-β induced nuclear fluorescence ( compare Figures 1E and 1F ) . Together , these experiments demonstrate that our Smad2-Venus transgenic embryos and Smad BiFC constructs report the activation of the TGF-β signal transduction pathway in the zebrafish embryo . We first investigated endogenous nodal signalling in zebrafish embryos at 5–6 hours post fertilisation ( hpf ) , when they express ntl and experience endogenous nodal signalling [17] . Observation of Smad2-Venus transgenic embryos at 6 hpf revealed a gradient of nuclear fluorescence that was high at the margin and decreased towards the animal pole ( Figure 2A ) , indicating that there is a gradient of nodal signalling in the developing embryo . This impression was confirmed by use of Smad2/4 BiFC , where high levels of nuclear fluorescence were observed in marginal cells , with intensity gradually decreasing as distance from the margin increased ( Figure 2B ) . This pattern of nuclear fluorescence was not observed in embryos injected with BiFC constructs lacking the TGF-β phosphorylation site ( VNS2ΔSXS/VCS4 ) ( Figure 2C ) , in embryos expressing the nodal antagonist lefty [4] ( 96%; n = 25 ) ( Figure 2D ) , or in MZoep embryos ( 100%; n = 15 , unpublished data ) . Smad2-Venus transgenic embryos do not exhibit detectable nuclear fluorescence in the yolk syncytial layer ( YSL ) of the embryo ( Figure 2A ) , and nor do we observe Smad2/4 BiFC fluorescence in YSL nuclei of embryos co-labelled with a fluorescent histone marker ( Figure 2E ) . These observations have allowed us to use Volocity software ( Improvision ) to quantify nuclear Smad2-Venus and Smad2/4 BiFC fluorescence intensity from the margin to the animal pole at different stages , defining the average intensity and average position of the YSL nuclei as zero ( Figure 3A and 3B ) . We found that the most marginal nuclei , nearest the YSL , had the greatest Smad2-Venus fluorescence ( Figure 3A ) and the greatest Smad2/4 BiFC ( Figure 3B ) . The nuclear fluorescence decreased in cells closer to the animal pole , some 200 μm away . Interestingly , nuclei positioned close to each other frequently had very different levels of nuclear Smad2-Venus and Smad2/4 BiFC fluorescence ( Figure 3A and 3B; see also Figure 2A and 2B ) . One possibility is that these differences reflect local variation in effective nodal concentrations . Alternatively , there may be cell cycle–dependent variations in signal level associated with the intense fluorescence during cell division ( Figure 1D , arrow , and Video S1 ) . We went on to investigate the spatial and temporal patterns of Nodal signalling by allowing embryos to continue development after imaging and then noting the positions of the imaged cells relative to the shield . This analysis exploited the superior signal-to-noise ratio of the Smad2/4 BiFC technique ( see Figures 1 and 2 ) . In preliminary experiments , analysis of lateral nuclei revealed that cells have higher levels of Smad2/4 signalling at 6 hpf compared to 5 hpf ( Figure 3C; blue points are 5 hpf and red are 6 hpf ) . To improve our understanding of the spatio-temporal aspects of these signalling events , we calculated the average nuclear Smad2/4 BiFC intensity in 25-μm intervals from the margin towards the animal pole in several different embryos ( Figure 3D–3H ) . We defined regions as dorsal , lateral , or ventral if the imaged cells were positioned within the dorsal quarter , lateral two quarters , or ventral quarter of the embryo , respectively . This analysis was performed for dorsal , lateral , and ventral cells at 5 and 6 hpf . As observed in individual embryos ( Figure 3C ) , equivalently positioned cells have greater nuclear BiFC intensities at 6 hpf compared with 5 hpf , consistent with the idea that these cells experience increasing levels of nodal signalling during this period ( Figure 3D–3F ) . When dorsal , lateral , and ventral cells were compared , we observed that lateral and ventral cells experience near identical levels of nodal signalling but dorsal cells experience higher levels ( Figure 3G and 3H ) . To place our observations in the context of normal development , we studied the expression profile of the nodal target gene ntl ( Figure 4A–4C ) . ntl is first activated on the dorsal side of the embryo at 4 hpf ( Figure 4A ) . Expression then spreads laterally , and by 5 hpf transcripts are detectable 3–5 cells deep throughout the margin ( Figure 4B ) . By 6 hpf the ntl expression domain has doubled , and is now approximately 12–14 cells deep ( Figure 4C ) . The expansion of the ntl domain observed over this period reflects the increasing levels of nuclear BiFC fluorescence and of Smad2/4 signalling ( Figure 4D compare blue and black trend lines ) . Our BiFC results show that the highest levels of Smad signalling occur at the dorsal side of the zebrafish embryo near the margin , where gsc is expressed ( Figures 3G , 3H , and 4D , compare red and black trend lines ) . Consistent with this observation , work in zebrafish and Xenopus indicates that activation of gsc requires higher levels of nodal or activin-like signalling than are required to induce Brachyury [2 , 18] . These results suggest , in contrast to previous proposals [13] , that a gradient of nodal signalling specifies the dorso-ventral axis of the zebrafish embryo . To explore this point in more detail , we expressed increasing amounts of sqt in the embryo . Our results showed that as levels of sqt increased , the domain of gsc expression expanded both animally and ventrally , as exogenously introduced sqt supplemented levels of the endogenous protein ( Figure 5B-5D and 5F ) . In an effort to correlate , in a quantitative manner , sqt signalling with Smad2/4 BiFC and gsc expression , we injected embryos at the one-cell stage with increasing amounts of sqt mRNA . At 6 hpf , we then measured Smad2/4 nuclear BiFC in the animal pole cells of some of the embryos and processed the remaining embryos for gsc expression . Injection of 1–4 pg of sqt mRNA resulted in an expansion of the gsc expression domain , but few embryos expressed gsc at the animal pole ( Figure 5G , boxed area ) . Injection of 5 pg of sqt mRNA resulted in a significant increase in the percentage of embryos that expressed gsc in animal pole cells ( Figure 5H ) , suggesting that the threshold for activation of gsc lies between 4 and 5 pg of sqt mRNA . Quantification of nuclear Smad2/4 BiFC fluorescence in the animal pole cells of injected embryos demonstrated that as the levels of sqt increased , so did fluorescence intensity ( Figure 5I ) . Based on these data , our results indicate that the threshold for the activation of gsc expression is represented by a nuclear Smad2/4 BiFC intensity between 0 . 60 and 0 . 65 ( Figure 5I ) . The only cells to experience endogenous levels of Smad2/4 BiFC that exceed this threshold are dorsal marginal cells ( Figure 3H ) . Our results are consistent with the idea that nodal signalling patterns the dorso-ventral axis of the zebrafish embryo as well as the animal-vegetal axis . But is nodal signalling the prime mover for dorso-ventral patterning in the zebrafish , or do sqt and cyc act downstream of BMP family members ? Embryos lacking BMP signalling become dorsalised and fail to form ventral tissues [19] , so it is possible that the dorso-ventral axis is first established by the ventral activation of the BMP signal transduction pathway , and it is this that directs the spatial distribution of nodal signalling and the dorsal activation of genes such as gsc . In this model , all dorso-ventral patterning would depend on BMP signalling , so to address the idea we injected embryos at the one-cell stage with a dominant negative BMP receptor ( dnBMPr ) and then studied the expression of gsc . Injected embryos became elongated ( Figure 6A and 6B; 190/203 elongated at 4-somite stage ) and appeared strongly dorsalised [20]; by 24 hpf almost all had died , with the survivors displaying slightly weaker dorsalised phenotypes ( 9 = c4 , 9 = c3 , and 7 = c2 ) . gsc expression was unaffected in embryos injected with RNA encoding dnBMPr ( Figure 6C–6F ) , indicating that the establishment of dorso-ventral patterning and the spatial distribution of nodal signalling is independent of BMP signalling .
Our results are consistent with the idea that nodal signalling patterns the animal-vegetal axis of the zebrafish embryo , with changes in the distribution and intensity of Smad signalling being reflected in changes in the spatial expression pattern of the nodal target gene ntl ( Figure 4 ) . In addition , we observe that over expression of sqt causes the expression domain of gsc to extend towards the animal pole ( Figure 5E and 5F ) . However , in contrast to previous conclusions based on cell lineage and gene expression experiments [13] , our data also suggest that nodal signalling plays a role in patterning the dorso-ventral axis of the zebrafish . In particular , we note that there are higher levels of Smad2/4 BiFC fluorescence in dorsal regions than in lateral and ventral regions ( Figures 3G , 3H , and 4D ) and that gsc , whose expression requires higher levels of nodal signalling than does ntl [2] , is expressed in these regions of elevated fluorescence . Consistent with this model , our correlation of Smad2/4 BiFC intensity with ectopic gsc expression ( Figure 5H and 5I ) demonstrates that the only cells to go above the gsc threshold are dorsal marginal cells . In addition , we note that over half of the cells of the prospective endoderm , a tissue whose formation also requires high levels of nodal signalling , are located dorsally [27] . If high levels of nodal signalling are indeed required for dorsal fates and lower levels for lateral and ventral tissues , then increased nodal signalling should produce a ventral shift in dorsal fates and loss or attenuation of nodal signalling should result in a dorsal shift of ventral fates . Consistent with this model , increased nodal signalling expands the expression domain of gsc in a ventral direction ( Figure 5A–5D ) and loss of nodal signalling results in a dorsal shift of the ventral marker gata2 [5] . Similarly , fate mapping experiments demonstrated that cells fated to become pronephros and midbrain , which in wild-type embryos are located in ventral and lateral positions respectively , shift towards the dorsal side of sqt-/-;cyc+/- embryos [13] . However , some ventral markers , such as spt and vox , are not expanded dorsally in embryos with reduced nodal signalling [13] . It is likely that these genes are regulated by BMP family members [5]; if BMP signalling is attenuated , ventral tissues fail to form and embryos become dorsalised [19] . Significantly , we found that the expression of a dominant negative BMP receptor had no effect on the expression of gsc ( Figure 6C–6F ) . This suggests that the elevated levels of nodal signalling at the dorsal side of the embryo occur independently of BMP signalling . Previous work has demonstrated that Wnt/β-catenin signalling is also required for the specification of dorsal cell fates and that ectopic activation of β-catenin induces the expression of gsc [13 , 28] . However , sqt is not expressed in embryos with disrupted β-catenin signalling , and β-catenin cannot induce gsc expression in sqt mutant embryos [13 , 28] . Together with our correlation of nodal signalling and gsc expression , these results indicate that the effects of β-catenin are mediated by nodal signalling . In combination with the results described above , our data therefore indicate that patterning of the zebrafish dorso-ventral axis involves high levels of BMP signalling in ventral tissues and high levels of nodal signalling in dorsal regions , effectively setting up a double gradient . It is also possible , as in the Xenopus embryo [29] , that BMP ventralises the embryo only after the onset of gastrulation . As discussed above , the dynamic expression pattern of ntl ( Figure 4B and 4C ) reflects the spatial changes in nodal signalling that occur in the margin of the zebrafish embryo between 5 and 6 hpf ( Figure 4D ) . Expression of both cyc and sqt declines between 5 and 6 hpf [13] , so it is likely that the increased level of signalling experienced by cells positioned away from the margin at 6 hpf derives from nodal ligand that has traversed cell tiers 1–6 during this period . At 6 hpf , Smad signalling extends farther towards the animal pole in dorsal regions of the embryo than in lateral and ventral regions ( Figure 3H ) , yet ntl is expressed in approximately the same number of cell tiers throughout the margin , and does not spread into prospective neural tissue at the dorsal side [30] . This suggests that ntl expression is repressed in the prospective neural plate , perhaps , as in Xenopus , in a Sip-1–dependent manner [31 , 32] . It is possible that the Smad signalling that occurs in the neural plate provides positional information to this tissue; injection of increasing concentrations of lefty results in the gradual loss of hindbrain structures , whereas prospective forebrain tissues are converted into hindbrain structures following expression of cyc [4] . At 5 hpf , the expression of sqt and cyc is uniform throughout the margin of the zebrafish embryo [13] , so what might cause the activation of Smad signalling to be higher in dorsal regions ? Evidence suggests that the duration of signalling as well as the concentration of the morphogen may determine cell fate [17 , 33 , 34] , and it may be significant that expression of sqt both commences on the dorsal side of the embryo and persists for longer in this region [13] . The elevated level of Smad2/4 BiFC in dorsal regions may therefore reflect both signal intensity and signal duration in the developing embryo . Our observations of transgenic embryos expressing Smad2-Venus indicate that Smad2 is associated with the centrosome , and comparison with results obtained with Smad1 [35] suggest that this might represent Smad2 that is destined for degradation . This is under investigation . We also noted that Smad2-Venus entered the nucleus shortly before nuclear envelope breakdown , and in this respect , its behaviour resembled that of cyclin B1 , which translocates to the nucleus after phosphorylation by Polo-like kinase 1 [36–38] . We do not yet know if the translocation of Smad2-Venus is regulated by phosphorylation , but if it were , this newly phosphorylated Smad2 might then be able to associate with Smad4 and form a complex on the chromosomes . We do not understand the significance of such an association , although one possibility is that it ensures an equal distribution of Smads between daughter cells , as is thought to occur for Sara-containing endosomes in the developing fly wing [39] .
The Smad2-Venus fusion was generated by PCR amplification of Venus and cloning into a pCS2-Smad2 plasmid , thus generating a fusion of Venus to the N terminus of Smad2 . The Smad2-Venus fusion was then subcloned into a miniTol vector containing the Xenopus EF1α , mcFos promoter . Transgenic embryos were generated by injecting embryos at the one-cell stage with 15 pg of Smad2-Venus miniTol plasmid and with 12 . 5 pg of transposase RNA . Injected embryos were raised to adulthood and then outcrossed to generate stable transgenic lines . All constructs were injected in volumes of 2 nl into the yolk of zebrafish embryos at the one-cell stage , and embryos were then incubated at 28 °C . Where stated , embryos were injected with 100 pg of histone CFP [12] , 50 pg of VNSmad2 , 50 pg of VCSmad4 , 300 pg of Lefty ( Antivin ) [4] , 1 pg of Taram-A-D , 2 . 5–10 pg of sqt [2] , or 800 pg of truncated dominant negative BMP receptor [40] . Zebrafish Smad2 and Smad4 open reading frames were amplified by PCR , cloned into the BiFC constructs [12] , and sequenced . The VNS2ΔSXS construct was created by introducing a stop codon into the VNS2 plasmid using PCR based mutagenesis with the primers 5′-TTAGGACATACTTTAGCAGCGTACGGAGGGGGAGCCCATC- 3′ and 5′-GATGGGCTCCCCCTCCGTACGCTGCTAAAGTATGTCCTAA-3′ . All RNA was synthesised using SP6 mMessage mMachine according to the manufacturer's instructions ( Ambion ) . Whole mount in situ hybridisation was performed essentially as described [41] , using probes specific for ntl [30] and gsc [42] . For imaging , embryos were de-chorionated and embedded in 0 . 3% agarose . Images were obtained with Perkin Elmer spinning disc and Olympus FV1000 inverted confocal microscopes using 40× lenses . All quantifications were performed by sequential imaging of CFP and Venus fluorescence using the Olympus FV1000 microscope . Ten 1-μm Z sections of the cells nearest the lens ( based on focal plane ) were imaged . Following imaging embryos were incubated at 28 °C until 6–7 hpf . The agarose dish was then placed in hot water to melt the agarose , the embryos were removed from the agarose using forceps , and the positions of the imaged cells in relation to the shield was noted . Individual Z sections were used for the quantification of animal pole cells . Fluorescence intensity was quantified using Volocity software ( Improvision ) . Individual nuclei were identified using a protocol to mark objects with intensities between 10 and 100% in the CFP ( histone ) channel . Quantifications were analysed using Microsoft Excel . For each image , the nuclei of the YSL were identified and the average distance and intensity of these nuclei was subtracted from all nuclei in that image . Video S1 was made using the Perkin Elmer spinning disc microscope .
|
One of the earliest events in vertebrate embryonic development is the patterning of the embryo into three germ layers: the ectoderm , mesoderm , and endoderm . Morphogens are signalling molecules that act in a concentration-dependent manner to induce the formation of different cell types . Members of the nodal family are thought to form a morphogen gradient in the developing zebrafish embryo and to be essential for pattern formation . Mesoderm and endoderm are believed to develop due to high levels of nodal signalling , while cells experiencing the lowest concentrations of nodal signalling become ectoderm . Although this idea is widely accepted , the formation of a nodal morphogen gradient has never been observed directly , and we have therefore used two different approaches to visualise the intensity of nodal signalling within individual cells . Our approaches have allowed us to visualise a gradient of nodal signalling activity in the developing zebrafish embryo . Quantification of the levels of nodal signalling experienced by individual cells confirms that nodal signalling patterns the animal-vegetal axis of the zebrafish embryo and , in contrast to previous studies , also suggests that it plays a role in patterning the dorso-ventral axis of the zebrafish embryo .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"cell",
"biology",
"developmental",
"biology"
] |
2009
|
Visualisation and Quantification of Morphogen Gradient Formation in the Zebrafish
|
Codon pair bias deoptimization ( CPBD ) has enabled highly efficient and rapid attenuation of RNA viruses . The technique relies on recoding of viral genes by increasing the number of codon pairs that are statistically underrepresented in protein coding genes of the viral host without changing the amino acid sequence of the encoded proteins . Utilization of naturally underrepresented codon pairs reduces protein production of recoded genes and directly causes virus attenuation . As a result , the mutant virus is antigenically identical with the parental virus , but virulence is reduced or absent . Our goal was to determine if a virus with a large double-stranded DNA genome , highly oncogenic Marek’s disease virus ( MDV ) , can be attenuated by CPBD . We recoded UL30 that encodes the catalytic subunit of the viral DNA polymerase to minimize ( deoptimization ) , maximize ( optimization ) , or preserve ( randomization ) the level of overrepresented codon pairs of the MDV host , the chicken . A fully codon pair-deoptimized UL30 mutant could not be recovered in cell culture . The sequence of UL30 was divided into three segments of equal length and we generated a series of mutants with different segments of the UL30 recoded . The codon pair-deoptimized genes , in which two segments of UL30 had been recoded , showed reduced rates of protein production . In cultured cells , the corresponding viruses formed smaller plaques and grew to lower titers compared with parental virus . In contrast , codon pair-optimized and -randomized viruses replicated in vitro with kinetics that were similar to those of the parental virus . Animals that were infected with the partially codon pair-deoptimized virus showed delayed progression of disease and lower mortality rates than codon pair-optimized and parental viruses . These results demonstrate that CPBD of a herpesvirus gene causes attenuation of the recoded virus and that CPBD may be an applicable strategy for attenuation of other large DNA viruses .
The attenuation by codon pair ( bias ) deoptimization ( CPBD ) has enabled rapid and highly efficient attenuation of a wide variety of RNA viruses , including Enterovirus C ( poliovirus ) [1] , Influenza A virus ( IAV ) [2–4] , Human immunodeficiency virus type 1 ( HIV-1 ) [5] , Human respiratory syncytial virus [6] , Indiana vesiculovirus [7] , and Dengue virus [8] . Some of the recoded viruses have shown 100 , 000-fold attenuation in comparison to their virulent parental viruses , and have been successfully used as highly protective experimental vaccines in mice and ferrets in the case of IAV [3 , 9] . In contrast to existing attenuation methods , CPBD-based vaccines can be designed within minutes and produced synthetically within days . The attenuation by CPBD is based on the observation that some codon pairs are found in protein coding sequences significantly less or more frequently than expected . The attenuation by CPBD involves reshuffling of the available codons in viral genes with the goal to maximize the number of codon pairs that are underrepresented in the protein coding sequences in the respective host . The recoded gene encodes the same protein and has the same codon bias as the parental gene , but its codon pair bias is perturbed . As a result , the mutated virus is antigenically identical with the pathogenic parent , but its virulence is reduced or absent . While the exact mechanism of attenuation by CPBD remains unknown , it is suggested that underrepresented codon pairs create unfavorable conditions for protein translation , modifications , or folding , which results in the decrease of protein production of recoded genes [2 , 3] . Because codon pairs , which contain CpG and TpA dinucleotides at the codon pair boundary , are among the most underrepresented codon pairs in eukaryotes , CPBD also inadvertently and markedly increases the number of CpG and TpA dinucleotides in codon pair-deoptimized genes [10 , 11] . An alternative theory proposes that the inadvertent change of dinucleotide frequencies is responsible for viral attenuation by making the recoded viral genes susceptible to recognition by the innate immune response [10 , 11] . A recent study showed that the host zinc-finger antiviral protein ( ZAP ) could be the long-suspected and enigmatic antiviral factor [12] , because it binds to CpG-rich RNA and targets them for degradation by the RNA exosome , thereby inhibiting viral replication [13] . Viruses have evolved effective countermeasures that block the antiviral responses mediated by ZAP [14 , 15] , and/or reduced the level of CpG dinucleotides in their genomes to become undetectable by ZAP . The strategy of attenuation by codon pair deoptimization has been successfully employed in attenuation of a variety of RNA viruses [1–4 , 6–9 , 16 , 17] , but it has never been tested on large double-stranded ( ds ) DNA viruses , such as asfarviruses , poxviruses or herpesviruses . The goal of this study was to determine whether a large ds DNA virus , Gallid alphaherpesvirus 2 ( GaHV-2 ) , also known as Marek’s disease virus ( MDV ) , can also be efficiently attenuated by CPBD . MDV is the causative agent of Marek’s disease ( MD ) , a highly contagious lymphoproliferative and immunosuppressive disease of chickens . Under field conditions , most chickens are infected with MDV within first days of life and the virus causes up to 100% mortality in unvaccinated hosts . Live attenuated vaccines play a crucial role in controlling MD . All chickens that are commercially raised are vaccinated in-ovo or immediately after hatching to prevent MD . While vaccination prevents or reduces MD symptoms , it does not prevent super-infection with virulent MDV strains , and permissive vaccines in turn might lead to the selection of more virulent viruses . [18–20] . Before chickens became farmed in large complexes , MDV strains of mild virulence were causing only minor problems . Intensive farming practices provided MDV with the opportunity to infect large and naïve populations of chickens and evolve towards greater virulence [19 , 21 , 22] . During the last few decades , increasingly virulent MDV strains emerged and caused rampant disease and high mortality in flocks vaccinated with the best MD vaccines available [21 , 23 , 24] . Thus , despite 40 years of vaccine development , MD still jeopardizes poultry and egg production on a world scale . One of the most important goal of current MD research , therefore , is the development of a vaccine superior to the gold standard , CVI988-Rispens [22] . To determine whether CPBD is a suitable approach for MDV attenuation , we recoded UL30 that encodes the catalytic subunit of the viral DNA polymerase ( Pol ) . The UL30 ORF was codon pair- optimized , -randomized , or -deoptimized . Corresponding to the level of codon pair deoptimization , the mutant viruses had either a lethal phenotype , or were severely attenuated in vitro and in vivo . Nevertheless , the oncogenic potential of the virus was not completely eliminated . In contrast , virus with codon pair-optimized UL30 formed bigger plaques than the parental virus in vitro , but was not more pathogenic than the parental virus in vivo . The results of this study imply that CPBD might be an applicable strategy for attenuation of other herpesviruses , and other large ds DNA viruses .
Codon pair bias has been found in every species examined and was shown to differ between species [1 , 11] . Because MDV infects domestic chickens , we hypothesized that MDV , in the process of co-evolution , might have adapted to codon pair bias of chicken to achieve optimal translation efficiency of its genes . To determine codon pair bias in chicken protein coding genes , we calculated a codon pair score ( CPS ) for each of the 3 , 721 codon pair combinations ( 61 × 61 codons , excluding the stop codons ) using the method described by Coleman et al . [1] . A positive CPS value means that a given codon pair is found in chicken protein coding DNA sequences more often than it would be expected based on the individual frequencies of codons that form the given codon pair . Similarly , a negative CPS value means that a codon pair is underrepresented in the chicken ORFeome . When we compared the chicken CPS with the human CPS values [1 , 11] , we found out that the CPS of the two species are very similar ( S1 Fig ) . Using the obtained chicken CPS , we calculated an average CPS , also labelled as codon pair bias ( CPB ) score , for each of the 15 , 762 predicted chicken protein coding sequences that we used in the analyses of CPS [11] . A negative CPB means that a coding sequence contains mostly under-represented codon pair combinations . We plotted each chicken gene’s calculated CPB score against its length in codon pairs to visualize the distribution of CPB scores of chicken genes ( Fig 1A ) . The majority of chicken protein coding genes have positive CPB scores , and the mean of all 15 , 763 CPB scores is 0 . 076 , which is almost identical with the mean CPB score of 0 . 075 of all human genes [1] . We used chicken CPS to calculate CPB of all predicted protein coding genes of MDV , and we observed that MDV genes had significantly lower CPB scores ( median CPB -0 . 06 ) than the chicken genes , which suggests that encoding of MDV genes is not considerably influenced the codon pair bias of its host ( Fig 1A ) . We developed a program that uses the calculated chicken CPS and recodes a given protein coding sequence to a sequence with a desired CPB value . The recoding program reshuffles the synonymous codons of a given sequence to maximize ( = optimization ) , minimize ( = deoptimization ) or preserve ( = randomization ) the CPB score of the sequence . Because only synonymous codons are swapped during recoding , the recoded sequence contains the same codon bias and encodes the same protein as the parental sequence . The algorithm also controls the free energy of folding RNA in a narrow range to prevent formation of large secondary structures , for example hairpins , as a consequence of reshuffling . It is important to keep the free energy of recoded genes in a narrow range to ensure that reduction of protein production is caused by codon pair deoptimization and not by extensive secondary RNA structures . We used the developed program to recode 112 predicted protein coding ORFs of MDV to minimize or maximize the CPB score ( S2 Fig ) . The median CPB of parental , codon pair-deoptimized and codon pair-optimized genes was -0 . 06 , -0 . 45 and 0 . 28 , respectively . Our goal was to determine whether CPBD is an applicable strategy for attenuation of herpesviruses . We selected the very virulent RB-1B strain of oncogenic MDV as our model virus , and recoded gene UL30 , which encodes the catalytic subunit of DNA polymerase ( Pol ) , to study the effect of codon pair randomization , optimization and deoptimization on protein production , viral replication and attenuation . We recoded the UL30 gene , because DNA Pol is essential for virus replication in vitro and in vivo , and we expected that reduction or increase of DNA Pol protein production might change biological properties of the mutant viruses , which could be easily observable in cell culture and in vivo . Because it was impossible to predict how strong an effect of codon pair deoptimization of UL30 on viral fitness would be , and because we conjectured that codon pair deoptimization of the entire UL30 gene might lead to a lethal phenotype , we recoded ORF UL30 such that we could produce , if needed , MDV mutants with different levels of codon pair deoptimization , optimization , or randomization . Therefore , we divided the UL30 coding sequence ( 3 , 663 nt ) in silico into three equally long segments ( 1 , 221 nt ) . The segments were recoded separately–producing genes with three codon pair-optimized ( OOO ) , -deoptimized ( DDD ) or -randomized ( RRR ) segments ( Fig 1B ) . Because the coding sequences of the UL30 and the essential UL31 gene overlap at their 3’ ends , we recoded only the first 3 , 459 nucleotides ( 1 , 153 codons ) , and left the last 204 nucleotides of the UL30 ORF intact . As a result , only the first 1 , 017 nucleotides of the third UL30 segment were recoded ( Fig 1B ) . The unaltered sequence contains the overlapping coding sequences ( 77 nucleotides ) and the polyadenylation signal of the UL31 gene . Independent recoding of the UL30 segments enabled us to generate chimeric UL30 genes , in which any of the segments are recoded but the codon bias of the entire ORF is preserved . We constructed UL30 genes with one ( DWW , WDW , and WWD ) , two ( DDW , WDD , and DWD ) , or three ( DDD ) codon pair-deoptimized segments ( Fig 1B ) . The characteristics of the parental and recoded sequences are summarized in Table 1 . As expected , reshuffling of synonymous codons introduced several hundred silent mutations in the recoded sequences . As observed previously [4 , 10 , 11] , codon pair optimization reduced , and codon pair deoptimization markedly increased , the number of CpG dinucleotides in recoded sequences ( Table 1 ) . All sequences contain exactly the same codons , and therefore also same codon bias—which can be characterized by codon adaptation index ( CAI ) [25] , but the order of codons in individual sequences is different . When two fully recoded sequences are compared to each other , on average 55% of all codons , which occupy the corresponding positions , are different . Because most of the synonymous codons share the first two and differ only in the third nucleotide , sequences are more similar to each other on the nucleotide than on the codon level . As expected , the parental , UL30-WWW and the UL30-OOO are the most alike on the nucleotide and codon level ( Table 1 ) . They contain the highest number of identical nucleotides ( 81 . 1% ) and codons ( 51 . 5% ) at the same positions . In contrast , UL30-OOO and UL30-DDD genes contain the least number of identical nucleotides ( 74 . 8% ) and codons ( 34 . 6% ) at the same positions ( Table 1 ) . To evaluate the effect of the recoding on UL30 protein production , we constructed expression plasmids pUL30-EGFP , in which UL30 expression was driven by the immediate-early promoter ( IE ) of Human cytomegalovirus , and UL30 genes were C-terminally tagged with EGFP . We transfected plasmids containing UL30-WWW , -RRR , -OOO , -DWW , -WDW , -WWD , -DDW , -DWD , -WDD and -DDD genes into DF-1 , HEK 293T , HeLa and Vero cells . We obtained similar results in different cells lines ( Fig 2 and S3–S6 Figs ) . While production of EGFP from the codon pair-randomized construct RRR was slightly reduced in comparison to the level of the parental WWW construct , the codon pair-optimized OOO construct produced more , and codon pair-deoptimized constructs produced less EGFP than the parental construct ( Fig 2 ) . In addition , the level of EGFP correlated with the level of codon pair-deoptimization of UL30: the genes with one codon pair-deoptimized segment produced only slightly less EGFP , but EGFP production was markedly reduced in genes that carried two or three codon pair-deoptimized segments . To quantify protein production of the parental and recoded genes , we constructed dual expression plasmids , in which expression of UL30-EGFP fusion gene and TagBFP was driven by two different promoters . The plasmids were transfected into HEK 293T cells and we analyzed mRNA and protein production from the parental and recoded genes 24 h after transfection . We used HEK 293T cells because they are highly transfectable , and because codon pair bias of human and chicken is highly similar ( S1 Fig ) . qPCR analysis showed that codon pair-optimized gene produced slightly more , and codon pair-randomized and -deoptimized variants produced less mRNA than the parental UL30 gene , yet , the differences were not significant ( Fig 3A ) . We used the ratio of the EGFP to TagBFP fluorescence as a measure of protein production from different UL30 variants . Protein production from the UL30-OOO-EGFP was increased , and protein production from the codon pair-deoptimized genes was significantly reduced in comparison to the UL30-WWW-EGFP per amount of TagBFP ( Fig 3B ) . These results confirmed that codon pair optimization had a positive effect , and codon pair deoptimization had negative effect on protein production of the recoded genes . To further investigate whether changes in UL30 RNA levels could be responsible for the observed differences in UL30 protein levels , we determined RNA expression from the parental and recoded UL30 genes during virus replication . Because MDV is strictly cell associated in vitro and in vivo , it is impossible to achieve synchronous infection of permissive cells by infection . To circumvent this inherent problem , we transfected viral BAC DNA into CEC and quantified the levels of viral UL29 , UL30 , and UL42 mRNA by qPCR 24 h post-transfection . We used the ratios of UL30 to UL29 , and UL30 to UL42 to determine if recoding of UL30 affected RNA expression levels . We compared the expression levels of the UL30 to the levels of UL29 and UL42 because all three genes belong to the same kinetic gene expression class and represent early ( β ) genes . Furthermore , UL42 encodes the processivity factor of DNA polymerase and forms with UL30 the functional DNA polymerase . Quantification of RNA expression showed that codon pair deoptimization had a negative effect on RNA expression during virus replication , but only the UL30-DDD mutant produced significantly less UL30 RNA than the parental virus ( Fig 4A ) . We constructed MDV viruses in which the parental UL30 was replaced with the recoded genes . The mutants were made by 2-step Red en passant mutagenesis of pRB-1B , the infectious bacterial artificial chromosome ( BAC ) clone of MDV RB-1B [26] . In addition , from each mutant clone , we constructed a revertant by restoring the parental sequence of the UL30 region . The BAC mutants were examined by restriction fragment length polymorphism ( RFLP ) analysis , and sequencing of the mutated UL29-UL31 region . Initially , we constructed BAC mutants with fully codon pair-randomized , -optimized and -deoptimized genes ( pUL30-OOO , pUL30-RRR and pUL30-DDD ) . The BACs were transfected into CEC and we recovered parental vWWW , vRRR and vOOO viruses , but we failed to recover infectious progeny of the vDDD mutant despite five independent transfections and five successive blind passages of independent pUL30-DDD BAC clones . Because the vDDD revertant construct could be recovered , we concluded that the vDDD had a lethal phenotype because the level of codon pair deoptimization was too high . We then generated additional mutants in which only one ( pUL30-DWW , pUL30-WDW and pUL30-WWD ) or two segments ( pUL30-DDW and pUL30-WDD ) of the UL30 were codon pair- deoptimized ( Fig 1B ) . Mutant viruses were reconstituted by transfection of CEC , and we noted that the efficiency of reconstitution was variable among mutants . While 2 to 3 passages were necessary to observe a cytopathic effect ( CPE ) of the codon pair-deoptimized mutants , the CPE of the parental , vOOO , vRRR and revertant viruses could be readily observed immediately after transfection or in passage 1 after transfection . Replication properties of mutant viruses in CEC were assessed by plaque size assays and multistep growth kinetics . As expected , viruses with two codon pair-deoptimized UL30 segments formed significantly smaller plaques than the parental vWWW virus . In contrast , mutants with only one segment deoptimized ( vDWW , vWDW , vWWD ) , the randomized vRRR mutant and all revertant viruses formed plaques with sizes that were virtually identical to those of vWWW ( Fig 4B ) . Unexpectedly , vOOO virus replicated more efficiently than the parental vWWW , as the plaques were bigger than those of the parental virus ( Fig 4B ) . The plaques formed by vDDW , vDWD and vWDD were not only smaller than those of the parental virus , but also differed in morphology ( Fig 4C , S7 Fig ) . Plaques that were formed by the codon pair-deoptimized viruses had fewer infected cells and were interspersed by many uninfected cells . Infected cells in these plaques also stained less efficiently than infected cells of the parental virus . Next , we determined the replication of vWWW , vRRR , vOOO , vDWW , and vDDW viruses in CEC by multi-step growth kinetics ( Fig 5 ) . All mutant viruses , with the exception of the vDDW ( Fig 5A ) and revertant viruses ( Fig 5B ) , replicated with kinetics that were comparable to those of the parental virus . The vDDW grew to significantly lower titers than the parental virus ( Fig 5A ) . To test the genetic stability of recombinant viruses , we passaged the parental and mutant viruses sequentially in CEC at low multiplicity of infection . After 20 passages , we sequenced the UL30 region and determined plaque sizes of the passaged viruses . We did not detect any mutation in the recoded region , nor changes of the virus phenotype ( S8A Fig ) or virus replication properties ( S8B Fig ) , confirming that recoded viruses were genetically stable . Because recoding influenced protein production of the UL30 genes , and because the vDDW replicated less efficiently in CEC than the parental virus , we hypothesized that recoded genes may also influence virus replication and tumorigenesis in vivo . To determine the pathogenic potential of the mutant viruses , we infected 1-day-old specific-pathogen-free chickens with parental vWWW , three mutants ( vOOO , vDWW , vDDW ) , and the corresponding revertant ( vOOO-Rev , vDWW-Rev or vDDW-Rev ) viruses . To determine if recoding affected replication of the viruses in chickens we monitored levels of viral DNA in the peripheral blood by qPCR until 28 days post infection ( p . i . ) . Analysis of the blood samples showed that birds infected with different viruses had similar viral loads at different times after infection , indicating that all viruses replicated within the host with similar kinetics ( Fig 6A ) . However , animals that were infected with the vDDW viruses showed delayed progression of MD and lower mortality rates , albeit the differences were not statistically significant ( Fig 7A ) . At final necropsy , animals infected with vDWW and vDDW had fewer tumors than animals infected with parental virus vWWW ( Fig 7B ) . In contrast to the parental and revertant viruses , contact birds that were housed together with chickens infected with vDWW and vDDW did not develop any tumors ( Fig 7C ) . We concluded from the results that MDV viruses with codon pair-deoptimized UL30 , despite being severely attenuated in vitro , are still pathogenic for highly susceptible chickens , because codon pair deoptimization reduced the incidence of visceral tumors , but did not completely abrogate virus-induced tumorigenesis .
In the work presented here , we studied if codon pair deoptimization of an essential MDV gene can result in attenuation of the virus in its natural host . Herpesviruses , as many other large DNA viruses , are equipped with their own DNA replicase , which renders them independent from the host DNA replication machinery . The catalytic subunit of the herpesvirus DNA polymerase is a pivotal enzyme responsible for genome replication , and , thus , for successful transmission of genetic information from one generation to the next . As the DNA polymerase is essential for virus replication , we expected that alteration of DNA polymerase levels should result in phenotypic changes that would be easily observable in cell culture , for example in the form of reduced virus spread , and in vivo . The exact molecular mechanisms that are responsible for attenuation by codon pair deoptimization remain unknown . However , it has been hypothesized that reduced translatability of codon pair-deoptimized genes might be the key factor responsible for attenuation of recoded viruses [1 , 2] . To test this hypothesis , we examined several MDV mutants that carried codon pair-randomized , -optimized and -deoptimized UL30 variants . In line with our expectations , codon pair optimization or random reshuffling of codons in UL30 did not negatively affect the fitness of recoded viruses . In contrast , a virus that carried a fully codon pair-deoptimized UL30 displayed a lethal phenotype in cell culture . To identify why a fully codon pair-deoptimized virus was unable to replicate in cell culture , we constructed additional viruses with reduced levels of codon pair deoptimization . We constructed chimeric viruses in which approximately one-third ( vDWW , vWDW , vWWD ) , or two-thirds ( vDDW , vWDD ) of UL30 were codon pair-deoptimized . Viruses of each of these two groups displayed similar phenotypes in vitro: while viruses with one codon pair-deoptimized UL30 segment replicated with parent virus-like kinetics , viruses in which two-thirds of the ORF was codon pair-deoptimized formed smaller plaques and replicated with significantly reduced kinetics when compared to the parent virus . These additional mutants also showed that none of the recoded segments itself , through potential negative effects on neighboring gene expression , was the cause of the observed lethal phenotype . The three types of codon pair-deoptimized mutants displayed a full spectrum of potential virus attenuation in vitro: viruses with one codon pair-deoptimized UL30 segment were very similar to the parental virus , the replication capacity of viruses with two codon pair-deoptimized segments was severely impaired , and a fully codon pair-deoptimized virus produced no viable progeny after transfection in susceptible cells . The results imply that the level of codon pair deoptimization correlated with the capacity of virus to replicate in vitro . To better understand the effect of recoding on UL30 expression , we quantified RNA and protein production from the recoded UL30 genes in cells transiently transfected with expression plasmids but also during lytic infection of permissive cells in the virus background . The experiments showed that codon pair deoptimization had negative effect on RNA levels of UL30 after transient expression but also during virus replication ( Figs 3A , 3B and 4A ) . In addition , we found a correlation between protein levels determined by flow cytometry and mRNA levels determined by qPCR . These results are in agreement with results from previous studies , which showed that codon pair deoptimization can lead to decreased mRNA levels [3 , 4] . It is speculated that the reduction of mRNA levels could be caused by decreased expression of recoded genes , or increased mRNA degradation caused by the suboptimal , or stalled translation [4] . Still , because the differences in RNA levels measured in transient transfection experiments were not statistically significant ( Fig 3A and 3B ) , and only the MDV mutant with the UL30-DDD gene produced significantly less UL30 RNA when compared to the parental gene ( Fig 4A ) , we consider it unlikely that altered mRNA levels alone are responsible for the observed reduction in protein quantities and replication of viral mutants . The recoded viruses were tested in vivo to determine if attenuation by codon pair deoptimization is a viable strategy for attenuation of MDV . In addition to the parental virus , we tested a codon pair- optimized , and two codon pair-deoptimized viruses ( vOOO , vDWW and vDDW , respectively , Fig 7 ) . From the tested viruses , only the virus with two codon pair-deoptimized segments , vDDW , exhibited significantly reduced replication kinetics in vitro when compared to the parental virus . Surprisingly , qPCR showed that all tested mutant viruses replicated in the natural host , and the level of replication was similar among different mutants ( Fig 6A ) . As we predicted based on our in vitro data , animals that were infected with the recoded virus vDDW developed fewer tumors , and showed delayed progression of MD in comparison with the parental virus; yet , surprisingly , the overall mortality at 90 days p . i . was similar to chickens infected with the parental wild type virus . Despite the clear differences in MD progression and tumor formation , the codon pair deoptimization of RB-1B UL30 alone was not sufficient to render the virus fully attenuated in vivo . Our study showed that codon pair deoptimization did not restrict bird-to-bird transmission of recoded MDV . The recoded viruses were shed from the feather follicles , as evidenced by the detection of replication of recoded viruses in contact birds by qPCR ( Fig 6B ) . Sentinel birds that were housed with vDWW- and vDDW-infected chickens did not develop tumors . Yet , we suspect that tumor formation was only delayed in these animals , because vDWW and vDDW viruses replicated efficiently in contact chickens ( Fig 6B ) . Until now , the effect of codon pair deoptimization has been studied only in RNA viruses , which have relatively small genomes and a relatively small number of protein coding genes . As a result , codon pair deoptimization even of a single gene results in recoding of a relatively large proportion of total coding capacity of such viruses [1 , 2 , 4 , 6 , 8] . Previous studies with poliovirus [1] , Influenza A virus [2] and Dengue virus [8] showed that recoding of multiple viral genes has a cumulative effect on virus attenuation . These studies also showed that the level of virus attenuation is not a function of the extent of viral genome deoptimization . Different genes contribute unequally to virus attenuation , and the contribution of each gene must be evaluated empirically . Consequently , it might be possible that codon pair deoptimization of a different essential MDV gene , or a combination of several genes , might result in satisfactory attenuation of MDV in vivo . An alternative speculation , namely that a codon pair deoptimization is an unsuitable method of attenuation for MDV , could be drawn based on the observation that a virus , which was severely impaired for replication in vitro ( vDDW , Fig 5 ) , still retained a high level of virulence in vivo ( Fig 7 ) . Because MDV establishes latency in T-cells as early as 7 days p . i . and that the latently infected T-cells become later neoplastically transformed , it allows us to speculate that no satisfactory level of virus weakening can be achieved by existing attenuation approaches–at least none that would abrogate the ability of MDV to transform infected cells and ultimately cause tumors . This would mean that a successful attenuation of MDV may be achieved , for example , by impairment of MDV replication by codon pair deoptimization and a deletion of the principal MDV oncogene meq . However , because we managed to drastically reduce protein production of an essential gene without killing the virus , we expect that codon pair deoptimization is a suitable strategy for attenuation of large DNA viruses , preferably those that do not result in neoplastic transformation and for which pathogenesis relies mostly , if not exclusively , on lytic replication . Only a limited number of studies explored the effect of codon pair optimization on viral properties [1] . Transient transfections showed that utilization of overrepresented codon pairs boosted protein production from the recoded , codon pair-optimized gene ( Figs 2 and 3B ) . Interestingly , the virus carrying such a gene formed bigger plaques than the parental virus ( Fig 4B and 4C ) , but replicated with parental-like kinetics in vitro ( Fig 5 ) , and was not more virulent than the parental virus in vivo ( Fig 7 ) . Thus , similar to codon pair-optimized poliovirus [1] , recoding did not result in a virus that would replicate better than the parent in cell culture or in vivo . However , because recoding can lead to increased protein production , it remains to be determined if codon pair optimization , or other forms of protein recoding , could result in a virus that would be more virulent or pathogenic than the wild type .
All animal experimentation was done in full accordance with the EU legislation for the use of animals for scientific purposes ( Directive 2010/63/EU ) and German law ( paragraph 8 Tierschutzgesetz ) . Animal housing , welfare and experimentation are under constant monitoring from an independent governmental institution . Animal experiments were approved by the Landesamt für Gesundheit und Soziales in Berlin , Germany ( approval G0218-12 ) . Fertile , specific-pathogen-free chicken eggs ( Lohmann Tierzucht , Germany ) were incubated in house , and 10-day-old embryos were used for production of primary chicken embryo cells ( CEC ) . We used 15 , 762 predicted chicken protein coding genes ( Gallus gallus , breed Red Jungle fowl , line UCD001 , version 4 . 0 ) to quantify the level of underrepresentation/overrepresentation of each of the 3 , 721 possible codon pairs ( 61 × 61 sense codons ) in the chicken ORFeome by calculating their codon pair scores ( CPS ) [1] . CPS is defined as the natural log of the ratio of the observed to the expected number of occurrences of a particular codon pair , and overrepresented codon pairs have positive CPS [1] . Using the calculated CPS we then calculated average CPS ( CPB scores ) for each of the 15 , 762 chicken and 112 MDV genes ( S2 Fig ) . We used the calculated chicken CPS to develop a computer program that can recode a given protein coding sequence to a new sequence with the desired CPB value . The program can reshuffle the available codons of a given sequence in order to preserve ( codon pair randomization ) , minimize ( codon pair deoptimization ) , or maximize ( codon pair optimization ) CPB of a given sequence . The recoding preserves codon bias , amino acid sequence , and the folding free energy of the recoded sequence . Since there are many possibilities how a certain protein can be encoded , recoding of a gene to the maximal level of codon pair optimization/deoptimization is computationally difficult and time-consuming . An approximate , near-optimal solution of this problem , which is more than sufficient for our purpose , can be found quickly by heuristic or metaheuristic approaches . Our recoding program , similar to the algorithm of Coleman et al . [1] , was designed to utilize simulated annealing , a fast metaheuristic algorithm , to locate a good approximation of the absolute CPB extreme [27] . Our algorithm also controls the free energy of folding RNA in a narrow range to prevent formation of extensive secondary structures ( S9 Fig ) . It is important to keep the free energy of recoded genes in a narrow range in order to ensure that reduction of protein expression is caused by codon pair deoptimization and not by extensive secondary RNA structures . To ensure absence of extensive secondary structures in encoded RNA , we scanned the recoded sequences using the mFold program [28] exactly as described [1] . Briefly , from the coding sequences , we generated an array of overlapping fragments , which were 100 nucleotides long and had an 80 nucleotide overlap with each other . Then , we calculated the folding free energy for the produced fragments , and , when necessary , recoded all fragments that had free energy lower than -30 Kcal/mol by additional codon reshuffling to elevate the free energy of those particular regions . The final recoded sequences have a similar distribution of free folding energy over the length of the UL30 sequence and also a similar mean folding free energy ( S9 Fig ) . The recoded sequences were synthesized ( BioBasic , Canada ) and cloned in the pUC57 vector ( pUL30-RRR , pUL30-OOO or pUL30-DDD ) . CAI of the parental and recoded genes was calculated based on the codon composition ( codon bias ) that is present in 15 , 762 predicted chicken protein coding genes [25] . Primary chicken embryo cells ( CEC ) were prepared from 10-day-old specific-pathogen-free embryos [29] and were cultured in minimal essential medium ( MEM ) with Earle’s salts , 100 U/ml penicillin , 100 μg/ml streptomycin and 1–10% fetal bovine serum ( FBS ) . Human embryonic kidney 293T ( ATCC CRL-3216 ) , chicken embryo DF-1 ( ATCC CRL-12203 ) , human epithelial carcinoma HeLa ( ATCC CCL-2 ) and African green monkey kidney Vero ( ATCC CCL-81 ) cells were grown Dulbecco’s modified MEM ( DMEM ) with Earle's salts , 100 U/ml penicillin , 100 μg/ml streptomycin and 10% FBS . To construct the pUL30-EGFP plasmids , the WT and recoded UL30 ORFs were cloned into pEGFP-N1 ( Clonetech ) between the NheI and AgeI restriction sites in frame with EGFP , so the DNA Pol is produced as fusion protein to the N-terminus of EGFP . To construct the pmKate2 plasmid the EGFP gene in pEGFP-N1 was replaced with mKate2 gene ( Evrogen ) encoding a far-red fluorescent protein . To assess protein production from the parental WWW and recoded UL30 by flow cytometry we constructed pVITRO2-TagBFP-UL30-EGFP plasmids . The UL30 ORFs were fused C-terminally with EGFP and cloned into a dual expression plasmid pVITRO2-MCS ( InvivoGen ) under the control of human ferritin H/mouse elongation factor 1 promoter . The gene of blue fluorescent protein mTagBFP ( Evrogen ) was cloned into the second MCS , under the control of human ferritin L/chimpanzee elongation factor 1 promoter and was used for normalization of transfection efficiency . The UL30 mutant viruses were generated based on the pRB-1B , an infectious BAC clone of the highly oncogenic MDV strain RB-1B [30] , by two-step Red-mediated en passant mutagenesis in Escherichia coli as previously described [26] . Normally , during en passant mutagenesis , the target gene is replaced with the gene of interest in two steps . Here , the pRB-1B UL30 BAC mutants were generated in three steps . First the parental UL30 ORF was replaced with an ampicillin resistance marker , which was replaced in the second step with a recoded UL30 gene and a kanamycin selection marker . In the final step the kanamycin selection marker was removed by homologous recombination . The initial deletion of the parental UL30 was necessary to prevent undesired recombination between the recoded and parental variants , which are ~80% similar on nucleotide level . Primers used for the construction of the mutant and revertant BAC clones are shown in Supplementary data ( S1 Table ) . The MDV BAC clones were analyzed by RFLP analysis , PCR and DNA sequencing of the target region . To recover infectious viruses the purified BAC DNA and a Cre recombinase expression vector ( pCAGGS-NLS/Cre ) were co-transfected into CEC with polyethyleneimine [31] . Briefly , 1 μg of BAC DNA was diluted in 100 μl of serum-free MEM , mixed with 10 μl of 2 mg/ml linear polyethylenimine ( MW = 25 , 000 ) , incubated at room temperature for 30 min , and added to CEC grown to 80% confluency in a well of a 6-well plate . After 4 h of incubation at 37°C , the transfection mixture was removed and replaced with fresh medium . The expression of Cre recombinase ensured that BAC cassette flanked by loxP sites was efficiently removed from the infectious clones . The removal of the BAC vector from the MDV genome was confirmed by PCR as described previously [32] . Viruses were titrated on fresh CEC and viral stocks were stored in liquid nitrogen . Viral DNA was isolated from the infected cells with the method of Hirt [33] . MDV was passaged serially in 100 mm dishes . In each new passage 5 × 106 freshly seeded CEC were infected with 1 , 000 PFU of MDV , which corresponds to a multiplicity of infection of 0 . 0002 . Multi-step growth kinetics were conducted as described previously [34] . Briefly , CEC ( 1×106 ) seeded in a well of a 6-well plate were infected with 100 PFU of each virus . All infections were performed in duplicates . Cells were trypsinized 1 , 2 , 3 , 4 , 5 and 6 days post infection and 10-fold serial dilutions were inoculated onto fresh CEC . Plaques were stained ( see below ) and counted six days post infection . To analyze cell-to-cell spread of viruses plaque areas were determined as described [35] . Briefly , CEC ( 1×106 ) were mixed with infected CEC containing 50 PFU and seeded into a well of a 6-well plate . After 6 days , plaques were visualized by indirect immunofluorescence ( see below ) . For each virus , images of 90 randomly selected plaques were taken at 100-fold magnification . Plaque areas were measured using ImageJ software ( http://rsbweb . nih . gov/ ) , from which , assuming that ideal plaques would have a circular shape , plaque diameters were calculated . Plaque diameters were plotted and analyzed using GraphPad Prism 7 . 02 . Plaque sizes were determined in three independent biological experiments in double-blind fashion . Expression plasmids pVITRO2-TagBFP-UL30-EGFP were transfected into subconfluent HEK 293T cells grown in 6-well plates with Lipofectamine 2000 ( Thermo Fisher Scientific ) in duplicates . Total RNA was isolated 24 h post transfection with RNAeasy kit ( Qiagen ) . DNA was removed with DNaseI for 30 min at 37°C and 1 μg of RNA was reverse transcribed into cDNA with MMLV reverse transcriptase and random hexamers ( Thermo Fisher Scientific ) . The cDNA was quantified by qPCR using the SYBR Green I system in an AB StepOnePlus Real-time PCR system ( Thermo Fisher Scientific ) . The copy numbers were determined based on calibration curves generated for both genes based on known concentrations of pVITRO2-TagBFP-UL30-EGFP plasmid which was used as a standard . To assess transcription of recoded UL30 genes in the viral background 1 μg of BAC DNA was transfected in duplicates into CEC grown in 6-well plates using polyethylenimine as described above . To synchronize infection of permissive cells the transfection mixture was incubated with cells only for 30 min . RNA was isolated 24 h post transfection as described above . The gene-specific primers and probes were used to quantify the viral transcripts of UL29 , UL30 and UL42 genes ( S1 Table ) using the Luna Universal One-Step RT-qPCR Kit ( NEB ) . The copy numbers were determined from the calibration curves generated for all three genes based on known concentrations of pRB-1B . Infected cells were fixed with 2% paraformaldehyde , permeabilized with 0 . 1% Triton-X 100 , and blocked with 3% bovine serum albumen ( BSA ) in PBS . The cells were then incubated with anti-MDV chicken serum ( dilution 1:2000 ) for 1 h , and then with goat anti-chicken IgG-Alexa Fluor 488 secondary antibody ( dilution 1:2000 , Invitrogen ) for 45 min to visualize plaques . Images were taken at 100- and 200-fold magnification using an inverted fluorescence microscope ( Axiovert S100 , Zeiss ) . To analyze protein production from recoded UL30 variants ( WWW , RRR , OOO , DWW , WDW , WWD , DDW , DWD , WDD and DDD ) DF-1 , HEK 293T , HeLa or Vero cells were contransfected with 500 ng of pUL30-EGFP and 500 ng of pmKate2 plasmid . As a control pmKate2 and pEGFP-N1 were transfected alone . Images were taken 24 h post transfection at 200- and 400-fold magnification using an inverted fluorescence microscope ( Axiovert S100 , Zeiss ) . Protein production was quantified by using a dual expression pVITRO2-TagBFP-UL30-EGFP plasmids . Plasmids were transfected into HEK 293T cells . 24 h post transfection cells were washed twice with PBS and resuspended in FACS buffer . Fluorescence of TagBFP and EGFP were measured in a CytoFlex flow cytometer ( Beckman Coulter ) . Color compensation was done using a TagBFP and EGFP positive control to eliminate artifact due to the overlap of TagBFP and EGFP emission . Cell debris and duplets were excluded from further analyses . One-day-old specific-pathogen-free Valo chickens ( Lohmann Tierzucht , Germany ) were infected intraperitoneally with 2 , 000 PFU of vWWW , vOOO , vDWW , vDDW , vOOO-Rev , vDWW-Rev or vDDW-Rev . Each experimental group was composed of sixteen infected chickens and eleven naïve chickens . The uninfected chickens were housed with the infected chickens to determine if mutant viruses were able to transmit via natural route by shedding . Chickens that were infected with three revertant viruses were housed together in one room , and therefore shared only one group of contact birds . Water and food were provided ad libitum . Animals were monitored for MD symptoms on a daily basis . After manifestation of clinical symptoms or termination of the experiment at 90 days after infection , animals were examined for tumors by necropsy . Whole blood was taken by wing vein puncture from 8 randomly selected chickens of each group . Infected and contact chickens were sampled 4 , 7 , 10 , 14 , 21 and 28 days p . i . and 21 , 28 , 35 and 42 days p . i . respectively . Blood was taken from the same animals , and dead animals were not replaced . DNA was isolated out of the chicken blood using an E-Z96 96-well blood DNA isolation kit ( Omega Biotek ) . The MDV genome copies in the chicken blood were quantified by qPCR using primers and probe for MDV gene ICP4 [35] . Number of cellular genome copies of the inducible nitric oxide synthase ( iNOS ) was used for normalization [35] . The recoded UL30 sequences are available under GenBank accession numbers MF671698 ( UL30-RRR ) , MF671699 ( UL30-OOO ) and MF671700 ( UL30-DDD ) , and also in the Supplementary data ( S1 Appendix ) .
|
Codon pair bias deoptimization ( CPBD ) enables highly efficient attenuation of viruses . In contrast to other methods , live attenuated virus vaccine candidates can be rationally designed and produced within days . The technique involves recoding of viral genes , while preserving their codon bias and amino acid sequence . Recoding increases the number of codon pairs that are statistically underrepresented in protein coding sequences of the viral host , and involves swapping of available synonymous codons . While CPBD has been used to attenuate RNA viruses , it has never been applied on large double-stranded DNA viruses , such as poxviruses , asfarviruses , or herpesviruses . We used CPBD to attenuate an oncogenic Marek’s disease herpesvirus . The mutant viruses contained a recoded UL30 gene , which encodes DNA polymerase . The UL30 was either codon pair-optimized , -randomized , or -deoptimized . Corresponding to the level of codon pair deoptimization , the mutant viruses had either a lethal phenotype or were severely attenuated in vitro and in vivo . Nonetheless , viral oncogenicity was not completely eliminated . Virus with codon pair-optimized UL30 had characteristics of the parental virus in vitro and in vivo . The results of our study imply that CPBD might be an applicable strategy for attenuation of other herpesviruses and potentially other large double-stranded DNA viruses .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"transfection",
"sequencing",
"techniques",
"dengue",
"virus",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"bac",
"cloning",
"pathogens",
"microbiology",
"vertebrates",
"cloning",
"animals",
"vector",
"cloning",
"viruses",
"rna",
"viruses",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"birds",
"medical",
"microbiology",
"microbial",
"pathogens",
"gamefowl",
"viral",
"replication",
"fowl",
"molecular",
"biology",
"poultry",
"nucleotide",
"sequencing",
"eukaryota",
"flaviviruses",
"virology",
"viral",
"pathogens",
"biology",
"and",
"life",
"sciences",
"chickens",
"amniotes",
"organisms"
] |
2018
|
Attenuation of a very virulent Marek's disease herpesvirus (MDV) by codon pair bias deoptimization
|
Little is known about the natural history of dengue in Papua New Guinea ( PNG ) . We assessed dengue virus ( DENV ) -specific neutralizing antibody profiles in serum samples collected from northern and southern coastal areas and the highland region of New Guinea between 1959 and 1963 . Neutralizing antibodies were demonstrated in sera from the northern coast of New Guinea: from Sabron in Dutch New Guinea ( now known as West Papua ) and from four villages in East Sepik in what is now PNG . Previous monotypic infection with DENV-1 , DENV-2 , and DENV-4 was identified , with a predominance of anti-DENV-2 neutralizing antibody . The majority of positive sera demonstrated evidence of multiple previous DENV infections and neutralizing activity against all four serotypes was detected , with anti-DENV-2 responses being most frequent and of greatest magnitude . No evidence of previous DENV infection was identified in the Asmat villages of the southern coast and a single anti-DENV-positive sample was identified in the Eastern Highlands of PNG . These findings indicate that multiple DENV serotypes circulated along the northern coast of New Guinea at different times in the decades prior to 1963 and support the notion that dengue has been a significant yet neglected tropical infection in PNG for many decades .
Dengue is a mosquito-borne disease of humans caused by infection with the dengue viruses ( DENV ) . An estimated 390 million infections occur each year in tropical and subtropical countries of which about 98 million are symptomatic; in endemic countries dengue is associated with significant morbidity and mortality . The frequency , magnitude , and geographic range of dengue epidemics began to increase dramatically after the Second World War when major demographic and ecological changes resulted in increased transmission of the dengue viruses and the disease is now endemic in more than 100 tropical and subtropical countries [1–3] . Incidence has increased consistently in Southeast Asia and the Western Pacific in the last decade , and more than 70% of the global dengue disease burden is currently borne by people who live in this region [4] . Dengue is caused by infection with any one of four dengue viruses ( DENV ) , RNA viruses which form their own antigenic complex within the Flaviviridae family [5] . The four serotypes , DENV-1 –DENV-4 , are further classified into genetically distinct groups or genotypes with sequence divergence of up to 6% [6] . Infection is believed to confer lifelong immunity to the homologous DENV serotype and short-lived cross-protective immunity to the other three serotypes [7] , and people who live in endemic areas may be infected up to four times in their lifetime . Symptomatic DENV infection typically presents as a non-specific acute febrile illness ( dengue fever , DF ) which usually develops 3–8 days following a bite from an infected Aedes mosquito , most commonly Aedes aegypti or Aedes albopictus . Severe dengue ( previously known as dengue hemorrhagic fever , DHF , and dengue shock syndrome , DSS ) is an acute vascular permeability syndrome which occurs in a subset of patients with DF; mortality rates in severe dengue range between 2 . 5%-20% [8] . The risk of developing severe dengue is greatly increased in a secondary infection with heterologous DENV [9 , 10] . Dengue transmission has been documented in the Asia-Pacific region since the 1940s when the four prototype DENV were identified in Japan ( DENV-1Mochizuki , 1943 ) , and Hawaii ( DENV-1Hawaii , 1944 ) , New Guinea ( DENV-2NGC , 1944 ) , and the Philippines ( DENV-3H-87 and DENV-4H-241 , 1956 ) [11–13] . DENV-2NGC was isolated from US soldiers who fought in New Guinea in 1942–1944 [14] . US Military records indicate there were a large number of suspected infections in New Guinea—approximately 27 , 000 cases—and several epidemic foci [15] . No severe disease was recorded , however dengue was considered to be a major cause of loss of troop strength . In the years following the Second World War dengue transmission was identified in serosurveys conducted on the island of New Britain in 1971 [16] and in the Bismarck Archipelago in 1975 [17] . More recently an analysis of patients presenting with acute febrile illness in Madang Province on the northern coast of PNG in 2007–2008 identified DENV infection ( measured by IgG , IgM , and NS1 antigen ) in 8% of cases [18]; no severe dengue was identified in these patients . Although dengue surveillance is not conducted in PNG , genetic analysis of serum samples collected from febrile travelers entering northern Australia from PNG between 1999 and 2010 has identified importation of DENV-1 , DENV-2 , and DENV-3 [19 , 20] , indicating that multiple serotypes have circulated in recent times . Dengue epidemics occur frequently in the Asia-Pacific region and DHF/DSS-associated epidemics have been recorded in many countries in this region since the 1950s [4 , 21 , 22] . A recent analysis of dengue surveillance data collected between 1968 and 2013 in Indonesia , a country which experiences frequent , large-scale dengue epidemics and shares geographic borders with PNG , found that the incidence of DHF has increased substantially over the past 45 years [23] . Severe dengue has not been reported in PNG despite the likely transmission of multiple DENV serotypes and the potential for introduction of DENV circulating in neighbouring countries such as Indonesia , and the reasons for this are not understood . We undertook a retrospective analysis of DENV seroprevalence in PNG using serum samples collected between 1959 and 1963 in order to identify the DENV serotypes which circulated before that time , and their geographic distribution . These findings will clarify the length of time dengue has been present in PNG , inform our understanding of the natural history of dengue in PNG , and provide a basis for future surveillance and control efforts .
Ethics approval for this study was granted by the Medical Research Advisory Committee , Ministry of Health , Government of Papua New Guinea; the Institutional Review Board , Binghamton University ( 54-6-2 ) ; and the Human research Ethics Committee , University of Western Australia ( RA/4/1/4050 ) . Archival human sera were obtained from the National Institutes of Health-National Institute of Neurological Disorders and Stroke ( NIH-NINDS ) Serum Archive of Binghamton University , New York , USA . Samples were collected in New Guinea ( now known as PNG ) from East Sepik ( 1959 ) ; Chimbu ( 1962 ) ; and Eastern Highlands ( 1962 ) including villages in the hinterlands of present-day Morobe Province ( Table 1 ) . Additional samples were collected in Dutch New Guinea ( now known as West Papua ) in 1962–1963 , from the northeast coast Central Highlands and southern coast ( Table 2 ) . The samples were stored at –20°C or -80°C at the NIH and Binghamton University prior to analysis . A serum microneutralization ( MN ) assay was developed , based on approaches described elsewhere [24–26] and used to measure anti-DENV antibodies in the archival serum samples . This approach was selected to allow simultaneous assessment of antibody to all four reference DENV serotypes ( DENV-1Hawaii-2001; DENV-2NGC; DENV-3H-87 and DENV-4H-241 ) in samples with limited volumes ( <0 . 1 mL ) . Vero cells at a concentration of 5 x 105/mL in in DMEM medium ( DMEM ( Invitrogen ) supplemented with L-glutamine , penicillin and streptomycin ) with 10% heat-inactivated FBS ( DMEM-10 growth medium ) were prepared and 0 . 1 mL was added to all wells of 96-well flat-bottomed tissue culture plates ( BD Falcon ) for a final input of 5 x 104 cells/well , and the plates were incubated in a humidified incubator at 37°C in 5% CO2 for 48 hours until wells were confluent . Test sera were diluted 1:10 in DMEM medium and heat-inactivated at 56°C for 30 minutes , then serially diluted two-fold to a dilution of 1:1280 in 96-well round-bottom plates . An equal volume of titrated virus was added to the prepared sera , for a final dose of 50 TCID50 , and the plates were incubated at 37°C in 5% CO2 for 1 hour . Growth medium was removed from the Vero cell monolayer and replaced with the serum-virus mixture . Sera were tested in duplicate against all four DENV serotypes simultaneously . Plates were incubated in a humidified 37°C incubator in 5% CO2 for 9–10 days during which time cell monolayers were monitored; appearance of the cells was compared to media-only wells on the same plates . At the end of the incubation culture supernatant was removed , cells were fixed in cold PBS:acetone ( 1:1 ) , and virus replication was assessed by fixed-cell ELISA . Absence of infectivity , indicated by OD415/490 values approaching background ( cell-only control well ) levels , indicated the presence of DENV-specific antibodies in the serum sample and thus constituted a positive neutralization reaction . The assay cutoff value for neutralization was established as the mean OD415/490 of serum only ( no virus ) control wells plus 2 standard deviations . Wells with higher values were deemed negative for neutralization and the neutralization titer was the reciprocal of the highest serum dilution with a mean OD below the cutoff value . OD415/490 for cell-only control wells was <0 . 3 and for virus control wells was >1 . 0 . Data were presented as Geometric Mean Titers ( GMT ) . Standard anti-DENV-1-4 sera NIBSC 05/248 ( National Institute for Biological Standards and Control [NIBSC] , Potter’s Bar , Hertfordshire , United Kingdom ) tested against homologous reference DENV consistently produced MN titers of 10–20 ( S1 Table ) and thus the cutoff value for a positive test result was a titer ( reciprocal serum dilution ) of 10 . Testing of these Standard DENV-1-4 sera in a classical 90% plaque reduction assay [27–28] also produced PRNT-90 values of 10 , affirming the sensitivity of the MN assay as utilized in the present study . Cross-neutralization experiments in which Standard DENV-1-4 sera were each tested against heterologous reference DENV consistently produced negative results; each Standard serum was tested against all four DENV reference viruses used in the MN assay ( S1 Table ) . Serum samples from individuals with well-defined monotypic or multitypic DENV infection [27–29] consistently produced monotypic MN titers greater than 10 to the homologous virus and serum samples from individuals with diagnosed other flavivirus infection ( JEV , MVEV , and KUNJV ) were tested against DENV-1-4 and were always negative ( S1 Table ) . There was insufficient archival sample volume to perform MN tests against JEV , MVEV and KUNV . Serum samples from 20 individuals with no history of DENV infection , designated as negative controls , who were anti-DENV seronegative by NS1 ELISA ( Platelia Dengue NS1 Antigen ELISA kit; Bio-Rad , Australia ) and hemagglutination inhibition ( HI ) assay ( PathWest Laboratory Medicine WA ) consistently produced MN titers less than 10 .
A total of 179 serum samples collected between 1959 and 1963 were analyzed . Sera from 108 males and 71 females were collected from villages and hamlets in PNG ( formerly Papua and New Guinea ) ( 98 sera ) between 1959 and 1962 and West Papua ( formerly Netherlands ( Dutch ) New Guinea ) ( 81 sera ) between 1962 and 1963 . Field notes from the time of collection identified three geographic regions in Papua and New Guinea: East Sepik ( 52 sera ) , Chimbu ( 13 sera ) , and Eastern Highlands ( 33 sera ) ; and three geographic regions in Dutch New Guinea ( West Papua ) : Northeast coast ( 25 sera ) , Central Highlands ( 4 sera ) , and Southern Coast ( 52 sera ) . These data are summarized in Tables 1 and 2 . Anti-DENV neutralizing antibodies were detected in 39 of 98 ( 40% ) sera from PNG; the remaining 59 were seronegative for all four DENV serotypes . All seropositive sera , with the exception of one sample , were from villages in East Sepik . The additional positive sample was from Po village in the southern border area of Eastern Highlands with Morobe , where villagers were known to hunt and maintain gardens at lower altitudes . Since we do not know the travel history of the 22-year-old man from whom the sample was obtained we cannot be sure that he acquired his DENV infection in the southern lowlands . These data are summarized in Table 3 . Of the 39 positive samples , a minority ( 6/39 , 15% ) demonstrated a monotypic neutralizing antibody response to DENV-1 ( 4 sera ) , DENV-2 ( 1 sample ) , or DENV-4 ( 1 sample ) . The remaining 33/39 sera ( 85% ) neutralized multiple DENV serotypes . Of these 33 samples , 20 ( 61% ) sera neutralized all four DENV , 10 ( 30% ) neutralized 3 serotypes , and 3 ( 9% ) neutralized 2 serotypes . Anti-DENV-2 was detected in 32/33 samples closely followed by anti-DENV-1 ( 30 sera ) and anti-DENV-4 ( 30 sera ) with anti-DENV-3 positive sera least frequent ( 24 sera ) . Among DENV-seropositive sera anti-DENV-2 titers were generally highest with titers up to 160 and DENV-2 GMT of 24 . 2 was higher than for the other serotypes . In comparison , responses for the other three serotypes demonstrated lower magnitude titers with anti-DENV-1 GMT of 16 . 3; anti-DENV-3 GMT 18 . 9; and anti-DENV-4 MN GMT 17 . 1 . ( Table 4 ) . Among the East Sepik villages seroprevalence rates were highest for Suonambo ( 16/21; 76% ) , Arisili ( 8/11; 73% ) and Wahlen ( 7/13; 54% ) , which between them contributed 79% ( 31/39 ) of positive sera . Sera were available from 2 main areas: the northeastern coast around Hollandia ( now known as Jayapura ) in West Papua , and the southern coast to the west of the PNG-West Papua border along the Arafura Sea , including Omba village in the hinterlands of the Central Highlands . In total 81 sera were tested including 52 from 14 villages on the southern coast , 25 from the village of Sabron on the northeastern coast , and 4 from Omba . These data are summarized in Table 2 . Anti-DENV antibody was detected exclusively in 20 of 25 sera ( 80% ) collected in Sabron on the northeast coast . These sera accounted for 25% ( 20/81 ) of West Papuan samples . Sabron serum samples showed monotypic anti-DENV-2 ( 9/20; 45% ) and anti-DENV-4 ( 1/20; 5% ) responses , with the remaining 10 of 20 ( 50% ) positive samples showing evidence of multiple previous infections: 3 sera neutralized all 4 DENV serotypes , 5 sera neutralized 2 serotypes , and the remaining 2 sera neutralized 3 serotypes . Anti-DENV-2 responses were identified most frequently , with 19 of the 20 positive sera showing some evidence of previous infection with this serotype . Anti-DENV-2 GMT was 16 . 1 , whereas anti-DENV-1 GMT was higher ( 21 . 8 ) despite the lack of monotypic anti-DENV-1 neutralizing capacity . Lower magnitude responses were seen for the 8/20 sera with anti-DENV-4 responses where MN titers were always 10 , regardless of the responses detected to other serotypes . Anti-DENV-3 responses were the least frequent and were detected ( with MN titers of 10 or 20 ) in the 3 serum samples which were positive for all 4 serotypes , suggesting that this anti-DENV-3 response was cross-neutralization by antibody to the other 3 prevalent serotypes . The lack of monotypic anti-DENV-3 supports this contention; it does not exclude the possibility that individuals with previous DENV-3 infection were not sampled . These data are summarized in Table 5 . A summary of the geographic distribution of DENV seroprevalence in Papua New Guinea and West Papua is shown in Fig 1 .
We undertook a retrospective analysis of DENV seroprevalence in what is now Papua New Guinea and West Papua using serum samples collected between 1959 and 1963 , and our findings suggest that multiple DENV serotypes circulated prior to this time period . DENV seroprevalence was largely restricted to the northern Pacific coast of PNG and West Papua and there was no evidence of previous infection in the villages along the southern Casuarine Coast , or in the Highlands . Monotypic infection with DENV-1 , DENV-2 , and DENV-4 was identified in 16 of 59 ( 27 . 1% ) seropositive individuals . Our findings are consistent with the work done by Sabin [12 , 14] during the Second World War which led to isolation of DENV-2NGC from febrile soldiers deployed in the areas around the northern coastal town of Hollandia ( now called Jayapura , in West Papua ) . Monotypic DENV-4 , which was not isolated until 1956 , in the Philippines [13] , is an interesting finding and implies there was more widespread circulation of the dengue viruses in Oceania in the period prior to the early 1960s than previously understood . The majority of seropositive sera ( 43/59; 72 . 9% ) neutralized multiple DENV serotypes , suggesting that most exposed individuals had been infected more than once before the time of sampling . Assessment of prior DENV infection using serum neutralization assays is subject to well-recognized limitations associated with the relatively high degree of homology among the four DENV serotypes . Anti-DENV antibody epitopes may be unique to each serotype , or they may be shared across all four DENV . Primary infection induces cross-neutralizing antibodies to heterologous DENV serotypes which may be detected in the first few weeks following the acute phase of infection [7] however heterotyoic neutralizing antibody responses have been shown to decline over time [30 , 31] . Several months after repeat infection with heterologous DENV the breadth of cross-neutralizing responses has stabilized , and the profile of multi-serotype responses is likely maintained for life . The monotypic anti-DENV responses identified in the present study likely reflect a true single DENV infection , whereas the multitypic responses suggest a history of infections with at least 2 DENV serotypes . Our results suggest transmission of at least three DENV serotypes along the northern coast of New Guinea in the years prior to 1962 . The mountains of the New Guinea Highlands , which reach heights of almost 5000 meters , bisect New Guinea island and separate the northern and southern lowlands and coastal areas . The rugged terrain of the island has contributed to the development of geographically isolated population groups who speak more than 800 distinct languages . Interaction between groups up until the second half of the 20th century was generally limited [32 , 33] , and this may have restricted the introduction and transmission of acute viral diseases which require relatively close human contact . In addition , continuous contact with colonial authorities and other foreign forces did not occur until around the middle of the 20th century for many groups; the first permanent Dutch government post and trade operations were established among the Asmat peoples , sampled in the present study , on the Casuarine Coast in 1953 . At this time inter-group warfare had not ended and the fierce reputation of the Asmat people [34] coupled with the difficult coastal marshland terrain contributed to isolation of the local populations . Adels and Gajdusek [33] assessed measles seroprevalence in this population and identified a virgin soil epidemic in 1961 , supporting our finding that the apparent lack , or rarity , of dengue transmission along the southern coast was likely a consequence of the isolation of the Asmat villages; the lack of sustained contact with the outside world limited introduction of DENV from external sources as it did for measles . There are no jungle-dwelling non-human primates in West Papua or Papua New Guinea to act as a dengue reservoir and to sustain a sylvatic cycle that might occasionally spill over into humans , as has been recorded in neighbouring Malaysia [35–37] . In studies conducted during a DHF epidemic in Central Java in the late 1970s [38] Aedes aegypti and Aedes albopictus , trapped in the rural district of Bantul , were both shown to be competent vectors for DENV-1-4 isolated from DHF patients in Java . There are no published descriptions of Ae . aegypti or Ae . albopictus along the Casuarine Coast during the time of the present study , in 1962; in the earliest reports from New Guinea island Ae . albopictus was first identified on the northern coast in Jayapura in 1962 and Alexishafen in Madang Province in 1972 , and on the southern coast in the Western Province of PNG along the southern Fly River coastal fringe in 1992 [39] . However , serosurveys conducted in the 1970s identified evidence of previous Chikungunya virus transmission along the Casuarine Coast [40] . Chikungunya and dengue are both vectored by Ae . albopictus mosquitoes and these findings suggest that the potential for dengue transmission existed at the time of sampling for the present study but that transmission did not occur . Ae . albopictus is considered to be a relatively inefficient dengue epidemic vector and has a short flight range [41] and these factors may have constrained the introduction of DENV into the geographically isolated people of the Casuarine Coast prior to 1962 . This study has shown that multiple dengue virus serotypes circulated in the northern coastal areas of West Papua and Papua New Guinea prior to 1963 . It is not possible to identify when the viruses were introduced into New Guinea or where they originated; however , a large number of dengue-like illnesses occurred in US soldiers deployed in this area during the Second World War [15] . There are no reports describing similar illness occurring among local populations in this period and we cannot determine whether DENV was imported into northern New Guinea with troop movements , or if multiple serotypes were already circulating at that time and spread rapidly within susceptible troops . Villagers on the northern PNG coast may have interacted with peoples of neighbouring countries or with visitors from other regions more often compared with inhabitants of the Asmat and Casuarine coasts in the south , who were uniformly seronegative in the present study . Despite the large number of dengue cases among US soldiers deployed along the northern PNG coast in the 1940s there are no reports of severe dengue among the soldiers . The lack of dengue surveillance in PNG in the decades since then has meant that incidence and prevalence data are lacking . Awareness of the disease may not be high throughout PNG but it is nevertheless surprising that severe dengue has not yet been identified in the more than six decades since DENV was first isolated from New Guinea , especially since our present findings indicate that residents of the northern New Guinea coast had experienced multiple DENV infections in the years prior to 1963 . Antibody preservation in serum samples stored for many decades is a major limitation of this study . Serum specimens analyzed in the present study were collected in New Guinea between 1959–1963 . The samples were kept cool until they were transferred for storage at -15 deg C [42]; sera were eventually stored at -20 or -80 deg C in the Serum Archive Facility at Binghamton University in New York [43] . Sample integrity was previously demonstrated in studies assessing chloroquine resistance in Plasmodium falciparum in archived sera collected from East Sepik in 1961–1962 [44] . Immunoglobulin deterioration over more than five decades of transport and storage might reduce the sensitivity of our assays , and thus our findings may represent an underestimation of the true prevalence of anti-DENV neutralizing antibodies in the study areas . Although DHF has never been reported in Papua New Guinea it occurs in West Papua , where the first reported DHF epidemic occurred in Jayapura in 1993–1994 on the northern coast [45] , attributed to DENV-1 , DENV-2 , and DENV-3 . Emergence of severe dengue is consistent with the apparent long-term transmission of multiple serotypes in this area identified in the present study , a similar pattern to that seen in other endemic settings . In contrast the first known DHF epidemic on the southern coast of West Papua occurred in Merauke in 2001 [46] , attributed to DENV-3 . Merauke is a town with an estimated population of 78 , 000 located on the southern coast of West Papua bordering PNG to the east and what is now the Asmat Regency to the west . An historical review of hospital records from 1995 to 2000 revealed no evidence of previous epidemic or sporadic DHF and the authors suggested that the outbreak represented the first instance of epidemic dengue virus transmission . There are no data on DENV transmission in the southern coastal areas of PNG east of Merauke , and no reports of DF or severe dengue , although dengue , including severe dengue , has been reported in the neighbouring Torres Strait Islands [19 , 47] which stretch from the southern coast of PNG to the northern coast of Australia . The reasons for the apparent lack of severe dengue in PNG are unclear and require further investigation . The pathogenesis of severe dengue disease is not well understood though DHF incidence is clearly associated with sequential infection with multiple DENV serotypes [9 , 10] , and in-vitro studies in humans have identified immune enhancement as a mechanism for DHF immunopathogenesis [27 , 48 , 49] . In this model DENV genetic heterogeneity resulting in B- and T-cell-specific epitope antigen variation , the sequence of DENV infections , host HLA profiles , and host genetic variation are among the major virus and host factors which influence clinical outcome . A long history of transmission of multiple DENV serotypes as suggested in our present study would predict a higher incidence of severe dengue than has been observed to date . Similar findings have been reported for Haiti [50] where extensive DENV transmission has been recorded in the absence of detectable severe disease . In summary we have identified serological evidence for transmission of dengue along the northern coastline of New Guinea island as measured by neutralizing antibody to multiple DENV serotypes in adults up to 72 years of age , sampled between 1959–1963 . These findings confirm that multiple DENV serotypes circulated in this region at different times in the decades prior to the sampling period and support the notion that dengue has been a significant yet neglected tropical disease in PNG for more than 70 years . Dengue burden needs to be quantified , and molecular epidemiological studies need to be undertaken to assess the origins and distribution of DENV genotypes so as to develop an understanding of disease transmission and pathogenicity in PNG , particularly in the context of emerging dengue vaccines .
|
Dengue is a mosquito-borne disease caused by infection with any of the four dengue virus serotypes ( DENV-1 –DENV-4 ) , which are transmitted in more than 100 tropical and subtropical countries . The current global dengue burden , and dengue mortality , is greatest in the southeast Asian and western Pacific region where more than 70% of people at risk of infection reside . All four DENV serotypes have been reported to circulate in this region and each DENV serotype has been associated with high rates of morbidity and mortality . Sequential infection with heterologous DENV serotypes is associated with more severe dengue disease ( previously known as dengue hemorrhagic fever and dengue shock syndrome ) and co-circulation of multiple DENV serotypes is frequently observed in endemic countries . Substantial variation in local capacity for systematic surveillance and reporting among countries in the region means dengue burden is likely underestimated . We tested archival serum samples collected more than 50 years ago in Papua New Guinea in order to begin to assess the true burden of dengue , in a country where severe dengue has not been reported and DF is rare . Serological evidence for previous monotypic and multitypic DENV infection in adults living along the northeastern coast of PNG between 1959–1963 indicates dengue was transmitted prior to this period . The contribution of dengue to acute febrile illness in PNG , and the reasons for the apparent lack of severe disease , should be investigated .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"dengue",
"virus",
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"immune",
"physiology",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"immunology",
"geographical",
"locations",
"microbiology",
"tropical",
"diseases",
"spatial",
"epidemiology",
"australia",
"animals",
"viruses",
"rna",
"viruses",
"neglected",
"tropical",
"diseases",
"antibodies",
"insect",
"vectors",
"antibody",
"response",
"immune",
"system",
"proteins",
"infectious",
"diseases",
"aedes",
"aegypti",
"proteins",
"medical",
"microbiology",
"dengue",
"fever",
"epidemiology",
"papua",
"new",
"guinea",
"microbial",
"pathogens",
"disease",
"vectors",
"insects",
"immune",
"response",
"arthropoda",
"people",
"and",
"places",
"biochemistry",
"mosquitoes",
"flaviviruses",
"oceania",
"viral",
"pathogens",
"physiology",
"biology",
"and",
"life",
"sciences",
"viral",
"diseases",
"species",
"interactions",
"organisms"
] |
2017
|
Serological evidence for transmission of multiple dengue virus serotypes in Papua New Guinea and West Papua prior to 1963
|
In vertebrates , the conserved Wnt signalling cascade promotes the stabilization and nuclear accumulation of β-catenin , which then associates with the lymphoid enhancer factor/T cell factor proteins ( LEF/TCFs ) to activate target genes . Wnt/β -catenin signalling is essential for T cell development and differentiation . Here we show that special AT-rich binding protein 1 ( SATB1 ) , the T lineage-enriched chromatin organizer and global regulator , interacts with β-catenin and recruits it to SATB1's genomic binding sites . Gene expression profiling revealed that the genes repressed by SATB1 are upregulated upon Wnt signalling . Competition between SATB1 and TCF affects the transcription of TCF-regulated genes upon β-catenin signalling . GATA-3 is a T helper type 2 ( TH2 ) specific transcription factor that regulates production of TH2 cytokines and functions as TH2 lineage determinant . SATB1 positively regulated GATA-3 and siRNA-mediated knockdown of SATB1 downregulated GATA-3 expression in differentiating human CD4+ T cells , suggesting that SATB1 influences TH2 lineage commitment by reprogramming gene expression . In the presence of Dickkopf 1 ( Dkk1 ) , an inhibitor of Wnt signalling , GATA-3 is downregulated and the expression of signature TH2 cytokines such as IL-4 , IL-10 , and IL-13 is reduced , indicating that Wnt signalling is essential for TH2 differentiation . Knockdown of β-catenin also produced similar results , confirming the role of Wnt/β-catenin signalling in TH2 differentiation . Furthermore , chromatin immunoprecipitation analysis revealed that SATB1 recruits β-catenin and p300 acetyltransferase on GATA-3 promoter in differentiating TH2 cells in a Wnt-dependent manner . SATB1 coordinates TH2 lineage commitment by reprogramming gene expression . The SATB1:β-catenin complex activates a number of SATB1 regulated genes , and hence this study has potential to find novel Wnt responsive genes . These results demonstrate that SATB1 orchestrates TH2 lineage commitment by mediating Wnt/β-catenin signalling . This report identifies a new global transcription factor involved in β-catenin signalling that may play a major role in dictating the functional outcomes of this signalling pathway during development , differentiation , and tumorigenesis .
Wnt growth factors regulate a variety of developmental processes by altering specific gene expression patterns [1] . Wnt proteins are secreted molecules that coordinate cell-to-cell interactions in many different cell types by binding to a member of the Frizzled ( Fz ) family of transmembrane receptors [2] . Binding of Wnt to Fz ( Table S1 ) elicits a complex cascade of molecular events culminating in the inhibition of the negative regulatory kinase GSK-3β [3] . Phosphorylation of β-catenin by GSK-3β targets it for degradation via the β-TrCP ubiquitin ligase-proteasome pathway [4] . Dephosphorylated β-catenin accumulates inside the nucleus [5] where it associates with the lymphoid enhancer factor/T cell factor ( LEF/TCF ) transcription factors to induce target gene transcription [6] . In vertebrates , β-catenin acts as a transcriptional activator , which is required to overcome the transcriptional repression by repressor complexes [7] . The C-terminus of β-catenin is indispensable for the transactivation function , presumably since it harbours binding sites for transcriptional coactivators such as p300/CBP and TBP [7] , [8] . Thus , recruitment of chromatin remodelling factors on TCF's genomic targets to modulate the gene transcription appears to be the principal function of stabilized β-catenin [8] . Within the thymus , thymocyte maturation involves a series of discrete phenotypic stages that correspond to developmental checkpoints and are subsequently referred to as CD4−CD8− ( DN ) , CD4+CD8+ ( DP ) , and CD4+CD3+ or CD8+CD3+ ( SP ) . In addition to the well-studied T cell receptor ( TCR ) -mediated signals and the Notch pathway , thymic epithelial cells also contribute towards thymocyte development and differentiation by producing Wnt [2] , [9] . The role of Wnt signalling in T cell development is highlighted by the fact that soluble Fz receptors have been shown to block early thymocyte development in fetal thymic organ culture at the DN-to-DP transition [10] . Furthermore , ablation of β-catenin from T cells resulted in poor β-selection and the DN stage of T cell development was impaired [11] . The SP cells then further differentiate into various functional T cell populations such as the T helper ( TH ) cells . Upon antigenic stimulus , naïve CD4+ TH cells differentiate into distinct effector cell lineages referred to as TH1 and TH2 . This differentiation is mediated by signalling proteins such as STAT4 and STAT6 and is accompanied by upregulation of marker transcription factors such as GATA-3 and c-Maf in TH2 cells and T-bet in TH1 cells [12] . The transcription factors GATA-3 and T-bet dictate differentiation of CD4+ T cells into TH2 and TH1 cell types , respectively [13] , [14] . However , although the contributions of various transcription factors and signalling pathways towards the TH cell differentiation have been studied extensively , there is no report demonstrating the direct role of Wnt signalling in TH cell differentiation . The T-lineage-enriched chromatin organizer special AT-rich sequence binding protein 1 ( SATB1 ) was shown to regulate distant genes by selectively tethering matrix attachment regions ( MARs ) to the nuclear matrix [15] . Furthermore , SATB1 acts as a “docking site” for several chromatin modifiers , including ACF , ISWI , and HDAC1 [16] , [17] , and these chromatin modifiers were suggested to suppress gene expression through histone deacetylation and nucleosome remodelling at SATB1-bound MARs [16] . SATB1 organizes the T helper 2 ( TH2 ) cytokine and MHC class-I loci into distinct chromatin loops by tethering MARs to the nuclear matrix at fixed distances [18] , [19] . The densely looped and transcriptionally active chromatin structure organized by SATB1 is essential for coordinated expression of the TH2 cytokine genes [18] . Moreover , SATB1 seems to play a role in dynamic organization of the transcriptionally poised chromatin [20] . SATB1 also regulates gene expression by recruiting various chromatin modifiers to promoters [21] . Interaction between SATB1 and partner proteins is frequently mediated by its N-terminal PDZ-like domain , which is also important for SATB1 homodimerization [22] , [23] . Additionally , SATB1 possesses a MAR-binding domain in its C-terminal half comprising a cut domain ( CD ) and a homeodomain ( HD ) that together contribute towards recognition and high affinity binding of MARs [23] , [24] . SATB1 regulates a large number of genes involved in T cell proliferation , development , and differentiation [21] , [25] . SATB1 itself is differentially expressed in various subsets of TH cells [26] , however the role of SATB1 in their differentiation has not been demonstrated . Interestingly , many of SATB1's target genes such as c-Myc [15] and Bcl-2 [27] are also targeted by Wnt/β-catenin [28] , suggesting a functional overlap between Wnt/β-catenin pathway and SATB1 . In this study , we set out to unravel the molecular mechanisms contributing towards the functional overlap between SATB1 and β-catenin pathways , especially with respect to their target gene regulation . We show that SATB1 physically interacts with β-catenin and recruits it to its genomic targets . Interestingly , deacetylation of SATB1 upon Wnt/β-catenin signalling leads to an increase in its occupancy on genomic targets . Recruitment of β-catenin alters the transcription of SATB1's target genes in thymocytes . Chromatin immunoprecipitation ( ChIP ) analysis of SATB1 binding sites ( SBSs ) in promoters of multiple genes revealed that interaction with β-catenin modulated SATB1 function on its target genes by increasing its occupancy and altering histone H3 lysine 9 ( H3K9 ) acetylation . β-catenin-responsive genes are also targeted by SATB1 , suggesting that both are functionally linked in the Wnt/β-catenin signalling pathway . Additionally , β-catenin-induced TCF responsive genes are dysregulated upon SATB1 overexpression , suggesting competitive interaction of TCF/LEF and SATB1 with β-catenin . Furthermore , SATB1 directly regulates GATA-3 expression and dictates TH2 lineage commitment via Wnt/β-catenin signalling . As a functional consequence of abrogating function of SATB1 or β-catenin or Wnt signalling , we demonstrate that the production of signature TH2 cytokines such as IL-4 , IL-10 , and IL-13 is reduced , indicating that TH2 differentiation is affected . Thus , these studies establish SATB1 as a determinant of TH2 differentiation and downstream effector in the Wnt/β-catenin signalling pathway .
Gene expression profiles of cells expressing phosphorylation- or acetylation-defective mutants of SATB1 [21] indicated shared target genes with the β-catenin signalling pathway ( Figure S1 ) . SATB1 is expressed abundantly in thymocytes [15] , and therefore we monitored the subcellular localization of β-catenin and SATB1 in thymocytes upon Wnt induction . Surprisingly , the intranuclear immunostaining pattern of β-catenin also resembled the “cage-like” architecture of SATB1-containing nuclear domains in thymocytes ( Figure 1A , Videos S1 and S2 ) [15] . Optical sectioning revealed that at least part of these signals colocalized across the depth of the nucleus , indicating that they occupy similar areas within the thymocyte nucleus ( Figure 1A ) . To test whether SATB1 and β-catenin interact physically , we performed in vitro pulldowns using immobilized β-catenin . When SATB1 was passed on GST-β-catenin and GST immobilized on Sepharose beads , SATB1 eluted specifically from the GST-β-catenin affinity matrix , suggesting their physical interaction ( Figure 1B ) . Coimmunoprecipitation analysis using nuclear extract from Jurkat T lymphoblastic cells indicated that β-catenin and SATB1 can be immunoprecipitated by antibodies against each other ( Figure S2 ) . We then tested the effect of Wnt signalling on the physical association of SATB1 and β-catenin . LiCl and 6-bromoindirubin-3′-oxime ( BIO ) are potent GSK-3β inhibitors that mimic Wnt signalling by stabilizing β-catenin [29] . Upon BIO treatment , increased interaction was observed in human thymocytes ( Figure 1C ) . In vitro pulldown experiments using the various truncated and tagged proteins ( Figure S3 ) indicated that the PDZ-like domain of SATB1 and carboxyl ( C ) terminal region of β-catenin ( 577–781 amino acids ) are indispensable for this interaction ( Figure 1D ) . Additionally , in vivo co-immunoprecipitation of β-catenin with Flag:PDZ ( 1–204 aa ) but not with Flag:CD+HD ( 330–763 aa ) indicated that β-catenin interacts with the PDZ-like domain of SATB1 ( Figure 1E ) . This interaction is very specific since expression of Flag-PDZ is at least 5-fold less as compared to that of Flag:CD+HD ( Figure 1E , compare lanes 5 and 6 ) . Finally , luciferase reporter based mammalian two-hybrid assay further independently confirmed that the PDZ-like domain of SATB1 is involved in its interaction with β-catenin in vivo . Although overexpression of the C-terminal region of β-catenin itself led to transactivation ( Figure 1F , bar 5 ) , along with the PDZ-like domain the reporter activity is enhanced further ( bar 9 ) , suggesting their functional association . To further assess the involvement of the C- terminal regions of β-catenin towards its interaction with SATB1 , we generated deletion constructs of β-catenin where N-terminus harbouring truncation consisted of aa 1–137 and C-terminus that of aa 666–781 . GST pulldown assay revealed that the C-terminal region encompassing aa 577–781 of β-catenin is important for its interaction with SATB1 , and immunoprecipitation data using 1–137 aa and 666–780 aa regions also confirmed the importance of extreme C-terminus ( 666–781 aa ) in this interaction ( Figure 1G ) . Collectively , these results indicate that the N-terminal PDZ-like domain of SATB1 interacts with the C-terminal region of β-catenin and that this interaction occurs in the nucleus . To investigate whether SATB1 and β-catenin collaborate functionally , we monitored the effect of SATB1 and β-catenin interaction on SATB1-mediated gene expression . For in vivo reporter assay , we used a construct in which the well-characterized SATB1 binding MAR from the IgH locus is cloned into a promoterless vector [30] . Overexpression of β-catenin enhanced the IgH MAR-driven luciferase reporter activity in the wild-type but not the mutant IgH MAR , suggesting that this effect is mediated by SATB1 ( Figure S4 ) . The T41A mutant of β-catenin used in reporter studies cannot be phosphorylated by GSK-3β [31] and is therefore constitutively stabilized and activated . We next monitored the effect of SATB1:β-catenin interaction on the transcription of multiple genes that are known to be targets of SATB1 [21] as well as Wnt signalling in immature T cells [32] . Wnt signalling was induced in human thymocytes by treating them with soluble Wnt3a ligand for 48 h , and the transcriptional activity of SATB1 targets was monitored by quantitative RT-PCR . In prolonged suspended cultures , thymocytes are known to undergo cell death . However , during the course of these treatments over 48 h in culture , thymocyte viability was not significantly reduced even without using the OP9-DL1 co-culture system ( Figure S5A ) . The transcription status of multiple Wnt- and SATB1-responsive genes was reversed upon induction of Wnt signalling by addition of Wnt3a in thymocyte culture ( Figure 2A , bar 2 ) . Thus , genes that are downregulated by SATB1 such as BCL-2 [27] , CHUK , PPM1A , c-Myc , and IL-2 [21] were upregulated upon overexpression of β-catenin or addition of Wnt3a in a manner akin to the known Wnt targets FOSB and BCL-XL [32] , suggesting that these are also β-catenin targets . Furthermore , inhibition of Wnt signalling by Dickkopf 1 ( Dkk1 ) treatment led to transcriptional repression of all SATB1 regulated genes that are also targeted by Wnt signalling ( Figure 2A , bar 3 ) . Among the SATB1 regulated genes , ThPOK ( ZBTB1-7B ) showed a reciprocal pattern of regulation . ThPOK was downregulated upon Wnt3a treatment and dramatically upregulated ( more than 20-fold ) upon inhibition of Wnt signalling by Dkk1 . Among the known Wnt targets in thymocytes , few genes such as TRAIL and PP2A [32] were downregulated upon Wnt3a treatment and upregulated upon Dkk1 treatment . More importantly , TIMP1 that is neither targeted by SATB1 nor targeted by β-catenin is not affected by Wnt3a or Dkk1 treatment of thymocytes , indicating that this effect is specific to SATB1 and Wnt/β-catenin targets ( Figure 2A ) . ERBB2 is a target of SATB1 but not that of β-catenin and therefore is not affected upon Wnt3a treatment , further demonstrating that to overcome the SATB1 mediated repression , SATB1:β-catenin functional interaction is essential to SATB1 genomic targets ( Figure 2A ) . Gene expression analysis using thymocytes co-cultured with OP9-DL1 cells supplemented with IL-7 and Flt3 ligand revealed that transcription of Wnt targets PP2A and BCL-XL is not significantly different than those cultured on the control OP9 cells ( Figure S5B ) . These results indicated that under the culture conditions employed , the effects on gene expression are not due to differential effects of these treatments on thymocyte viability . Thus , interaction with β-catenin dramatically alters the transcription status of SATB1 regulated genes , many of which are also known Wnt targets . To study how SATB1 directly regulates expression of these genes , we performed in vitro binding assays to assess the binding status of SATB1 on various regulatory regions from these genes and observed that SATB1 binds to at least one site located within the upstream 1 kb regions of multiple genes ( Figure S6 ) . To confirm whether the upregulation of transcription is mediated by SATB1 and factors recruited by SATB1 , we performed ChIP analysis of promoters of these genes . Chromatin was isolated from untreated control , Wnt3a or Dkk1 treated human thymocytes and subjected to immunoprecipitation using antibodies to SATB1 , β-catenin , p300 , and H3K9 acetylation ( H3K9Ac ) . The DNA purified from immunoprecipitate was subjected to PCR amplification using oligonucleotide primers flanking various regions within the upstream 1 kb regulatory regions of CHUK , PPM1A , and IL-2 promoters and the SBS within the major breakpoint region of BCL2 [27] . ChIP analysis revealed that not all of the in vitro binding sites were occupied by SATB1 in vivo ( Figure 2B ) . Strikingly , upon Wnt3a treatment the occupancy of SATB1 was enriched by 1 . 5- to 4-fold on promoters of genes that are downregulated by SATB1 . This change in occupancy is an outcome of Wnt signalling since treatment with Dkk1 resulted in a 1 . 5- to 6-fold decrease in SATB1 occupancy on its in vivo binding sites ( Figure 2B ) . The changes in occupancy of β-catenin , p300 , and H3K9Ac also mirrored that of SATB1 , suggesting that SATB1 occupancy is the primary event leading to recruitment of β-catenin and p300 ( Figure 2B ) . These changes in occupancy were highly specific and restricted to SBSs since the regions in proximal CHUK and distal PPM1A promoters that were not bound by SATB1 in vitro and in vivo did not show any significant change in the occupancy of SATB1 , β-catenin , and p300 ( Figure 2B ) . Thus , Wnt signalling results in increased occupancy of SATB1 on its targets , which then recruits β-catenin and p300 to upregulate target genes . To ascertain whether transcription regulation upon Wnt signalling is a direct consequence of the interaction between SATB1 and β-catenin or is an indirect effect of their interaction , we used the deletion constructs of β-catenin where C-terminus deleted truncation ( ΔC-term ) consisted of aa 1–576 and only C-terminus harbouring construct that of aa 577–780 . We tested the possibility of whether overexpression of this interaction region works as a dominant negative for SATB1 and β-catenin interaction . Towards this , we overexpressed 1–576 and 577–780 regions of β-catenin as red fluorescent protein ( RFP ) fusions in Human Embryonic Kidney ( HEK ) -293T cells and induced β-catenin signalling by adding LiCl . SATB1 immunoprecipitation followed by immunoblot analysis with anti-β-catenin confirmed that they form a complex in LiCl-treated cells ( Figure 3A , lane 1 ) . However , upon overexpression of the C-terminus ( 577–780 ) , endogenous β-catenin no longer associated with SATB1 as evidenced by a lack of signal in the coimmunoprecipitation assay ( Figure 3A , lane 3 ) . Overexpression of the ΔC-term ( 1–576 ) did not lead to the displacement of endogenous β-catenin ( Figure 3A , lane 2 ) . This result suggests that endogenous full-length β-catenin and overexpressed C-terminus of β-catenin compete with each other to bind with SATB1; hence the latter serves as a dominant negative for their association . Interestingly , quantitative RT-PCR analysis under such conditions yielded reduction in the activating effect of stabilized β-catenin on c-Myc that is known to be repressed by SATB1 [15] . However , overexpression of the ΔC-term β-catenin did not alter the expression profile of c-Myc ( Figure 3B ) . Thus , overexpression of C-terminus alters the effect of β-catenin on transcription of SATB1 targets in vivo . To investigate the molecular mechanism of this effect , we tested the occupancy of SATB1 and β-catenin at the c-Myc promoter . ChIP analysis demonstrated the occupancy of both SATB1 and β-catenin specifically at the SBS of c-Myc promoter in vivo but not at another region 10 kb upstream of the start site ( Figure S7 ) . Next , we independently overexpressed VP16-fused truncations 1–137 and 666–780 of β-catenin representing its extreme N- and C-termini , respectively , in HEK 293T cells followed by treatment with LiCl for 24 h . Extreme terminal regions of β-catenin were overexpressed in this assay to avoid any interference from proteins interacting with the arm repeats such as TCF . ChIP analysis using SATB1 and β-catenin antibodies followed by quantitative ChIP PCRs revealed that overexpression of C-terminal region of β-catenin ( 666–780 ) led to the reduced occupancy of β-catenin on c-Myc SBS ( Figure 3C , bar 6 ) . In contrast , overexpression of N-terminal region of β-catenin ( 1–137 ) failed to exert the same effect ( Figure 3C , bar 5 ) . Since the epitope of anti-β-catenin used here spans only the N-terminal region , we used an antibody for the VP16 fusion tag to monitor whether the C-terminus is indeed recruited by SATB1 . Quantitative ChIP PCRs yielded about 5-fold enrichment of the C-terminus of β-catenin at the c-Myc SBS ( Figure 3C ) , providing compelling evidence for the in vivo displacement of full-length β-catenin . We next monitored transcription activity under the control of a concatemerized TCF binding site–driven reporter TOP-Flash and its non-TCF binding mutant version FOP-Flash [33] . Interestingly , although SATB1 does not directly bind to the consensus TCF binding site ( Figure S8 ) , overexpression of SATB1 led to repression of TCF reporter–driven transcription ( Figure 3D , compare bars 1 and 2 ) , indicating that SATB1 may compete with TCF for forming a complex with β-catenin . Strikingly , silencing of SATB1 upregulated TCF-driven transcription ( bar 3 ) , which was dramatically induced when β-catenin was simultaneously overexpressed , suggesting a role of SATB1 in titrating β-catenin in vivo ( Figure 3D , bar 8 ) . This increase in reporter activity was significantly more than with β-catenin alone ( bar 4 ) . We used the T41A constitutively active mutant of β-catenin for the TCF reporter assay . Next , we used the RFP-fused N-term ( 1–576 ) and C-term ( 577–780 ) regions of β-catenin as described above to test whether the Wnt-responsive reporter activity is a direct consequence of the interaction between SATB1 and β-catenin . Overexpression of the N-term ( 1–576 ) β-catenin that interacts with TCF but does not interact with SATB1 upregulated the TCF-reporter activity , although to a much lower extent ( bar 5 ) . The activating effect of this arm repeat containing region of β-catenin may be explained by its ability to sequester inhibitors such as ICAT that inhibit TCF-β-catenin interaction . However , upon overexpression of only the C-terminal region ( 577–780 ) of β-catenin that interacts with SATB1 but not with TCF , the reporter activity was completely abolished ( bar 6 ) , presumably due to the dominant negative effect exerted by the titration of endogenous p300 . Such interaction may result in the diminished availability of CBP/p300 for association with the TCF:β-catenin complex on TCF responsive elements , resulting in the non-functional interaction between TCF and β-catenin . Alternatively , the pool of endogenous full-length β-catenin that is displaced from SATB1 upon overexpression of C-terminal of β-catenin may not be available for binding to TCF and the subsequent activation of target genes and may simply be turned over . Collectively , these findings provide direct evidence that SATB1-β-catenin interaction is required for mediating the Wnt-dependent effects on target gene expression . Since we tested transcription of a number of genes that are targets of both SATB1 and TCF-1 , we tested whether they interact with each other . Immunoprecipitation of thymocyte extract with anti-TCF-1 did not pulldown SATB1 , suggesting a lack of their physical interaction ( Figure 4A , lane 2 ) . The TCF family of proteins consists of four members that are expressed as multiple isoforms through alternative splicing [34] . TCF-4 is the largest of all TCF family proteins containing all known functional domains , and therefore we employed it in the binding studies such that all possible domains including those that are not present in other TCF isoforms are also tested . TCF-4 is not known to play any role in T cell development [35] but plays an important role in epithelial cells such as HEK-293T [36] . We therefore used extract derived from control and LiCl treated HEK 293T cells for immunoprecipitation using anti-TCF-4 . Immunoblot analysis using anti-SATB1 revealed an absence of interaction between SATB1 and TCF-4 even upon LiCl treatment ( Figure 4A , lanes 5 and 6 ) . After establishing that these two proteins do not interact with each other , we then explored whether they compete for association with β-catenin . We performed an in vitro competition experiment by loading TCF-4 on immobilized GST-β-catenin and GST-Arm proteins . The bound TCF-4 was then challenged separately with three functional domains of SATB1 , namely PDZ , HD , and CD , and the TCF-4 remaining bound to the columns was then eluted and monitored by immunoblot analysis . Strikingly , although TCF-4 protein could be detected in eluates after incubation with HD or CD ( Figure 4B , lanes 3 and 4 ) , it was virtually undetectable after incubation with PDZ domain ( lane 2 ) , suggesting dissociation of TCF-4 from the β-catenin affinity matrix . In contrast , TCF-4 bound to the GST-Arm repeat affinity column was not displaced by PDZ domain ( Figure 4B , lane 7 ) , indicating that direct interaction between PDZ of SATB1 and the C-terminal region of β-catenin is required for its ability to compete with TCF-4 . To monitor if TCF-4 and SATB1 compete in vivo , we performed co-immunoprecipitation analysis using extracts from HEK-293T cells transfected with control Flag vector or Flag-TCF4 and then treated with LiCl . Immunoprecipitation with anti-SATB1 yielded a marked reduction in β-catenin pulldown upon TCF-4 overexpression as compared to the vector control ( Figure 4C ) . This observation provides strong evidence towards in vivo competition between SATB1 and TCF-4 for interaction with β-catenin . To test whether in vivo competition between these two transcription factors for recruitment of β-catenin affects gene expression , we monitored expression of two TCF responsive genes , integrin β1 and cyclin D1 [37] , [38] . Overexpression of the T41A constitutively active β-catenin upregulated these genes ( Figure 4D , bar 2 ) . Expression of both these genes was negatively affected upon SATB1 overexpression presumably through titration of TCF-associated β-catenin ( bar 6 ) . When SATB1 and β-catenin were overexpressed simultaneously , the transcription of Integrin β1 and Cyclin D1 was reduced by about 2-fold ( bar 3 ) , indicating that overexpression of β-catenin partially derepresses the SATB1-mediated downregulation . However , overexpression of N-term ( 1–576 ) β-catenin does not yield a similar effect upon coexpression of SATB1 ( Figure 4D , compare lanes 4 and 5 ) , presumably since it lacks the SATB1-interacting C-terminal region of β-catenin . This result provides compelling evidence for requirement of SATB1: β-catenin interaction towards observed effects on gene regulation . Considered together , these data suggest that SATB1 mediates β-catenin signalling by competitively recruiting β-catenin and thereby also affects the transcription of TCF-regulated genes . The DNA binding activity of SATB1 is regulated by its post-translational modifications such as phosphorylation and acetylation . In particular , acetylation has a drastic negative influence on DNA binding activity of SATB1 [21] . To investigate the molecular mechanism of increased occupancy of SATB1 on its in vivo genomic targets , we monitored the acetylation level of SATB1 during the course of β-catenin accumulation . We performed immunoprecipitation of SATB1 in HEK 293T and Jurkat cells treated with LiCl over a time course of 36 h at an interval of every 6 h using SATB1 antibody followed by immunoblotting by pan-acetyl antibody . Interestingly , LiCl treatment led to a time-dependent decrease in acetylation of SATB1 in HEK-293T cells and Jurkat T cells ( Figure 5A ) . We observed two bands of acetylated proteins in this study . However , molecular weight of SATB1 matched with the upper band only . The lower band could be due to immunoprecipitation of a SATB1 interacting protein that is also deacetylated upon Wnt signalling . While SATB1 levels did not change during the time course; as expected , β-catenin was stabilized over 36 h ( Figure 5A ) . Similarly , upon BIO treatment we observed time-dependent deacetylation of SATB1 and accumulation of β-catenin over 36 h in primary human thymocytes ( Figure 5A ) . Interestingly , PCAF acetyltransferase that is known to acetylate SATB1 [21] is downregulated in a time-dependent manner upon BIO treatment of thymocytes ( Figure 5A ) , suggesting a probable mechanism for deacetylation of SATB1 during this time course . Consequently , IgH MAR and IL-2 promoter-mediated reporter activity was upregulated in a time-dependent manner upon LiCl treatment ( Figure 5B ) . To induce Wnt signalling in naïve T cells , we used the soluble Wnt3a ligand and performed IgH MAR-luciferase reporter assay . Upon treatment with Wnt3a , the reporter activity was stimulated more than 6-fold over 36 h , indicating stabilization of β-catenin and presumably its increased association with deacetylated SATB1 . Deacetylated SATB1 has higher affinity for DNA [21] and recruits various cofactors such as β-catenin and p300 upon Wnt signalling , leading to the upregulation of MAR-linked transcription . Both IgH MAR and IL-2 distal promoter reporter constructs harbour SBS [17] but do not contain the consensus TCF binding site; hence the observed upregulation is due to SATB1 only . Thus , these data argue that Wnt/β-catenin signalling induces deacetylation of SATB1 , which then promotes its association with β-catenin and thereby alleviates SATB1 mediated repression . Next , to evaluate occupancy of SATB1-β-catenin to its genomic targets in a time-dependent manner , we monitored their occupancy at promoters of various genes containing SBS by ChIP . Upon LiCl treatment , SATB1 occupancy on c-Myc promoter is enhanced and β-catenin also follows a similar pattern ( Figure 5C ) . Both IL-2 and c-Myc are negatively regulated by SATB1 [15] , [17] but are induced upon Wnt signalling [28] , presumably by direct recruitment of the SATB1:β-catenin complex to their upstream regulatory regions containing the SBSs ( Figure S7 ) . Moreover , the time-dependent increase in the occupancy of the SATB1:β-catenin complex is accompanied with an increase in H3K9 acetylation at the promoters indicative of transcriptional activation ( Figure 5C ) . Collectively , these findings further confirm that recruitment of β-catenin drastically alters the fate of SATB1-regulated genes . Expression of multiple members of the Wnt family has been documented in thymocytes [39] . However , expression of neither the Wnt molecules nor any of the downstream effectors has been documented in differentiating CD4+ T cells . To test whether Wnt signalling is active in CD4+ T cells , we isolated naïve CD4+ T cells from human umbilical cord blood and monitored transcription activity under the control of a concatemerized TCF binding site–driven reporter TOP-Flash in absence of any Wnt agonist . Control untreated CD4+ T cells showed low but consistent TCF reporter activity ( Figure 6A , bar 1 ) , suggesting that low levels of Wnt signals are produced by the T cells . Treatment with Dkk1 or transfection of siβ-catenin almost completely abolished the endogenous Wnt activity ( Figure 6A , bars 2 and 3 ) , suggesting that Wnt signalling is active in undifferentiated CD4+ T cells . Overexpression of the constitutively active T41A β-catenin yielded a dramatic increase in the TCF activity , indicating that CD4+ T cells are responsive to Wnt/β-catenin signals . The differential sensitivity of various subsets of thymocytes has been attributed to differential expression of the Wnt components [39] . To investigate whether CD4+ T cells isolated from cord blood and cultured ex vivo are responsive to Wnt signalling due to the Wnts produced by them , we examined the presence of several Wnts using RT-PCR and real-time quantitative RT-PCR . Most of the investigated Wnts appeared to be differentially expressed in these cells . Wnt 2 was the most prominent Wnt , followed by Wnt 4 and Wnt 5a ( Figure 6B ) . Wnt 5b and Wnt 1 were expressed at extremely low levels and detected only by quantitative RT-PCR ( Figure 6B ) . Wnt 5b is highly expressed in SP and DP thymocytes [39] , and therefore it was used as a reference for calculating fold expression of other Wnts in CD4+ T cells . In second assay , the naïve CD4+ T cells were first activated with plate-bound anti-CD3 and soluble anti-CD28 . TH1 and TH2 differentiation was then induced by adding cytokines IL-12 and IL-4 , respectively , and cells were differentiated for 72 h . Expression of the TH1 marker cytokines IFN-γ and IL-12 as well as TH2 marker cytokine IL-4 in the cell supernatants was analyzed using the protein bead array to confirm their polarization ( Figure S9 ) . Stabilization of β-catenin upon differentiation was monitored by immunoblot analysis of nuclear extracts from the two subtypes of TH cells . Endogenously stabilized β-catenin was detected in both TH1 and TH2 cells , although TH2 cells expressed 2-fold more β-catenin as compared to TH1 ( Figure 6C , compare lane 4 with lane 1 ) . To monitor the effect of Wnt signalling on the stabilization of β-catenin during differentiation , activated CD4+ T cells were differentiated for 72 h in the presence of soluble Dkk1 or Wnt agonist BIO . Treatment with Dkk1 resulted in reduction in endogenously stabilized β-catenin , which was more prominent in TH2 cells as compared to that of TH1 ( compare lane 6 with lane 3 ) . Treatment with BIO resulted in stabilization of β-catenin predominantly in TH2 cells ( Figure 6C , compare lane 4 with lane 5 ) , indicating that Wnt signalling was active in these cells . Thus , these results corroborate the finding that low levels of Wnt signals are produced by the TH cells themselves and that the downstream processes such as stabilization of β-catenin also occur in the two subtypes of TH cells , albeit to differential extent . To test whether Wnt signalling is involved in TH cell commitment , we monitored the effect of Wnt inhibitor Dkk1 on the transcription of the TH2 marker GATA-3 . Naïve CD4+ T cells were isolated from human umbilical cord blood and activated with plate-bound anti-CD3 and soluble anti-CD28 ( TH0 state ) [26] and differentiated in the presence of specific cytokines as described above . Soluble Dkk1 was added to the medium in which TH2 cells were cultured . Cells were differentiated for 72 h and supplemented with Dkk1 after every 24 h . Expression of GATA-3 was monitored as an indicator of the TH2 differentiation . Quantitative transcript profiling revealed that GATA-3 expression was suppressed by over 3-fold upon Dkk1 treatment in TH2 subset , suggesting that Wnt signalling is necessary for the upregulation of GATA-3 during differentiation of TH2 cells ( Figure 7A ) . To directly assess the role of SATB1 in TH cell differentiation , we altered their expression level by overexpression and siRNA mediated silencing and monitored the differentiation of CD4+ cells . Scrambled siRNA was used as a control . CD4+ cells transfected with siSATB1 or SATB1 expression construct were cultured for 24 h , activated with plate-bound anti-CD3 and soluble anti-CD28 , polarized by adding cytokines , and grown further for 24 h . The expression of TH2 marker GATA-3 was monitored by quantitative RT-PCR using total RNA extracted from these cells . Upon siRNA mediated silencing of SATB1 expression GATA-3 was downregulated in TH0 and TH2 cells ( Figure 7B ) . The expression of GATA-3 in scrambled siRNA treated control TH0 cells was considered as baseline ( Figure 7B , bar 1 ) . Upon SATB1 knockdown , GATA-3 expression was reduced by 2-fold in TH0 cells ( bar 2 ) . Notably , in TH2 cells GATA-3 was downregulated more than 11-fold upon SATB1 knockdown ( bar 6 ) . The downregulation of GATA-3 in TH2 was more striking as compared to the TH0 subset , indicating that SATB1 is required for GATA-3 expression in TH2 cells . The influence of SATB1 on GATA-3 expression was further confirmed by overexpression of SATB1 that led to a significant increase in the expression of GATA-3 in both subsets as compared to the respective controls ( compare bars 3 and 7 with 1 and 5 , respectively ) , suggesting that SATB1 positively regulates GATA-3 expression . Under these conditions , the expression of another TH2-specific transcription factor c-Maf was not altered significantly ( Figure S10 ) , suggesting that the regulatory effect of SATB1 is specific for GATA-3 at least during the early differentiation of TH cells . The differentiation of TH2 cells is characterized by secretion of signature cytokines that are encoded by the TH2 cytokine locus [40] . Therefore , to monitor the effect of Wnt/β-catenin signalling and SATB1 on expression of signature interleukins from differentiated TH cells , culture supernatants of TH cells that were subjected to GATA-3 expression profiling upon knockdown of SATB1 and β-catenin and also culture supernatants from Dkk1 treated TH cells were collected . The expression of secreted interleukins in the control , Dkk1 treated , and siRNA treated cell supernatants was monitored using multiplex protein bead array . Strikingly , signature TH2 cytokines such as IL-4 , IL-10 , and IL-13 were downregulated upon Dkk1 treatment ( Figure 7C ) and upon knockdown of SATB1 as well as β-catenin ( Figure 7D ) , independently confirming the requirement of SATB1 and Wnt/β-catenin signalling during TH cell differentiation . Knockdown of SATB1 also led to dramatic downregulation of the IL-4 transcript as well as intracellular IL-4 levels in TH2 cells ( Figure S11 ) . Together , these results argue that SATB1 and Wnt/β-catenin signalling positively regulate expression of signature interleukins in differentiating TH2 cells . To understand whether SATB1 and β-catenin regulate GATA-3 expression independently or they are functionally linked , we monitored the occupancy of SATB1 on the 1 kb upstream region from the transcription start site of GATA-3 and found one potential SBS . Direct binding of SATB1 within this region encompassing around 600–900 bp upstream of the transcription start site was established in vitro by EMSA ( Figure S12 ) . Next , we monitored the occupancy of SATB1 in vivo using human CD4+ TH cells in which expression of SATB1 was silenced using siSATB1 . Cells transfected with scrambled RNA ( Scr ) served as the control . Both siSATB1 and Scr transfected cells were differentiated into TH2 cell lineage by adding IL-4 over the next 72 h . Cells that were only activated by plate bound anti-CD3 and anti-CD28 and not treated with IL-4 were referred to as TH0 . We subjected these differentiated cells to ChIP analysis and monitored the occupancy of SATB1 and β-catenin at the SBS within the upstream region of GATA-3 promoter by quantitative PCR analysis of the eluted chromatin . As a control for the ChIP analysis , we used another region at −1 , 500 to −1 , 800 bp of GATA-3 promoter that was not directly bound by SATB1 ( non-SBS ) ( Figure S12 ) . The occupancy of SATB1 gradually increased on promoter from 24 to 72 h of differentiation in control ( Scr ) TH2 cells ( Figure 8A , bars 5 , 9 , and 13 ) . SATB1 knockdown resulted in drastic reduction of its occupancy on GATA-3 promoter ( bars 6 , 10 , and 14 ) . Strikingly , β-catenin occupancy also increased during the differentiation of control TH2 cells from 24 to 72 h and was significantly reduced upon SATB1 knockdown ( compare bars 7 , 11 , and 15 with bars 8 , 12 , and 16 , respectively ) . Such increase in the occupancy of SATB1 or β-catenin was not observed at the upstream non-SBS during the time course of differentiation ( bars 21–32 ) , suggesting that SATB1 recruits β-catenin at the SBS in GATA-3 promoter . To assess the contribution of β-catenin towards the positive influence of the SATB1:β-catenin complex on GATA-3 promoter , we inhibited the Wnt signalling by treating the differentiating TH cells with Dkk1 . The TH cells were differentiated as mentioned above and treated with Dkk1 for 72 h . The activated TH cells ( TH0 ) were used as controls . We differentiated cells for 72 h because the maximum occupancy of β-catenin and SATB1 on GATA-3 was observed upon 72 h of differentiation . Chromatin isolated from differentiated TH0 ( control ) and TH2 , Dkk1 treated , and untreated cells was subjected to immunoprecipitation using antibodies to SATB1 , β-catenin , and p300 . The occupancy of these proteins at the SBS within the GATA-3 promoter was monitored during Wnt-on and -off conditions by quantitative PCR of immunoprecipitated DNAs . Upon Dkk1 treatment ( Wnt-off ) , since the nuclear β-catenin level is reduced , its occupancy on GATA-3 promoter was also reduced significantly ( Figure 8B , compare bars 9 and 10 ) . Occupancy of SATB1 was reduced at the SBS upon Dkk1 treatment ( compare bars 7 and 8 ) , suggesting that Wnt signalling is required for increased SATB1 occupancy on its genomic targets . Furthermore , p300 occupancy is also reduced at the GATA-3 promoter upon Dkk1 treatment ( compare bars 11 and 12 ) , corroborating the observation that the SATB1:β-catenin complex recruits p300 . This further explains the observed downregulation of GATA-3 upon Dkk1 treatment ( Figure 7A ) . Such changes in the occupancy of SATB1 , β-catenin , and p300 were not observed at the non-SBS within GATA-3 promoter ( Figure 8A , lanes 19–24 ) , indicating that Wnt signalling-dependent recruitment of SATB1 at the SBS in GATA-3 promoter is required for upregulation of GATA-3 during TH2 differentiation .
The non-random occupancy of SATB1 across the entire genome presumably results in the formation of a characteristic “cage-like” network in mouse thymocytes that demarcates heterochromatin from euchromatin [15] , [20] . Furthermore , the “clustering” type occupancy exhibited by SATB1 at MARs and the upstream regulatory sequences of genes indicates a dichotomous role for SATB1 as a structural and regulatory chromatin component [19] . SATB1 presumably occupies the interchromatin territory or the nuclear space that contains genes that are transcriptionally poised [20] . What could be the consequence of sequestration of β-catenin by SATB1 in this nuclear compartment ? The immunostaining of SATB1 and β-catenin in primary thymocytes indicates that they colocalize in the interchromatin space and form a distinguished substructure inside the nucleus . Since SATB1 is the only chromatin-associated protein known to harbour a PDZ-like signalling domain and the C-terminus of β-catenin interacts with SATB1 via this domain , it may also be speculated that the Wnt signalling cascade may cross-talk with the PDZ signalling cascade . β-catenin does not bind to DNA by itself; therefore SATB1 must recruit it to its genomic targets . Once recruited by SATB1 onto its genomic targets , β-catenin may then recruit the chromatin-modifying and -remodeling complexes to transcribe the Wnt target genes . This aspect is similar to the TCF/LEF family HMG box transcription factors , which need accessory factors to activate transcription at the site of their recruitment [34] , [41] . The HMG box mediates sequence-specific binding to a core consensus sequence AGATCAAAGGG [42] . SATB1 also binds DNA in a sequence-specific manner to a 12-mer consensus sequence TATTAGTAATAT resembling the HD consensus [23] . Given that SATB1 and TCF have different recognition sites on the DNA , it is less likely that SATB1 can directly compete for TCF binding sites . Indeed , in vitro binding studies provide evidence that SATB1 does not bind directly to the TCF consensus element . However , since Wnt signalling also appears to affect the SATB1 responsive IgH-MAR-linked reporter in vivo , the possibility of SATB1's effect on TCF binding sites cannot be excluded . Many of the SATB1 regulated genes could also respond to Wnt signalling , and therefore this study has the potential to unravel hitherto unknown Wnt targets in T cells . In fact our gene expression profiling data unravelled ThPOK as a new Wnt target in thymocytes . ThPOK is repressed by SATB1 [21] ( Limaye and Galande , manuscript in preparation ) and is important for CD4 T cell development [43] . However , the precise mechanism of how only a specific subset of SATB1 targets is responsive to Wnt/β-catenin and not all targets requires further investigation . The cellular expression levels of TCF and SATB1 could play a decisive role in determining the outcome of their interaction with β-catenin . We found that SATB1 competes with TCF for sequestering β-catenin . Since not all of SATB1 targets are Wnt targets and vice versa , the choice among these effector proteins as a partner of β-catenin could dictate the developmental fate of cells . There could be functional redundancy between SATB1 and TCF because TCF is expressed in some cell types but SATB1 is not and vice versa . After 48 h of differentiation , TH1 cells predominantly express TCF-1 and SATB1 is expressed at low levels , whereas TH2 cells express SATB1 only [26] . It would be interesting to study the mechanism that determines the occupancy of SATB1 and TCF-1 on promoters when both these proteins are expressed simultaneously in the same cell . One possibility that we have tested is that of their competition to occupy the genomic targets by interacting with β-catenin . TCF interacts with β-catenin via arm repeats , whereas SATB1 interacts with its C-terminus and therefore this competition is not driven by the site of interaction on β-catenin . Thus , the competition could be driven by the post-translational modifications and associated partners of these proteins . The DNA binding sites for TCF and SATB1 are not similar , and therefore the competition is not at the level of DNA binding but at the level of protein-protein interactions . Strikingly , upon SATB1 overexpression , TCF responsive genes that are not SATB1 targets are negatively affected in a β-catenin-dependent manner , indicating titration of β-catenin . A number of studies have specifically addressed the relevance of Wnt signalling for lymphocyte development and typically involve genetic and/or biochemical analyses of TCF functions [2] , [32] , [35] , [39] , [44]–[48] . Wnt signalling is essential for T cell development in thymus , and it occurs in all thymocyte subsets but to different extents [2] , [35] , [39] , [45] . The differential sensitivity to Wnt signalling is attributed to differential expression of Wnt ( s ) and its receptor and that of activating signalling molecules such as the long form of TCF-1 and β-catenin in different subsets of T cells [39] . The high redundancy of various Wnts and Fz in thymus masks their importance in knockout mouse models [2] . However , the involvement of β-catenin in T cell development could be elucidated by overexpression of the negative regulators of the Wnt/β-catenin pathway [10] . Studies of the molecular mechanisms of regulation of gene expression in T cells upon Wnt/β-catenin signalling have also been restricted to the TCF/LEF family proteins [34] . Our findings of physical association between SATB1 and β-catenin and its role in regulation of transcription of multiple genes provides an unprecedented clue towards the role of Wnt/β-catenin signalling in T cell development and differentiation . SATB1 is known to orchestrate spatio-temporal expression of multiple genes during T cell development [21] , [25] . The effect of SATB1 deficiency on thymocyte development resembles that of β-catenin or TCF-deficient thymocytes [11] , [25] , [46] . Although Wnt signalling is known to occur in all thymocyte subsets , it occurs predominantly in the double negative ( DN ) subsets due to the high expression of activating Wnt transducers such as β-catenin and TCF-1 [39] . Interestingly , in SATB1 deficient mice , DN cells were greatly reduced in number [25] . Wnt/β-catenin signalling also plays a role in late stages of thymocyte development such as differentiation into CD4+CD8+ DP cells [47] , generation of mature CD8+ SP thymocytes [48] , and differentiation of CD4+ SP cells to TH2 cells ( this investigation ) . SATB1 presumably contributes to all of the above since in SATB1 null mice the development of thymocytes is arrested at the DP stage , very few CD8 SP cells are observed [25] , and SATB1 is upregulated early during TH cell differentiation [26] . However , SATB1 has not been directly shown to be important for TH cell development , and the role of Wnt signalling in this process is also unprecedented . We show that CD4+ T cells are receptive to Wnt signals presumably because they produce different Wnts themselves . The differential sensitivity of TH cell subtypes to Wnt signalling could be due to the fact that the downstream processes such as stabilization of β-catenin occur prominently in the TH2 subtype . Our study defines the role of SATB1:β-catenin collaboration in this important biological phenomenon . Further , SATB1 reprogrammes chromatin organization and gene expression profiles to promote breast tumor growth and metastasis [49] . Our study provides a plausible mechanism for such reprogramming via recruitment of β-catenin and chromatin modifying machinery . The implications of SATB1-mediated orderly deposition of β-catenin and its partners on chromatin towards global gene regulation upon Wnt signalling in various developmental systems and during tumorigenesis await further investigation . In light of our findings , we propose that in absence of Wnt signalling when β-catenin is phosphorylated and degraded , SATB1 acts predominantly as a repressor on most of the genes and recruits different chromatin remodelling complexes to its genomic targets . However , upon β-catenin stabilization , it associates with β-catenin inside the euchromatic nuclear compartment , and its binding to target DNA is enhanced along with recruitment of β-catenin and H3K9 acetylation . Thus our study provides a mechanism explaining how SATB1's regulatory functions are modulated from repressor to activator in a signal-dependent manner . SATB1 acts as a “landing platform” for chromatin remodelling factors [16] and depending upon its post-translational modification status may selectively use its associated partners to modulate target gene transcription [21] . Wnt signalling activates the SATB1 target genes , which are otherwise repressed by SATB1 . This switching mechanism is similar to that reported for the TCF:β-catenin complex , where interaction of β-catenin with TCF overcomes the repressor effect and multiple genes are activated [34] . Similarly , the β-catenin-HD containing protein Prop1 complex also works as a binary switch to simultaneously activate expression of the critical lineage-determining factor Pit1 and represses the gene encoding the lineage-inhibiting transcription factor Hesx1 acting via the TLE/Reptin/HDAC1 corepressor complexes [50] . This function resembles that of MyoD , CREB , and STAT-1 factors that simultaneously interact with several modifiers and selectively utilize their enzymatic activities for promoter stimulation [51] , [52] . Deacetylation of SATB1 upon Wnt signalling is a key event modulating transcription . Acetylation has been shown to abrogate the DNA-binding activity of SATB1 [21] and therefore deacetylation is mandatory for enhancing its occupancy on genomic targets . However , acetylation of these two proteins has opposing effects on their association and ability to regulate transcription ( Notani and Galande , manuscript in preparation ) . Thus , Wnt signalling leads to deacetylation of SATB1 that facilitates its interaction with β-catenin and thereby results in enhanced recruitment of SATB1 on its genomic targets . The Histone Acetyl Transferases ( HATs ) that acetylate these two proteins , namely PCAF and p300 , respectively , may therefore regulate SATB1-mediated transcription upon Wnt signalling . TCF , the transcription factor known to recruit β-catenin on Wnt target genes , requires p300 as a cofactor in order to work as an activator [7] , whereas SATB1 is dependent on PCAF for such functions [21] . Thus , post-translational modifications of TCF , SATB1 , and β-catenin may fine tune the transcriptional outcome of the Wnt signalling cascade . Interestingly , PCAF itself is downregulated upon Wnt signalling in human thymocytes , especially at later time points . However , at very early time points ( up to 9 h upon induction of Wnt signalling ) , PCAF is upregulated and so is the acetylation of SATB1 , leading to its dissociation from the CtBP1 corepressor [53] . Therefore , acetylation seems to be the molecular switch governing the ability of SATB1 to function as an activator or repressor of multiple genes in a Wnt-dependent manner . Role of SATB1 in regulation of TH2 specific cytokine locus has been suggested wherein SATB1 facilitates the expression of IL-4 , IL-5 , IL-13 , and c-Maf via formation of a densely looped , transcriptionally active , higher-order chromatin structure upon TH2 cell activation [18] . TH cell differentiation is mediated by signalling proteins , including STAT4 and STAT6 , resulting in expression of transcription factors such as GATA-3 and c-Maf ( in TH2 cells ) or T-bet ( in TH1 cells ) [12] . GATA-3 mediates TH2 responses through three different mechanisms: induction of TH2 cytokine production , selective growth of TH2 cells , and inhibition of TH1 promoting factors [13] , [54] . Moreover , GATA-3 expression is necessary and sufficient for TH2 polarization [55] . GATA-3 is a marker of TH2 cells and SATB1 is also upregulated in TH2 cells [26] . Therefore , to monitor the role of SATB1 in TH cell differentiation , we silenced SATB1 expression and monitored the TH cell differentiation by quantifying GATA-3 expression . Interestingly , upon siRNA mediated knockdown of SATB1 , GATA-3 expression is drastically decreased , indicating that TH2 polarization is SATB1-dependent . This is further substantiated by results of SATB1 overexpression in TH2 cells , which clearly establish SATB1 as a positive regulator of GATA-3 . The Wnt/β-catenin signalling is active in CD4+ TH cells as revealed by stabilization of β-catenin . The differential stabilization of β-catenin in TH1 and TH2 subtypes may explain the differential pattern of gene expression in these cells . We therefore used Dkk1 treatment to inhibit the active Wnt signalling in TH2 cells . Upon Dkk1 treatment , GATA-3 expression was reduced drastically , suggesting that GATA-3 is regulated in a Wnt/β-catenin signalling-dependent manner . Notably , c-Maf is not regulated by SATB1 or β-catenin under these conditions . Thus , at least during the early commitment and differentiation of human CD4+ TH cells , SATB1 does not seem to regulate c-Maf . SATB1 regulates GATA-3 by directly binding to its promoter elements and may not regulate c-Maf in this manner . However , further studies may be required to address the role of SATB1 in regulation of c-Maf during different stages of TH differentiation since positive regulation of c-Maf by SATB1 has been demonstrated in the differentiated mouse TH2 cell line ( clone D10 ) [18] . Most strikingly , the expression of signature TH2 cytokines such as IL-4 , IL-10 , and IL-13 is also positively regulated by both SATB1 and Wnt signalling , as evident from their downregulation upon silencing of SATB1 or β-catenin and inhibition of Wnt signalling by Dkk1 . The early time points in differentiation as employed in this study do not yield a very high amount of secreted cytokines in cell culture supernatants . Interestingly , at the RNA level as well as at the intracellular protein level , dramatic downregulation of IL-4 was observed . The secretion of IL-4 and other cytokines presumably surges after 72 h of differentiation . However , the silencing effect of siRNA would not last over such a long time period , and therefore effects of siSATB1 or siβ-catenin could not be monitored beyond 72 h in culture . Intriguingly , IL-10 gene is located outside of the TH2 cytokine locus and yet is co-ordinately regulated by SATB1 in a Wnt-dependent manner . In GATA-3-deficient TH2 cells , production of signature TH2 cytokines such as IL-4 and IL-13 as well as that of IL-10 was also reduced , indicating that GATA-3 is a major orchestrator of coordinated TH2 response [13] . Furthermore , ectopic expression of GATA-3 induced TH2 cytokine expression in both differentiating and irreversibly committed TH1 cells [56] . Therefore , it is established that change in GATA-3 expression directly affects TH2 differentiation . Thus , in the presence of the inhibitor of Wnt signalling ( Dkk1 ) or upon knockdown of β-catenin , TH2 differentiation seems to be affected due to the downregulation of GATA-3 . We therefore propose that SATB1 mediates TH2 differentiation by regulating GATA-3 expression and thereby the expression of signature TH2 cytokines in a Wnt/β-catenin-dependent manner . Interestingly , a recent study reported that nephric duct-specific inactivation of GATA-3 leads to massive ectopic ureter budding , suggesting that GATA-3 acts downstream of β-Catenin signalling to prevent ectopic metanephric kidney induction [57] . A recent study using TCF-1 and β-catenin deficient mice established that TCF-1 initiates TH2 differentiation of activated CD4+ T cells by promoting GATA-3 expression and suppressing IFN-γ expression [58] . Thus , there is increasing evidence that GATA-3 is regulated by Wnt signalling . Our study also provides a molecular mechanism for regulation of GATA-3 expression by the SATB1:β-catenin complex in a Wnt signalling-dependent manner . ChIP analysis of an SBS within the GATA-3 promoter in differentiating TH2 cells revealed a time-dependent increase in occupancy of SATB1 and β-catenin . Upon silencing of SATB1 , its occupancy was expectedly decreased , but remarkably , even that of β-catenin was decreased by several-fold , suggesting that SATB1 is the key factor mediating the recruitment of β-catenin on GATA-3 and presumably other targets . The recruitment of the SATB1:β-catenin complex on GATA-3 promoter is sensitive to Dkk1 treatment . Moreover , since β-catenin recruits p300 on SATB1 genomic targets , the occupancy of p300 is also reduced concomitantly as that of β-catenin , presumably leading to the observed downregulation of GATA-3 upon inhibition of Wnt signalling by Dkk1 treatment . Although control TH0 cells also show a similar Dkk1 sensitive pattern of occupancy of these three factors on the GATA-3 promoter SBS , their fold occupancy is significantly more in the TH2 cells . The occupancy of the SATB1:β-catenin complex is very low at a non-SBS located upstream on the GATA-3 promoter and is not sensitive to Dkk1 . As with multiple other genes , the differential occupancy of various factors on different regions of the same promoter seems to be the mechanism governing regulation of transcription for GATA-3 . The intriguing observation is that such differential recruitment is dependent on Wnt signalling . Thus , our results provide a molecular mechanism towards understanding the role of Wnt signalling in TH2 differentiation and define a unique role of SATB1 in this process .
Antibodies to β-catenin and SATB1 were purchased from BD Biosciences . Antibodies to PCAF and pan-acetyl lysine were from Santa Cruz Biotechnology; antibodies to N-terminus of β-catenin , VP-16 , and H3K9ac were obtained from Upstate Technologies . Anti-Flag was procured from Sigma-Aldrich Corp . siSATB1 and siβ-catenin duplex RNAs were from Santa Cruz Biotechnology , and BIO was obtained from Calbiochem . Recombinant Wnt3a , Dkk1 , and cytokines were obtained from R&D Systems . Anti-SATB1 used for IP and WB was raised in rabbit and was purified using immunoaffinity chromatography using standard procedures . Expression constructs for β-catenin mutants T41A and GST 1–12 arm repeats were kindly provided by Dr . C . Neuveut , and pGEX6P3-β-catenin was gifted by Dr . M . Dunach . Deletion construct pGEX6P3-β-catenin-1–535 was generated by digesting pGEX6P3-β-catenin with SpeI and XhoI and relegating the larger fragment into pGEX6P3 . pGEX6P3-β-catenin was digested with EcoRI , and the 1 Kb fragment was subcloned into pGEX6P2 to obtain pGEX6P2-β-catenin-425–781 . RFP-fused versions of truncated β-catenin were produced by subcloning regions of β-catenin cDNA corresponding to aa 1–576 and 577–780 in pDsRed Express C1 vector ( Takara Clontech ) . Various VP16-β-catenin truncations and Gal4DBD:SATB1 , and Gal4DBD:PDZ fusions were cloned in pACT and pBIND plasmids ( Promega ) , respectively . TOP and FOP reporter constructs were kind gifts by Dr . R . T . Moon . pTriEx-SATB1 , various SATB1 domain constructs , and the reporter constructs pGL3-Basic-IgH MAR and pGL3-Basic-IL-2 have been described previously [17] , [21] . Thymocytes were isolated from 3-wk-old Balb/C mice . Thymocytes were fixed using 2% paraformaldehyde and were permeabilized with 0 . 1% Triton X-100 followed by antibody staining of SATB1 and β-catenin ( BD Biosciences ) . The secondary antibodies used were conjugated to Alexafluor dyes 488 and 594 ( Invitrogen ) . DNA counterstaining was performed using DAPI . Cells were visualized under an upright fluorescence microscope ( model AxioImager Z1 , Carl Zeiss ) , and digital images were enhanced using the Apotome module ( Carl Zeiss ) . The sectional views of the stained cells revealing the signals from the nuclear interior were generated using the Axiovision software ( Carl Zeiss ) . GST:1–178 , GST:1–535 , GST:425–781 , and GST:Arm repeats of β-catenin and GST:PDZ , GST:CD+HD , GST:CD , and GST:HD of SATB1 were expressed and purified from Escherichia coli ( E . coli ) . Specific fusion proteins were cleaved on column by caspase-6 to obtain GST-free proteins as described [59] . Intact GST-fusion proteins were incubated with glutathione Sepharose 4B beads ( GE Healthcare ) for 2 h . 35S-labeled SATB1 was prepared by coupled in vitro transcription and translation as per manufacturer's instructions ( Promega ) . Full-length SATB1 was expressed as 6XHis-SATB1 in E . coli and was bound to Ni-NTA beads ( Qiagen ) . Bound proteins were washed with 1× PBS and incubated either with affinity purified recombinant proteins or with 100 µg nuclear extract from Jurkat or SW 480 cells or with in vitro transcribed and translated SATB1 for 2–3 h . The complexes were then washed with 1× PBS containing 0 . 1% triton X-100 and eluted using 2× SDS-loading buffer , resolved by SDS-polyacrylamide gel electrophoresis and analyzed by immunoblotting . HEK cell line 293T , colon cancer cell line SW480 , and T cell lymphoblastoid Jurkat cells were cultured in DMEM or in RPMI 1640 ( Invitrogen ) , respectively , supplemented with 10% FCS ( Invitrogen ) . Cells were transfected with indicated expression constructs using Lipofectamine 2000 ( Invitrogen ) . After transfections cells were cultured for 24 h and then either LiCl or BIO were added to media at final concentrations of 20 mM or 1 . 0 µM , respectively , and cells were cultured further for 24 h unless mentioned otherwise . Transactivation assays were performed after 48 h of transfections as described [17] . Briefly , cells were harvested and washed with 1×PBS , followed by lysis with Luclite reagent ( Perkin Elmer ) as per manufacturer's instructions . Luciferase counts were measured using TopCount ( Packard ) and plotted as relative activity using Sigmaplot ver . 10 . For monitoring Wnt/TCF activity , the TOPFlash and FOPFlash reporter assay system was used . HEK 293T cells or naïve CD4+ T cells were cotransfected with the indicated expression vectors and the TOPFlash or FOPFlash reporter constructs . Equal masses of DNA were used in each transfection , and vector DNAs were used for normalization of DNA if required in each transfection reaction . The ratio of luciferase activities in TOPFlash-transfected versus FOPFlash-transfected cells was calculated and plotted as the relative TCF activity . All reporter assays were repeated at least thrice using independently transfected HEK 293T cells unless mentioned otherwise , and bar graphs represent values +/− s . d . The statistical significance of observed differences was calculated by t test . Five hundred µg of nuclear extract was diluted three times using 1× PBS containing 0 . 1% Triton X-100 and was precleared using 1 . 0 µg of either rabbit or mouse IgG and protein A/G beads ( Pierce ) . Precleared lysate was subjected to immunoprecipitation by incubating with anti-SATB1 or anti-β-catenin for 2 h; protein A/G beads were then added and mixed further by incubating on an end-to-end rotator at 4°C for 4 h . Protein complexes were analyzed by immunoblotting with respective antibodies . ChIP was performed essentially as described [17] . Briefly , cells were crosslinked by addition of formaldehyde to 1% final concentration in media and incubation at room temperature for 10 min , neutralized with 125 mM glycine , and then subjected to sonication using Bioruptor ( Diagenode ) to fragment the chromatin to obtain 200–500 bp fragments . Sonicated chromatin was precleared with a cocktail containing 50% protein A/G beads slurry ( Pierce ) , Salmon sperm DNA , and BSA . Precleared chromatin was incubated with specific antibodies and respective Ig types were used as isotype controls . Protein A/G bead cocktail was then added to pull down the antibody-bound chromatin and was subjected to elution using sodium biocarbonate buffer containing SDS and DTT . Eluted chromatin was de-crosslinked and protein was removed by treating with proteinase K . Purified immunoprecipitated chromatin was subjected to PCR amplification using specific primers . Input chromatin was used as a control . Quantitative PCRs were performed using iCycler ( BioRad ) using iQ SYBR Green mix ( Bio Rad ) . ΔCt values were calculated using the formula: ΔCt = ( Ct Target−Ct Input ) . Fold differences in gene expression were calculated as follows: Fold difference = 2 ( Treatment−Loading Control ) /2 ( Control−Loading Control ) . Fold changes were calculated by dividing the level of expression of the experimental sample with the corresponding control sample . This study was conducted according to the principles expressed in the Declaration of Helsinki . Studies involving human samples were approved by the institutional ethics committee for medical research , and tissues were obtained according to the guidelines of the committee . All patients provided written informed consent for the collection of samples and subsequent analysis . Thymic tissue was obtained from local hospitals from children undergoing cardiac surgery who did not have any medical history of immunological abnormalities . The tissue was minced into fine pieces , and thymocytes were collected by grinding on tissue sieve in RPMI 1640 media . The thymocyte suspension was passed through a 70 µm sieve to remove clumps and debris . The thymocytes obtained in this manner were devoid of thymic epithelial cells as confirmed by lack of staining by epithelial/stromal cell markers EPCAM and UEA-1 . Thymocytes from 3–4 donors were pooled to reduce individual variation . Peripheral human mononuclear cells were isolated from the human umbilical cord blood procured from local hospitals using Ficoll-Paque ( GE Healthcare ) . CD4+ T cells were selectively isolated from PBMCs using human CD4+ T cell isolation ( BD Biosciences ) as per manufacturer's instructions . CD4+ T cells were cultured in RPMI 1640 media supplemented with 10% FCS ( Invitrogen ) . Cells were cultured in 24 well plates precoated with 0 . 5 µg/ml of anti-CD3 and in the presence of 0 . 5 µg/ml soluble anti-CD28 ( eBiosciences ) . TH1 polarization was induced using 2 . 5 ng/ml IL-12 and TH2 polarization with 10 ng/ml IL-4 ( R&D Systems ) . After 48 h of polarization , 17 ng/ml of IL-2 ( R&D Systems ) was added . Polarized cells were grown further for various time points from 24 to 72 h . Supernatants from the control and treated TH cells were collected by centrifugation . Levels of various cytokines in these supernatants were measured using the Bio-Plex TH1-TH2 kit and the Bio-Plex protein array reader ( Bio-Rad ) as per manufacturer's instructions . Levels of secreted cytokines were normalized against the respective media . Isolated T cells were differentiated ex vivo as described [26] . Thymocytes and CD4+ T cells were electroporated using the Nucleofector device ( Amaxa ) according to manufacturer's instructions . Cells were cultured in RPMI 1640 media supplemented with 10% FCS for 24 h . After 24 h cells were activated and polarized as described above . Wnt signalling was induced by addition of purified recombinant Wnt3a ( R&D Systems ) at 50 ng/ml or inhibited by addition of Dkk1 ( R&D Systems ) at 400 ng/ml . After 72 h of polarization in presence or absence of Dkk1 , cells were harvested and divided into three parts . Chromatin was prepared from two portions of cells as described above . Total RNA was isolated from the remaining one third portion of cells , and cDNA was prepared using the Cells-to-Signal kit ( Ambion ) as per the manufacturer's instructions .
|
In vertebrates the canonical Wnt signalling culminates in β-catenin moving into the nucleus where it activates transcription of target genes . Wnt/β-catenin signalling is essential for the thymic maturation and differentiation of naïve T cells . Here we show that SATB1 , a T cell lineage-enriched chromatin organizer and global regulator , binds to β-catenin and recruits it to SATB1's genomic binding sites so that genes formerly repressed by SATB1 are upregulated by Wnt signalling . Some of the genes known to be regulated by SATB1 ( such as genes encoding cytokines and the transcription factor GATA3 ) are required for differentiation of Th2 cells , an important subset of helper T cells . Specifically we show that siRNA-mediated knockdown of SATB1 downregulated GATA-3 expression in differentiating human CD4+ T cells . Inhibiting Wnt signalling led to downregulation of GATA-3 and of signature TH2 cytokines such as IL-4 , IL-10 , and IL-13 . Knockdown of β-catenin also produced similar results , thus together these data confirm the role of Wnt/β-catenin signalling in TH2 differentiation . Our data demonstrate that SATB1 orchestrates TH2 lineage commitment by modulating Wnt/β-catenin signalling .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"molecular",
"biology/histone",
"modification",
"cell",
"biology/nuclear",
"structure",
"and",
"function",
"cell",
"biology/cell",
"signaling",
"immunology/leukocyte",
"signaling",
"and",
"gene",
"expression",
"cell",
"biology/gene",
"expression"
] |
2010
|
Global Regulator SATB1 Recruits β-Catenin and Regulates TH2 Differentiation in Wnt-Dependent Manner
|
In eukaryotes , the degradation of cellular mRNAs is accomplished by Xrn1 and the cytoplasmic exosome . Because viral RNAs often lack canonical caps or poly-A tails , they can also be vulnerable to degradation by these host exonucleases . Yeast lack sophisticated mechanisms of innate and adaptive immunity , but do use RNA degradation as an antiviral defense mechanism . We find a highly refined , species-specific relationship between Xrn1p and the “L-A” totiviruses of different Saccharomyces yeast species . We show that the gene XRN1 has evolved rapidly under positive natural selection in Saccharomyces yeast , resulting in high levels of Xrn1p protein sequence divergence from one yeast species to the next . We also show that these sequence differences translate to differential interactions with the L-A virus , where Xrn1p from S . cerevisiae is most efficient at controlling the L-A virus that chronically infects S . cerevisiae , and Xrn1p from S . kudriavzevii is most efficient at controlling the L-A-like virus that we have discovered within S . kudriavzevii . All Xrn1p orthologs are equivalent in their interaction with another virus-like parasite , the Ty1 retrotransposon . Thus , Xrn1p appears to co-evolve with totiviruses to maintain its potent antiviral activity and limit viral propagation in Saccharomyces yeasts . We demonstrate that Xrn1p physically interacts with the Gag protein encoded by the L-A virus , suggesting a host-virus interaction that is more complicated than just Xrn1p-mediated nucleolytic digestion of viral RNAs .
Degradation of mRNAs is a process essential to cell viability . Degradation pathways eliminate aberrant mRNAs , and also act to control gene expression levels . This process typically begins with host enzymes that perform either deadenylation or decapping on mRNAs targeted for degradation [1] . Following decapping , mRNAs are typically degraded by the 5’ to 3’ cytoplasmic exonuclease , Xrn1 [2 , 3] . Alternatively , after deadenylation , mRNAs can be subject to 3’ to 5’ degradation by the cytoplasmic exosome [4–6] . Viral transcripts and viral RNA genomes usually do not bear the canonical 5’ methylated cap structures or the 3’ polyadenylated ( poly ( A ) ) tails typical of cellular mRNAs , making them vulnerable to destruction by these host mRNA degradation pathways . In fact , it has been observed that Xrn1 and components of the exosome efficiently restrict virus replication in eukaryotes as diverse as mammals and yeasts [7–11] . As a result , mammalian viruses have evolved diverse countermeasures to prevent degradation by these proteins [7 , 8 , 12–18] . Still unknown is whether the host proteins like Xrn1 and components of the exosome can co-evolve with viruses to circumvent viral countermeasures . While such tit-for-tat evolution is common in mammalian innate immunity pathways , mRNA degradation is essential to the host and would be expected to be subject to strong evolutionary constraint . Saccharomyces yeasts are known to harbor very few viruses [19] . Further , all yeast viruses are unable to escape their host cell , and instead are transmitted through mating or during mitotic cell division . Almost all described species of Saccharomyces yeasts play host to double-stranded RNA ( dsRNA ) viruses of the family Totiviridae [20 , 21] . In fact , most commonly used S . cerevisiae laboratory strains are infected with a totivirus named L-A ( Fig 1A ) [22] . When initially synthesized , the RNAs produced by the L-A virus RNA-dependent RNA polymerase lack both a cap structure [23 , 24] and a poly ( A ) tail [25] , and are vulnerable to degradation by yeast Xrn1 ( denoted Xrn1p ) [8] and the cytoplasmic exosome [14 , 26] . 3’-to-5’ degradation of viral RNAs by the cytoplasmic exosome is linked to the action of the SKI complex ( Ski2 , Ski3 , Ski7 , and Ski8 ) , which acts to funnel aberrant RNAs into the nucleolytic core of the exosome [5 , 6] . The disruption of exosome and SKI complex genes has been shown to cause higher expression of viral RNAs , higher virus genome copy number , and an overproduction of virus-encoded toxins ( i . e . the “superkiller” phenotype ) [11 , 14 , 27] . In addition , the 5’-to-3’ exonuclease Xrn1p degrades viral transcripts and genomes of several RNA viruses in yeasts [8 , 24 , 28] . Viruses and their hosts exist in a constant state of genetic conflict , where what is advantageous for one party is often disadvantageous for the other . Both genomes experience selection for mutations that benefit their own fitness but , particularly in yeast where viruses are strictly intracellular , the virus will be bounded in this process by the fact that if it begins to replicate too well , it may kill its host . Co-evolutionary battles between hosts and viruses play out in the physical interaction interfaces between interacting host and viral proteins ( reviewed in [30–32] ) . One party is selected to reduce these interactions , and the other party is selected to strengthen them . For instance , there are many examples showing that mammalian restriction factors are selected to better recognize their viral targets , while viruses are continuously selected to escape that interaction , or to encode an antagonist protein that neutralizes the restriction factor . Because there is often no stable equilibrium in these systems , this process of tit-for-tat evolution between host and virus can cycle over and over , causing unusual signatures of evolution in both the host and virus proteins engaged in this interaction . While host protein complexes ( host proteins interacting with other host proteins ) can sometimes become co-evolved , this process of within-species refinement of protein-protein interactions is not the same as the dynamic and recurrent selection for new amino acids at interaction interfaces between host and pathogen proteins . The two scenarios can be disentangled using a metric that looks for codons that have accumulated a significantly higher rate of nonsynonymous mutations ( dN ) than even synonymous mutations ( dS ) . The signature of dN/dS > 1 commonly results from the repeated cycles of selection that occur in genetic conflict scenarios [33] , but has not been shown to be driven by subtler processes like the refinement of within-host physical interactions . Highly diverged host proteins reinforce species barriers , making it difficult for viruses to move from their current host species into new host species ( for example , [34 , 35] ) . Since yeast have limited antiviral strategies , we reasoned that evolutionary pressure on the RNA quality control pathways to thwart the replication of RNA viruses might be especially intense . This led us to investigate the unique evolutionary scenario involving a restriction system employing proteins critical to RNA turnover and cellular homeostasis . In this study , we analyzed whether or not any components of the yeast RNA degradation pathways mentioned above are evolving under positive natural selection , potentially indicative of tit-for-tat coevolution with viruses . We identified this evolutionary signature in at least two genes involved in RNA metabolism , RRP40 and XRN1 , and then undertook an in depth functional analysis of XRN1 . To test the hypothesis that Xrn1p has been honed by co-evolution to target and restrict totiviruses , we made a series of S . cerevisiae strains where XRN1 is replaced with wild-type orthologs from other Saccharomyces species . All XRN1 orthologs fully complemented an XRN1 knockout strain of S . cerevisiae , as assessed by several assays . On the other hand , we found that XRN1 orthologs were different in their ability to control the replication of the L-A virus . Xrn1p from S . cerevisiae was most efficient at controlling the L-A virus that chronically infects S . cerevisiae , and Xrn1p from S . kudriavzevii was most efficient at controlling the L-A-like virus ( SkV-L-A1 ) that we discovered within S . kudriavzevii . All XRN1 orthologs were equivalent in their interaction with another virus-like parasite , the Ty1 retrotransposon . Our identification of signatures of positive selection and species-specific virus restriction suggests that XRN1 can be tuned by natural selection to better restrict totivirus in response to the evolution of these viruses over time . We show that the structure of Xrn1p affords the flexibility to change in response to selective pressure from totiviruses , while also maintaining cellular functions .
We first looked for evidence of positive selection ( dN/dS > 1 ) within the genes encoding the major components of the SKI complex , the exosome , and Xrn1p ( Fig 1B ) . Importantly , signatures of positive selection do not identify the genes that are most important for controlling viral replication . Rather , these statistical tests are designed to identify host proteins that are involved in direct physical interactions with viruses , and which also have the evolutionary flexibility to change in response to viral selective pressure , becoming species-specific in the process . For this reason , we would neither expect to identify signatures of positive selection in all genes known to be involved in controlling totiviruses , nor in all genes encoding components of a complex like the exosome . For each gene , we collected sequences from six divergent species of Saccharomyces ( S . cerevisiae , S . paradoxus , S . mikatae , S . kudriavzevii , S . arboricolus , and S . bayanus ) [36–38] and created a multiple sequence alignment . We then analyzed each alignment for evidence of codons with dN/dS > 1 using four commonly employed tests for positive selection [39 , 40] . We see some evidence for positive selection of specific codon sites in several of these genes , however , only XRN1 and the exosome subunit gene RRP40 passed all four tests ( Fig 1C and S1 Table ) . Other genes are determined to be under positive selection by some tests , and may be of interest to explore further . Of XRN1 and RRP40 , the impact of XRN1 on viral replication has been more directly substantiated [7–9 , 13–16 , 41–48] , so we focused our attention on this gene . However , it should be noted that RRP40 encodes a component of the cytoplasmic exosome , which , in conjunction with the SKI complex , is clearly linked to the restriction of L-A [11 , 14 , 27] . We next tested if S . cerevisiae XRN1 has been tailored by co-evolution with the L-A virus . Double-stranded RNA ( dsRNA ) purified from a S . cerevisiae xrn1Δ strain migrates as a distinct band of 4 . 6 kilobase pairs ( Fig 2A ) , which is consistent with the size of the L-A virus genome , and its identity was further confirmed by RT-PCR ( S1 Fig ) . We confirmed a strong reduction in dsRNA when the xrn1Δ strain was complemented with plasmid-mounted XRN1 from S . cerevisiae under the transcriptional control of its native promoter ( Fig 2A ) , consistent with the published role of Xrn1p as an L-A restriction factor [8 , 14 , 24 , 27] . On the other hand , catalytically-dead versions of Xrn1p ( E176G and Δ1206–1528 ) did not suppress L-A dsRNA levels ( Fig 2B ) , as has been previously described [49] . We next performed heterospecific ( other species ) complementation by introducing the XRN1 from S . mikatae , S . kudriavzevii , or S . bayanus into the S . cerevisiae xrn1Δ strain . These species were chosen as they are representative of the diversity found within the sensu stricto complex of Saccharomyces yeasts . Strikingly , no other Xrn1p was able to reduce L-A dsRNA to the same extent as Xrn1p from S . cerevisiae ( Fig 2A ) . Xrn1p from S . mikatae , the closest relative to S . cerevisiae in this species set , was capable of slightly reducing L-A dsRNA abundance . Xrn1p from S . bayanus and S . kudriavzevii appear to have levels of dsRNA similar to xrn1Δ , indicating little or no effect on L-A copy number . In summary , we find that XRN1 orthologs vary in their ability to restrict the S . cerevisiae L-A virus . This is somewhat surprising for a critical and conserved gene involved in RNA quality control , but consistent with the signatures of positive selection which suggest that certain parts of this protein are highly divergent between species . We next used a functional and quantitative assay to confirm the species-specific effects of XRN1 on virus replication . This assay exploits the dsRNA “killer virus” ( also known as M virus ) . The killer virus is a satellite dsRNA of L-A that is totally dependent on L-A proteins for replication . It uses L-A-encoded proteins to encapsidate and replicate its genome , and to synthesize and cap its RNA transcripts [12] . The killer virus encodes only a single protein , a secreted toxin referred to as the killer toxin [19 , 50 , 51] . The result is that “killer yeast” colonies , i . e . those infected with both L-A and the killer virus , kill neighboring cells via the diffusion of toxin into the surrounding medium ( Fig 2C ) . Importantly , resistance to the killer toxin is provided by the pre-processed , immature form of the toxin , supplying killer yeast cells with an antidote to their own poison [50] . It has been shown previously that Xrn1p can inhibit the expression of the killer phenotype by degrading uncapped killer virus RNAs [14 , 52] . Therefore , we use the presence and size of kill zones produced by killer yeasts as a quantitative measurement of killer virus RNA production in the presence of each Xrn1p ortholog . A strain of S . cerevisiae lacking XRN1 , but harboring both the L-A and killer virus ( xrn1Δ L-A+ Killer+ ) , was complemented with each XRN1 ortholog . Clonal isolates from each complemented strain were grown to mid-log phase , and 6 x 105 cells were spotted onto an agar plate seeded with a lawn of toxin-sensitive yeast . After several days’ incubation at room temperature , kill zones around these culture spots were measured and the total area calculated . The transformation of xrn1Δ L-A+ Killer+ with S . cerevisiae XRN1 produced an average kill zone that covered 0 . 68 cm2 ( n = 14 ) . However , transformation with XRN1 from S . mikatae , S . bayanus , or S . kudriavzevii produced significantly larger kill zones covering 0 . 92 cm2 ( n = 11 ) , 0 . 96 cm2 ( n = 17 ) and 0 . 97 cm2 ( n = 17 ) , respectively . The kill zone produced by xrn1Δ L-A+ Killer+ yeast expressing S . cerevisiae XRN1 was significantly smaller than those produced by yeast expressing any of the other XRN1 orthologs ( Tukey—Kramer test , p<0 . 05 ) ( Fig 2D ) . The smaller kill zones in the strain expressing S . cerevisiae XRN1 are consistent with lower levels of killer and L-A derived RNAs . In summary , this assay also supports a species-specific restriction phenotype for XRN1 . It has been observed that over-expression of XRN1 can cure S . cerevisiae of the L-A virus , presumably by degrading viral RNA so effectively that the virus is driven to extinction [8 , 28] . Therefore , we developed a third assay to test the ability of XRN1 orthologs to control L-A , in this case by assessing their ability to cure S . cerevisiae of the virus . Plasmids expressing HA-tagged and untagged Xrn1p were transformed into a killer strain of S . cerevisiae with its genomic copy of XRN1 intact . This was followed by the analysis of more than 100 purified clones for virus curing , that is , the absence of the killer phenotype as indicated by the loss of a kill zone when plated on a lawn of sensitive yeast . Importantly , the introduction of an empty plasmid fails to produce any cured clones ( n = 103 ) ( Fig 3A and 3B ) . Provision of an additional copy of S . cerevisiae XRN1 cured 49% of clones ( n = 159 ) ( Fig 3A and 3B ) . Cured clones remained cured ( i . e . non-killers ) when purified and tested again for their ability to kill sensitive yeasts ( n = 20 ) . Over-expression of XRN1 from S . mikatae , S . kudriavzevii , and S . bayanus was unable to efficiently cure the killer phenotype , resulting in only 12% ( n = 129 ) , 8% ( n = 120 ) , and 9% ( n = 123 ) cured clones , respectively ( Fig 3A , blue bars ) . The loss of L-A from cured strains was also verified by RT-PCR . We detected no L-A or killer RNAs within the four cured clones analyzed ( Fig 3C ) . These data show that XRN1 from all Saccharomyces species have the ability to cure the killer phenotype , however , XRN1 from S . mikatae , S . kudriavzevii , and S . bayanus are considerably less efficient than S . cerevisiae XRN1 . Taken together , we show that viral restriction by XRN1 is species-specific . These data are consistant with a model where viral restriction can be refined through sequence evolution in XRN1 . We next tested the presumption that the XRN1 orthologs are functionally equivalent for cellular processes when expressed within S . cerevisiae . We first confirmed that XRN1 orthologs successfully complemented the severe growth defect of S . cerevisiae xrn1Δ , by measuring the doubling time of S . cerevisiae xrn1Δ with or without a complementing XRN1-containing plasmid ( Fig 4A ) . The knockout of XRN1 also renders cells sensitive to the microtubule-destabilizing fungicide benomyl [49] , and we observed that all XRN1 homologs convey equal resistance to benomyl on solid medium ( Fig 4B ) . It has been previously reported that over-expression of XRN1 is toxic to S . cerevisiae , a phenotype that has been suggested to be due to a dominant negative interaction of Xrn1p with other essential cellular components , such as the decapping complex [49] . Growth upon medium containing 2% galactose was equivalently reduced for strains carrying GAL1 inducible XRN1 genes from each species , whereas the strain over-expressing GFP grew normally ( Fig 4C , right ) . Finally , the Ty1 retrotransposon is another intracellular virus that replicates within Saccharomyces species and often co-exists with L-A within the same cell . Interestingly , Xrn1p is not a restriction factor for Ty retrotransposons , but rather promotes their replication [43 , 47 , 48 , 53–57] . We found no significant difference between the mean values for retrotransposition in the presence of Xrn1p from S . cerevisiae , S . mikatae , S . kudriavzevii , or S . bayanus ( one-way ANOVA , F3 , 8 = 0 . 36 , p = 0 . 78 ) , indicating that the evolutionary differences between divergent XRN1 genes do not affect the ability of Ty1 to replicate within S . cerevisiae ( Fig 4D ) . Collectively , these data indicate the cellular functions of Xrn1p have remained unaffected during yeast speciation , while the interaction with L-A viruses has changed . We next mapped the region responsible for the species-specific restriction by XRN1 . To better understand the structural organization of Xrn1p from S . cerevisiae , we used Phyre [58] to generate a template-based homology model of the exonuclease using the solved structure of Kluyveromyces lactis Xrn1p ( Fig 5A ) . A linker region within the N-terminal domain , the far C-terminal domain , and domain D2 were not included in the model as there is a lack of information regarding the structural organization of these regions . Importantly , modeled domains contained three of the residue positions that we identified as evolving under positive selection ( S1 Table ) , and all of these ( blue ) fall in and near the D1 domain ( orange ) ( Fig 5A ) . As expected because of the selection that has operated on them , these residue positions under positive selection are more variable in sequence between species than are surrounding residues ( two are shown in Fig 5B ) . All residues under positive selection are surface exposed and are far from the highly conserved Xrn1p catalytic domain ( 96% identity , across the Saccharomyces genus ) and catalytic pocket ( red ) . The other sites of positive selection fall within the last 500 amino acids of Xrn1p , which is less conserved compared to the rest of the protein ( 83% identity , across the Saccharomyces genus ) ( Fig 5C ) . To define the importance of the two regions that we identified as containing signatures of positive selection , we replaced portions of S . kudriavzevii XRN1 with the equivalent portions of S . cerevisiae XRN1 , and assayed for a region of S . cerevisiae XRN1 that would convey the ability to cure the killer phenotype . We found that an XRN1 chimera encoding the last 775 amino acids from S . cerevisiae ( Sc-775 ) was sufficient to cure 56% of clones analyzed , and this was very similar to S . cerevisiae XRN1 ( 57% ) ( Fig 5D ) . Conversely , when the last 777 amino acids from S . kudriavzevii ( Sk-777 ) were used to replace the same region within S . cerevisiae XRN1 , only 9% of clones were cured ( S3 Fig ) . This focused our construction of further chimeras to the second half of the protein , which also contains all of the codons under positive selection and has less amino acid conservation between S . cerevisiae and S . kudriavzevii ( 82% protein identity , compared to the N-terminius of Xrn1p with 95% identity ) . Initial analysis of the highly diverged C-terminal tail revealed that the last 461 amino acids of S . cerevisiae Xrn1p were unable to convey efficient L-A restriction to S . kudriavzevii Xrn1p ( S3 Fig ) . For this reason , we focused further chimeric analysis on the region encompassing the D1 , D2 , and D3 domains [59] . We swapped into S . kudriavzevii Xrn1p the D2+D3 , D1 , or D1-D3 domains of S . cerevisiae Xrn1p , and saw increasing rescue of the ability to cure the L-A virus ( Fig 5D ) . All chimeric XRN1 genes were functionally equivalent with respect to their cellular functions , as all were able to establish normal growth and benomyl resistance in S . cerevisiae xrn1Δ ( S4 Fig ) . Species-specific restriction maps predominantly to D1 , with contribution from the neighboring D2 and D3 domains . Together , our data suggest that the exonuclease activity of Xrn1p is important for virus restriction and is preserved across species , but that evolution has tailored a novel virus interaction domain ( D1-D3 ) that targets the enzymatic activity of Xrn1p against L-A in a manner that changes over time . It’s hard to imagine that Xrn1p proteins from different species are differentially recognizing viral RNA , since they are all equivalent in their host functionalities within S . cerevisiae . We considered the possibility that there might be host-virus interactions beyond Xrn1p and the viral RNA . It has been shown that Xrn1p targets uncapped viral RNA transcripts rather than affecting dsRNA propagation [52] . As totivirus transcription only occurs in the context of a fully-formed capsid [60] and capsids are assembled entirely from the L-A Gag protein [61] , it would seem plausible that Xrn1p may interact directly with Gag to target virus-derived uncapped RNAs . We introduced epitope tags onto Xrn1p ( HA-tag ) and the major capsid protein of L-A , Gag ( V5-tag ) , and expressed both tagged and untagged versions of each protein from plasmids introduced into S . cerevisiae xrn1Δ ( Fig 6A ) . Bead-bound antibodies specific for either HA or V5 were used to immunoprecipitate Xrn1p or Gag , respectively . We found that Gag ( V5-tagged ) was able to immunoprecipitate Xrn1p ( HA-tagged ) from S . cerevisiae and S . kudriavzevii ( Fig 6A , top panel ) . Reciprocally , Xrn1p-HA from both S . cerevisiae and S . kudriavzevii were able to immunoprecipitate Gag-V5 ( Fig 6A , bottom panel ) . The interaction between Xrn1p and Gag appears not to be mediated by single-stranded RNAs , as their digestion by RNase A in the whole cell extract did not affect the co-immunoprecipitation of Gag by Xrn1p ( S5 Fig ) . We next performed these experiments with a monoclonal antibody specific to L-A Gag , so that endogenous L-A Gag protein could be immunoprecipitated . This reaction co-immunoprecipitated both S . cerevisiae and S . kudriavzevii Xrn1p ( Fig 6B ) . Qualitatively , the relative efficiencies of Gag interaction with both S . cerevisiae and S . kudriavzevii Xrn1p appear similar in all assays , which seems at odds with our model that suggests that evolutionary differences within Xrn1p are a direct determinant of totivirus interaction . There are several possible interpretations . First , Gag might be antagonizing Xrn1p rather than being the species-specific target of Xrn1 . Second , there may be a third component in this interaction which makes manifest the species-specificity of Xrn1p . Finally , a trivial explanation could be that coimmunoprecipitations are not very quantitative , and maybe there is in fact a difference in interaction with Gag between the Xrn1p of different species . Nonetheless , these data demonstrate a previously undescribed interaction that goes beyond Xrn1p interaction with viral RNA and warrants careful in vitro study . We next wished to test our findings against other related yeast viruses . Indeed , the S . cerevisiae totivirus L-A-lus has been shown to have limited susceptibility to XRN1 from a different strain of S . cerevisiae [28] . We also wanted to test viruses of other species , but the only fully characterized totiviruses within the Saccharomyces genus are from S . cerevisiae . To identify totiviruses of other species , we screened Saccharomyces species from the sensu stricto complex for the presence of high molecular weight viral RNAs , and discovered a ~4 . 6 kbp dsRNA molecule within S . kudriavzevii FM1183 isolated from Europe ( Fig 7A ) [38] . We cloned the 4 . 6 kbp dsRNA molecule using techniques described by Potgieter et al . [62] and sequenced the genome of the virus using Sanger sequencing . We named the virus SkV-L-A1 ( S . kudriavzevii virus L-A isolate number 1; Genbank accession number: KX601068 ) . The SkV-L-A1 genome was found to be 4580 bp in length , with two open reading frames encoding the structural protein Gag and the fusion protein Gag-Pol ( via a -1 frameshift ) ( Fig 7B ) . Conserved features of totiviruses were identified and include a conserved catalytic histidine residue required for cap-snatching ( H154 ) , a -1 frameshift region , packaging signal , and replication signal ( Figs 7B and S6 ) . Phylogenetic analysis of the Gag and Pol nucleotide and protein sequences firmly places SkV-L-A1 within the clade of Saccharomyces totiviruses represented by L-A and L-A-lus [28 , 63] , as opposed to the more distantly related Saccharomyces totivirus L-BC ( Figs 7C and S6 ) [64] . To test the effect of XRN1 upon SkV-L-A1 , plasmids expressing XRN1 orthologs were introduced via LiAc transformation into S . kudriavzevii infected with SkV-L-A1 . These plasmids were able to express XRN1 from each species , although we find that the expression is variable , with S . mikatae Xrn1p expressing at a level higher than the others ( Fig 7D ) . The expression of these proteins did not affect the overall growth rate or colony morphology of S . kudriavzevii ( S7 Fig ) . Because of the lack of an observable killer phenotype in this strain ( likely because a killer toxin-encoding satellite dsRNA is not present ) , heterospecific XRN1 were expressed within S . kudriavzevii and analyzed for their ability to spontaneously cure SkV-L-A1 , as we did previously with L-A in S . cerevisiae ( Fig 3 ) . We did not observe any virus curing by any orthologs of Xrn1p , but believe that this could be because the high-copy plasmids that we used in this experiment in S . cerevisiae are unable to drive Xrn1p expression in S . kudriavzevii high enough to actually cure the virus . However , we have observed previously that Xrn1p can reduce the abundance of totivirus RNAs ( Fig 2A ) , so we further analyzed the XRN1-transformed clones of S . kudriavzevii for changes in SkV-L-A1 RNA levels using reverse transcriptase quantitative PCR ( RT-qPCR ) . Total RNA was extracted from clones of S . kudriavzevii and converted to cDNA using random hexamer priming . cDNA samples were amplified using primers designed to specifically target SkV-L-A1 GAG and the cellular gene TAF10 . The empty vector control was used as the calibrator sample , and TAF10 expression was used as the normalizer to calculate the relative amount of SkV-L-A1 RNAs present within each XRN1 expressing S . kudriavzevii cell line using the comparative CT method [65] . We found that expression of XRN1 from S . kudriavzevii ( n = 10 ) reduced the relative levels of SkV-L-A1 RNAs by 40% ( Fig 7E ) , even though this Xrn1p was expressed at the lowest levels ( Fig 7D ) . This is in contrast to XRN1 from S . mikatae ( n = 9 ) and S . bayanus ( n = 8 ) that only showed a 13% increase or 15% decrease in SkV-L-A1 RNAs , respectively . S . cerevisiae XRN1 was able to reduce SkV-L-A1 RNAs by 27% and is noteworthy due to the close evolutionary relationship between SkV-L-A1 and other L-A-like viruses from S . cerevisiae ( Figs 7C and S6 ) . These data suggest that Xrn1p is a species-specific restriction factor in different Saccharomyces yeasts , and that coevolution of totiviruses and yeasts has specifically tailored the potency of Xrn1p to control the replication of resident viruses within the same species .
In the Saccharomyces genus , Xrn1p , the SKI complex , and exosome are all important for controlling the abundance of totivirus RNAs . We find that XRN1 and the exosome component RRP40 are somewhat unique in their strong signatures of positive natural selection . We speculated that positive selection might be driven by selection imposed by totiviruses . As speciation occurs and viruses mutate in unique ways in each lineage , new allelic versions of these antiviral genes that enable better control of totivirus replication would experience positive natural selection . Indeed , we found this to be the case , with S . cerevisiae Xrn1p restricting the S . cerevisiae L-A virus better than any other ortholog of XRN1 , and S . kudriavzevii Xrn1p restricting S . kudriavzevii SkV-L-A1 virus the best . The exact nature of the host-virus protein-protein interaction that is driving this evolutionary arms race is not clear . To thwart XRN1 , the totiviruses are known to synthesize uncapped RNAs with an exposed 5’ diphosphate , which is a suboptimal substrate for Xrn1p-mediated decay [24] . Further , it has been shown that the totivirus Gag protein has a cap-snatching activity that cleaves off caps from host mRNAs and uses them to cap viral transcripts , protecting them from Xrn1p degradation [12 , 14] . We have found that Xrn1p interacts with L-A Gag , and that this interaction is not mediated by the presence of single-stranded RNAs . What remains unknown is whether Xrn1p is targeting Gag as part of the restriction mechanism , or whether Gag is targeting Xrn1p as a counter defense . As we did not observe an obvious species-specific differences in the interaction between Xrn1p and L-A Gag by coimmunoprecipitation , we cannot clearly define the observed role of sequence variation in Xrn1p . This may be because of the low sensitivity of our assay system , or because direct binding of Xrn1p by L-A Gag is ubiquitous and that the rapid evolution of XRN1 results from another intriguing facet of virus-host interaction and antagonism . However , we now know that the interaction between L-A and Xrn1p goes beyond the simple recognition of L-A RNA by Xrn1p . We can speculate that Xrn1p may compete with Gag for access to uncapped viral RNAs as they are extruded into the cytoplasm , or that interaction with unassembled Gag allows the recruitment of Xrn1p to sites of virion assembly resulting in viral RNA degradation . Alternately , it is possible that the target of Xrn1p is simply L-A RNA , and that the interaction with Gag reflects a viral countermeasure where Gag is redirecting or otherwise altering the availability of Xrn1p to degrade L-A RNA . Indeed , there are several examples of mammalian viruses that redirect or degrade Xrn1p to aid in their replication [17 , 18 , 66] . The literature suggests that Xrn1p is a widely-utilized restriction factor against viruses , as it has been reported to have activity against mammalian viruses [9 , 16] , yeast viruses [8 , 24] , and plant viruses [46] . The potent 5’-3’ exonuclease activity of Xrn1p has resulted in viruses developing a rich diversity of strategies to protect their RNAs . For instance , Hepatitis C virus recruits MiR-122 and Ago2 to its 5’ UTR to protect its RNA genome from Xrn1p degradation [7 , 16] . The yeast single-stranded RNA narnavirus uses a different strategy to protect its 5’ terminus , folding its RNA to form a stem-loop structure that prevents Xrn1p degradation [8] . In some cases , viruses even depend on Xrn1p to digest viral RNA in a way that benefits viral replication , for example , preventing the activation of innate immune sensors [41] . Flavivirus ( West Nile and Dengue virus ) genomes also encode RNA pseudoknot and stem-loop structures that arrest the processive exonuclease activity of Xrn1p , producing short subgenomic flavivirus RNAs ( sfRNAs ) that are important for viral pathogenicity [13 , 67] . Members of the Flaviviridae , Herpesviridae , Coronaviridae , and yeast Totiviridae have all been shown to encode proteins that initiate endonucleolytic cleavage of host mRNAs , revealing exposed 5’ monophosphates that are substrates for Xrn1p degradation . This is thought to interfere with host translation and to produce uncapped RNA “decoys” that potentially redirect Xrn1p-mediated degradation away from viral RNA [14 , 15] . Xrn1p degradation , Xrn1p relocalization , virus-encoded capping enzymes , cap-snatching mechanisms , RNA-protein conjugation , recruitment of host micro-RNAs , cleavage of host mRNAs as “decoys” , and viral RNA pseudoknots are all utilized to prevent Xrn1p-mediated viral RNA destruction [7 , 8 , 12–18] . All of this evidence suggests that viruses can employ various methods to escape or harness the destructive effects of Xrn1 . Our data now suggests that Xrn1p in yeast is not a passive player in the battle against viruses , but rather that hosts can be selected to encode new forms of Xrn1p that can overcome virally encoded defense strategies . To rationalize the model of an antagonistic relationship between L-A and Saccharomyces species , it is important to consider the fitness burden of strictly intracellular viruses . Prevailing wisdom assumes that infection of fungi by viruses is largely asymptomatic and benign , especially when considering that their intracellular lifecycle ensures an evolutionary dead-end if they kill or make their host unfit . Indeed , within laboratory yeast strains , the association between L-A and S . cerevisiae appears to be at equilibrium , with no major biological differences between strains infected or not infected by L-A [68] . Therefore , the relationship between L-A and the Saccharomyces yeasts could be viewed as mutualistic or even commensalistic [68–70] . Mutualism is particularly striking in the context of the L-A / killer virus duo that provides the host cell with the “killer” phenotype , a characteristic that is broadly distributed throughout fungi [71] . If an infected yeast cell can kill other yeasts around it using the killer toxin , it no longer has to compete for resources within that environmental niche , an evolutionarily advantageous situation [51 , 69 , 70] . Indeed , there are other examples of host-virus mutualism in fungi [72 , 73] . However , there are many observations that lead one to believe that the relationship between intracellular viruses and their hosts is not benign and static . Firstly , there is a measurable fitness cost to killer toxin production by S . cerevisiae within unfavorable environmental conditions that inactivate the toxin , allow for regular cellular dispersal and/or are nutrient rich [69 , 70] . Secondly , virus infection of pathogenic fungi can also cause hypovirulence ( a reduction in fungal pathogenesis ) , an outcome that is being exploited to treat agricultural disease [74–77] . Thirdly , many wild and domesticated strains of S . cerevisiae are free of totiviruses ( and therefore also of killer ) , suggesting that there is selection against the ongoing maintenance of these viruses [20 , 28 , 71] . Fourthly , the continued maintenance of RNAi systems in fungi also correlates with the loss of the killer phenotype and is known to antagonize fungal viruses [71 , 78] . However , a virus of the fungi Cryphonectria parasitica has been shown to antagonize and escape restriction by RNAi without crippling its host [78] . This antagonistic relationship appears similar to the equilibrium of Saccharomyces yeasts and totiviruses , and suggests that in the absence of effective RNAi , additional antiviral defenses may be biologically relevant ( i . e . Xrn1p ) . In line with this view of a dynamic relationship between hosts and intracellular viruses , we show that totiviruses from different Saccharomyces species are best controlled by the Xrn1p of their cognate species , and that disruption of this equilibrium can result in excessive virus replication ( Fig 2 ) , virus loss ( Fig 3 ) , or a reduction in viral RNA ( Figs 2 and 7 ) . Signatures of positive selection that we have detected in Saccharomyces XRN1 are also consistent with a host-virus equilibrium that is in constant flux due to the dynamics of a back-and-forth evolutionary conflict ( Figs 2 and 6 ) . There are several examples of mammalian housekeeping proteins engaged in evolutionary arms races with viruses . ( By “housekeeping” we refer to proteins making critical contributions to host cellular processes , as opposed to proteins dedicated to immunity . ) In most of these other examples though , the housekeeping protein is hijacked by viruses to assist their replication in the cell ( rather than serving to block viral replication ) . For instance , many viruses hijack cell surface receptors to enter cells . We and others have shown that entry receptors are quite evolutionarily plastic , and that mutations can reduce virus entry without compromising host-beneficial functions of the receptor [34 , 79–83] . For example , the antagonistic interaction of Ebola virus ( and/or related filoviruses ) with the bat cell surface receptor , Niemann-Pick disease , type C1 ( NPC1 ) , has driven the rapid evolution of the receptor without affecting the transport of cholesterol , critical to the health of the host [34] . Numerous such examples highlight how essential housekeeping machineries , not just the immune system , are critical for protecting the cell from replicating viruses . This study highlights an interesting evolutionary conundrum that does not apply to classical immunity genes: as Xrn1p appears to be an antiviral protein , it must be able to evolve new antiviral specificities without compromising cellular health and homeostasis .
XRN1 from S . cerevisiae , including 1000 bp of the 5’ and 3’ UTRs , was amplified by PCR from genomic DNA prepared from S . cerevisiae S288C . This PCR product was cloned into the plasmid pAG425-GAL-ccdB by the “yeast plasmid construction by homologous recombination” method ( recombineering ) [84] to produce pPAR219 . Briefly , pAG425-GAL-ccdB was amplified by PCR to produce a 5000 bp product lacking the GAL-1 gene and the ccdB cassette . The PCR primers used to amplify pAG425-GAL-ccdB contained additional DNA sequence with homology to the UTRs of XRN1 from S . cerevisiae . Both PCR products were used to transform BY4741 , with correctly assembled plasmids selected for by growth on complete medium ( CM ) –leucine . The XRN1 open reading frame ( HA-tagged and untagged ) from S . mikatae , S . bayanus , or S . kudriavzevii was introduced into pPAR219 between the 5’ and 3’ UTRs from S . cerevisiae XRN1 using recombineering to produce pPAR225 , pPAR226 , and pPAR227 , respectively . As a negative control , NUP133 was cloned into the pPAR219 plasmid backbone to produce pPAR221 , which was used to allow growth of xrn1Δ on medium lacking leucine without XRN1 complementation . The LEU2 gene was replaced by TRP1 using recombineering techniques to produce the plasmids pPAR326 , pPAR327 , pPAR328 , and pPAR329 . Using PCR and recombineering , we also constructed chimeric XRN1 genes by exchanging regions of S . kudriavzevii XRN1 ( pPAR227 ) with the corresponding regions of S . cerevisiae XRN1 ( pPAR219 ) . XRN1 inducible plasmids were constructed by cloning PCR-derived XRN1 genes into pCR8 by TOPO-TA cloning ( Thermo Fisher ) . Utilizing Gateway technology ( Thermo Fisher ) , XRN1 genes were sub-cloned into the destination vector pAG426-GAL-ccdB for over-expression studies [85] . The same pCR8/Gateway workflow was also used to clone and tag GAG from a cDNA copy of the L-A totivirus ( pI2L2 ) to produce pPAR330 and pPAR331 . The DNA sequences from all constructed plasmids can be found in S2 File . A list of all relevant plasmids can be found in S2 Table . The S . cerevisiae killer strain ( BJH001 ) was created by the formation of a heterokaryon from the mating of the haploid strains BY4733 ( KAR1 ) and 1368 ( kar1 ) [86] . The resultant daughter heteroplasmon cells were selected by growth on CM—uracil and the ability to produce zones of growth inhibition indicative of the presence of L-A and the killer virus . The inability to grow on CM lacking histidine , leucine , tryptophan or methionine was also used to confirm the genotype of BJH001 . BJH006 was created by replacing XRN1 with the KANMX4 gene using homologous recombination within BJH001 [87] . A list of relevant yeast strains and species used in this study can be found in S3 Table . 1 x 109 yeast cells ( ~10 mL ) were harvested from a 24–48 h overnight culture grown to saturation . Strains of S . kudriavzevii and S . mikatae were grown at room temperature , all other strains were grown at 30°C . The flocculent nature of some strains of wild yeasts made it challenging to accurately determine the exact number of cells present in some cultures . In these cases , the size of the cell pellet was used as an approximate measure of cell number relative to S . cerevisiae . Harvested cells were washed with ddH2O , pelleted , and washed with 1 ml of 50 mM EDTA ( pH 7 . 5 ) . Cells were again harvested and the pellets suspended by vortexing in 1 ml of 50 mM TRIS-H2SO4 ( pH9 . 3 ) , 1% β-mercaptoethanol ( added fresh ) , and incubated at room temperature for 15 min . The cell suspension was centrifuged and the supernatant removed and the cell pellet suspended in 1 ml of BiooPure-MP ( a single-phase RNA extraction reagent containing guanidinium thiocyanate and phenol ) ( Bioo Scientific ) and vortexed vigorously . 200 μl of chloroform was added and vortexed vigorously before incubation for 5 min at room temperature . The aqueous phase and solvent phase were separated by centrifugation at 16 , 000 x g for 15 min at 4°C . The aqueous phase was transferred to a new tube and 1/3 volume of 95–100% ethanol added and mixed well by vortexing . The entire sample was loaded onto a silica filter spin column ( Qiagen plasmid miniprep kit ) and centrifuged for 30 s at 16 , 000 x g . The flow-through was discarded and the column washed twice with 750 μl of 100 mM Nacl/75% ethanol by centrifugation at 16 , 000 x g for 30 sec . The column was dried by centrifugation at 16 , 000 x g for an additional 30 sec . The dsRNA was eluted from the column by the addition of 100 μl of 0 . 15 mM EDTA ( pH 7 . 0 ) and incubation at 65°C for 5 min before centrifugation at 16 , 000 x g for 30 sec . dsRNA that was extracted from 1 x 109 yeast cells using our rapid extraction of viral dsRNA protocol was used as template for superscript two-step RT-PCR ( Thermo Fisher ) . cDNA was created using a primer specific for the negative strand L-A genomic RNA– 5’ CTCGTCAGCGTCTTGAACAGTAAGC . Primers 5’-GACGTCCCGTACCTAGATGTTAGGC and 5’-CTCGTCAGCGTCTTGAACAGTAAGC were used to specifically target and amplify cDNA derived from negative strand L-A virus RNAs using PCR with Taq ( New England Biolabs ) . The plasmid pI2L2 was used as a positive control for the RT-PCR reaction as it contains a cDNA copy of the L-A virus genome [88] . Alternatively , we collected total RNA from ~1 x 107 actively growing yeast cells using the RNeasy total RNA extraction kit ( Qiagen ) and synthesized cDNA using primers to target both the positive and negative strand of either L-A ( 5’-AAGATATTCGGAGTTGGTGATGACG and 5’-TCTCCGAAATTTTTCCAGACTTTATAAGC ) or killer virus ( 5’-GCGATGCAGGTGTAGTAATCTTTGG and 5’-AGTAGAAATGTCACGACGAGCAACG ) . The same primers were used to detect L-A and killer virus specific cDNAs using PCR with Taq polymerase ( New England Biolabs ) . We assayed Ty1 retrotransposition in S . cerevisiae xrn1Δ , using the previously described Ty1 retrotransposition reporter system [89] , and confirmed that XRN1 deletion causes a dramatic reduction in Ty1 retrotransposition ( ~50-fold ) [43] . To test the effect of XRN1 evolution on Ty1 replication , we introduced XRN1 from S . cerevisiae , S . mikatae , S . kudriavzevii , or S . bayanus into xrn1Δ and assayed Ty1 retrotransposition . Yeast lysates were prepared using the Y-PER reagent ( Thermo Fisher ) from 100 μl volume of log-phase yeast cells as per manufacturer’s instructions or by bead beating as described previously [90] . HA-tagged Xrn1p was detected via Western blot using a 1:5000 dilution of a horseradish peroxidase conjugated anti-HA monoclonal antibody ( 3F10—# 12013819001 ) ( Roche ) . Adh1p was detected using a 1:10000 dilution of rabbit polyclonal anti-alcohol dehydrogenase antibody Ab34680 ( Abcam ) . V5-tagged proteins were detected using a 1:5000 dilution of a mouse monoclonal antibody ( R960-25 ) ( Life Technologies ) . Native L-A Gag was detected using a 1:1000 dilution of a mouse monoclonal antibody ( gift from Nahum Sonenberg ) . Secondary antibodies were detected using ECL Prime Western Blotting Detection Reagent on a GE system ImageQuant LAS 4000 ( GE Healthcare Life Sciences ) . Nucleotide sequences from six species of Saccharomyces yeasts were obtained from various online resources , where available [36 , 38 , 91] . Maximum likelihood analysis of dN/dS was performed with codeml in the PAML 4 . 1 software package [39] . Multiple protein sequence alignments were created using tools available from the EMBL ( EMBOSS Transeq and Clustal Omega ) ( www . embl . de ) . Protein alignments were manually curated to remove ambiguities before processing with PAL2NAL to produce accurate DNA alignments [92] . DNA alignments were fit to the NSsites models M7 ( neutral model of evolution , codon values of dN/dS fit to a beta distribution , with dN/dS > 1 not allowed ) and M8 ( positive selection model of evolution , a similar model to M7 but with an additional site class of dN/dS > 1 included in the model ) . To ensure robustness of the analysis , two models of codon frequencies ( F61 and F3x4 ) and multiple seed values for dN/dS ( ω ) were used ( S1 Table ) . Likelihood ratio tests were performed to evaluate which model of evolution the data fit significantly better . Posterior probabilities of codons under positive selection within the site class of dN/dS > 1 ( M8 model of positive selection ) were then deduced using the Bayes Empirical Bayes ( BEB ) algorithm . REL and FEL analysis was carried out using the online version of the Hyphy package ( www . datamonkey . org ) S1 Table [40] . Analysis of XRN1 was performed using the TrN93 nucleotide substitution model and the following phylogenetic relationship ( Newick format ) : ( ( ( ( ( ( S . paradoxus-Europe , S . paradoxus-Far East ) , ( S . paradoxus-North America , S . paradoxus-Hawaii ) ) , S . cerevisiae ) , S . mikatae ) , S . kudriavzevii ) , S . arboricolus , S . bayanus ) ; GARD analysis found no significant evidence of homologous recombination within any dataset . MEGA6 was used to infer the evolutionary history of totiviruses using the Maximum Likelihood method . Appropriate substitution models were selected using manually curated DNA and protein alignments . The tree topologies with the highest log likelihood were calculated , with all positions within the alignment files containing gaps and missing data ignored . The reliability of the generated tree topologies was assessed using the bootstrap test of phylogeny using 100 iterations . Bootstrap values >50% are shown above their corresponding branches . YPD plates containing 15 μg ml-1 of benomyl were prepared as described previously [49] . Yeast strains expressing XRN1 or containing an empty vector were grown overnight at 30°C in CM—leucine . Cell numbers were normalized and subject to a 10-fold serial dilution before spotting onto YPD agar plates with or without benomyl , and grown at 37°C for 72 h . S . cerevisiae carrying multi-copy plasmids encoding XRN1 or GFP under the control of the GAL1 promoter were grown overnight at 30°C in CM—uracil with raffinose as a carbon source . Cell numbers were normalized and subject to a 10-fold serial dilution before spotting onto CM—uracil agar plates containing either 2% raffinose or galactose . Plates were grown at 30°C for 72 h . Plasmids encoding various XRN1 genes were used to transform BJH006 . Purified single colonies of killer yeasts were inoculated in 2 ml CM—leucine cultures and grown to mid-log phase . YPD “killer assay” agar plate supplemented with methylene blue ( final concentration 0 . 003% w/v ) and pH balanced to 4 . 2 with sodium citrate , were freshly inoculated and spread with S . cerevisiae K12 and allowed to dry . Thereafter , 1 . 5 μl of water containing 6 x 105 cells was spotted onto the seeded YPD plates and incubated at room temperature for 72 h . The diameter of the zones of growth inhibition were measured and used to calculate the total area of growth inhibition . The curing of the killer phenotype was measured by transforming S . cerevisiae BJH006 with approximately 100 ng of plasmid encoding various XRN1 genes using the LiAc method . The addition of 1000 ng or as little as 10 ng of plasmid had no affect on the percentage of colonies cured using this assay . After 48 h of growth , colonies were streaked out and grown for a further 48 h . Clonal isolates of killer yeasts were patched onto a YPD “killer assay” plate ( see kill zone measurement protocol ) that were previously inoculated with S . cerevisiae K12 , and incubated at room temperature for 72 h . The presence or absence of a zone of inhibition was used to calculate the percentage of killer yeast clones cured of the killer phenotype . PHYRE was used to create a template-based homology model of S . cerevisiae Xrn1p using the solved structure of K . lactis Xrn1p as a template [58 , 59] . The structure was determined with an overall confidence of 100% ( 36% of aligned residues have a perfect alignment confidence as determined by the PHYRE inspector tool ) , a total coverage of 81% , and an amino acid identity of 67% compared to K . lactis Xrn1p . PDB coordinates for the modeled structure can be found in S1 File . Structural diagrams were constructed using MacPyMOL v7 . 2 . 3 . Strains were grown in CM lacking the appropriate amino acids in order to retain the relevant plasmids . For co-immunoprecipitations involving L-A Gag-V5 and Xrn1p-HA , 50 mL cultures ( CM—tryptophan—leucine , 2% raffinose ) were used to inoculate 500 mL cultures ( CM—tryptophan—leucine , 2% galactose ) at OD600 ~0 . 1 . Cells were harvested at OD600 0 . 7 , after ~14 h of growth at 30°C with shaking . Cultures used for the immunoprecipitation of native Gag were grown in the same manner , but in CM—leucine medium containing 2% dextrose . Immunoprecipitation of yeast and viral proteins were performed as previously described [90] with the following modifications: 2–4 mg of protein was used per co-immunoprecipitation . Approximately 50 μg of protein was loaded for the whole-cell extract “input” , as determined by Bradford Assay ( ~2% of total input ) , and was compared to 10–20% of each co-immunoprecipitation . Sepharose beads were substituted for Dynabeads MyOne Streptavidin T1 or Dynabeads Protein G ( Thermo Fisher Scientific ) . For immunoprecipitation of Xrn1p-HA , we used an anti-HA-Biotin , rat monoclonal antibody ( 3F10—#12158167001 ) ( Roche ) , and for Gag-V5 a mouse monoclonal antibody ( R960-25 ) ( Life Technologies ) . RNase A was added to whole cell extracts at a concentration of ( 80 μg mL-1 ) and incubated with Dynabeads during immunoprecipitation for 2 hours at 4°C . RNAse is in excess in our co-IP experiments , because significant RNA degradation occurred at concentrations of RNase 8-fold lower than we used ( S5 Fig ) . RNA from samples with and without the addition of RNase A was recovered from yeast whole cell extracts after co-immunoprecipitation using Trizol according to manufacturer’s guidelines ( Thermo Fisher ) . The extent of RNA degradation was measured using a 2200 TapeStation Instrument and a RNA screentape , as per manufacturer’s instructions ( Agilent ) . An RNA integrity number ( RIN ) was calculated for each sample based upon criteria that reflect the quality of the RNA sample , as described previously [93] . dsRNAs were isolated from S . kudriavzevii as described above and processed according to the protocol of Potgieter et al . [62] , with the following modifications: Reverse transcription reactions were carried out using Superscript IV ( Thermo Fisher ) , PCR amplification was performed by Phusion polymerase ( Thermo Fisher ) , and cDNAs were cloned into pCR8 by TOPO-TA cloning ( Thermo Fisher ) before Sanger sequencing . S . kudriavzevii was transformed with plasmids expressing XRN1 from various Saccharomyces species , and an empty vector control using the LiAc method . The transformation was carried out at room temperature and heat shocked at 30°C . S . kudriavzevii transformants were recovered on CM—tryptophan and grown at room temperature . Clones were derived from two independent transformation reactions and grown to mid-log phase at room temperature . Total RNA was extracted from these cultures by first treating the cultures with Zymolase 100T ( final concentration 100 μg mL-1 ) for 2 hours at room temperature in buffer Y1 ( 1 M Sorbitol , 100 mM EDTA ( pH 8 . 0 ) , 14 mM β-mercaptoethanol ) . Yeast spheroplasts were treated with Trizol to extract total cellular RNA , followed by a digestion of residual DNA by Turbo DNase for 30 min at 37°C ( Thermo Fisher ) . The RNeasy RNA cleanup protocol was used to remove DNase from the RNA samples ( Qiagen ) , which were then stored at -80°C . RNA was converted to cDNA using Superscript III and random hexamer priming , as per manufacturers recommendations . cDNA samples were diluted 10-fold with distilled RNase-free water and used as templates for qPCR . Primers designed to recognize the RNAs corresponding to GAG of SkV-L-A1 ( 5’-TGCTTCTGATTCTTTTCCTGAATGG-3’ and 5’-GCCACTTACTCATCATCATCAAAACG-3’ ) and the cellular transcripts from TAF10 ( 5’-ATGCAAACAATAGTCAAGCCAGAGC-3’ and 5’-TCACTGTCAGAACAACTTTGCTTGC-3' ) were used to amplify cDNA using SYBR Green PCR Master Mix ( Thermo Fisher ) on a CFX96 Touch ( Biorad ) . TAF10 was used as a cellular reference gene to calculate the amount of viral cDNA within a given sample using the comparative CT method [65] .
|
Like other eukaryotes , Saccharomyces cerevisiae is chronically infected with viruses . It is fascinating to consider how S . cerevisiae deals with viral infection , because yeast have limited mechanisms of immunity . Our paper focuses on Xrn1p , an enzyme that is important for the destruction of irregular cellular RNAs in all eukaryotic cells . Xrn1p also degrades viral RNAs , owing to the fact that viral RNAs share biochemical characteristics with aberrant cellular mRNAs . Xrn1p was previously known to efficiently control the replication of a S . cerevisiae virus called “L-A . ” We find that two different L-A viruses of Saccharomyces yeasts are best controlled by the Xrn1p from their own host species compared to the Xrn1p from other species . Importantly , these Xrn1p from different species are functionally equivalent in all other ways . This would suggest that while the important cellular functions of Xrn1p have been conserved over millions of years , the interaction with L-A-like viruses has been dynamic and constantly redefined by evolution . The identification of species-specific host proteins , like Xrn1p , is recently being appreciated as a key criterion for understanding why viruses infect the species that they do .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"organismal",
"evolution",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"rna",
"extraction",
"microbiology",
"cloning",
"fungi",
"model",
"organisms",
"fungal",
"evolution",
"microbial",
"evolution",
"molecular",
"biology",
"techniques",
"extraction",
"techniques",
"research",
"and",
"analysis",
"methods",
"saccharomyces",
"mycology",
"pathogenesis",
"viral",
"replication",
"molecular",
"biology",
"viral",
"evolution",
"yeast",
"host-pathogen",
"interactions",
"virology",
"biology",
"and",
"life",
"sciences",
"yeast",
"and",
"fungal",
"models",
"saccharomyces",
"cerevisiae",
"evolutionary",
"biology",
"organisms"
] |
2016
|
XRN1 Is a Species-Specific Virus Restriction Factor in Yeasts
|
Staphylococcus aureus , a pathogen responsible for hospital and community-acquired infections , expresses many virulence factors under the control of numerous regulatory systems . Here we show that one of the small pathogenicity island RNAs , named SprD , contributes significantly to causing disease in an animal model of infection . We have identified one of the targets of SprD and our in vivo data demonstrate that SprD negatively regulates the expression of the Sbi immune-evasion molecule , impairing both the adaptive and innate host immune responses . SprD interacts with the 5′ part of the sbi mRNA and structural mapping of SprD , its mRNA target , and the ‘SprD-mRNA’ duplex , in combination with mutational analysis , reveals the molecular details of the regulation . It demonstrates that the accessible SprD central region interacts with the sbi mRNA translational start site . We show by toeprint experiments that SprD prevents translation initiation of sbi mRNA by an antisense mechanism . SprD is a small regulatory RNA required for S . aureus pathogenicity with an identified function , although the mechanism of virulence control by the RNA is yet to be elucidated .
Staphylococcus aureus is a member of the commensal flora that can be an opportunistic pathogen and a cause of nosocomial and community-acquired infections [1] . With the widespread use of antimicrobials , the incidence and spread of highly antibiotic-resistant S . aureus strains have increased rapidly in recent years and constitute a clinical and epidemiological challenge in hospitals all over the world . In order to survive and to establish an infection , S . aureus inhibits the attack of the host immune system , utilizing diverse escape mechanisms [2] . The staphylococcal protein A ( SpA ) recognizes the Fc domain of immunoglobulins which results in inverted tagging and blocking the C1q and Fcγ receptor binding sites [3] . S . aureus IgG binding protein ( Sbi ) is another immunoglobulin-binding protein expressed by S . aureus [4] . Sbi acts also as a complement inhibitor and forms a tripartite complex with host complement factors H and C3b [5] . S . aureus modulates the expression of virulence genes in response to environmental changes thanks to global regulatory elements . They are either two-component regulatory systems as the agr ( accessory gene regulator ) regulon which is a sensor of the population density [6] , or transcription factors as the SarA family of DNA binding-proteins [7] . These pathways allow the expression of virulence factor regulation during host colonization and dissemination . In addition to protein-mediated regulations , ribonucleic acids also possess regulatory functions in many bacterial pathogens [8] . Until now , RNAIII is the only S . aureus regulatory RNA with a demonstrated function . It is the effector of the global agr regulon that controls the synthesis of several virulence factors [9] , [10] . RNAIII regulates the expression of numerous mRNA targets at the translational and/or transcriptional levels [11] and also acts as an mRNA , containing a small ORF encoding the delta-hemolysin . Additional regulatory RNAs are expressed by S . aureus [12]–[14] . Their expression profiles vary among clinical strains and many of them , called Spr for ‘small pathogenicity island rNAs’ , are expressed from genomic pathogenicity islands containing virulence and antibiotic resistance genes . Their functions are so far unknown . This study was aimed at elucidating the role of one of them , SprD . The Sbi immune evasion protein was identified as a molecular target of SprD . We show that SprD interacts with the sbi mRNA by an antisense mechanism , occluding the Shine-Dalgarno ( SD ) sequence and the initiation codon . Moreover , we show that a small regulatory RNA SprD has a major implication during the intravenous ( i . v . ) infection of mice by a S . aureus clinical strain .
The expression of SprD was monitored during the growth of N315 ( agr− ) [15] , [16] and MRSA252 ( agr+ ) [17] , two S . aureus clinical isolates . SprD is already expressed during the early exponential ( E ) phase ( Figure 1 ) , in contrast to RNAIII that is transcribed at late exponential and stationary ( S ) growth phases . SprD expression increases during the E phase , up to the end of the E phase for N315 and MRSA252 strains . To evaluate the implications of the RNAIII in SprD expression during growth , an RNAIII deletion mutant ( ΔRNAIII ) was constructed in strain RN1 ( agr+ ) [18] . In contrast to N315 and MRSA252 , the SprD expression levels remain almost identical during cell growth in strain RN1 ( Figure 1C–D ) . Also , during growth SprD expression is similar in RN1wt and in RN1-ΔRNAIII isogenic strains ( Figure 1C–D ) , suggesting that the expression pattern of SprD is not influenced by the RNAIII . Therefore , the expression of SprD is independent of the presence of absence of RNAIII . SprD is expressed from the genome of a converting phage containing virulence factors [12] . In most S . aureus strains , sprD is situated in-between scn and chp , within the 8 kb innate immune evasion cluster ( IEC ) that contains the genes for modulation of the early immune response . Such a genomic localization , as well as its growth phase dependent expression , suggest that this RNA may regulate the expression of virulence factor ( s ) . In order to identify the target ( s ) of SprD , we analyzed whether SprD modifies the expression of extracellular proteins that contain many virulence factors . For this purpose , a sprD deletion mutant ( ΔsprD ) was constructed . We determined sprD 5′-end by RACE ( rapid amplification of cDNA ends ) at position C2007178 from the S . aureus N315 sequence [16] . Based on ( i ) 5′-end mapping , ( ii ) transcript size derived from Northern blot analysis [12] and ( iii ) transcription terminator prediction ( Figure S1 ) , sprD 3′-end was assigned at position G2007037 , implying that SprD has 142 nts . In S . aureus N315 , the sprD gene was substituted for an erythromycin resistance cassette by homologous recombination , abolishing the SprD expression ( Figure 2A ) . Complementation of ΔsprD was achieved with a pCN38ΩsprD plasmid expressing SprD from its endogenous promoter ( strain ΔsprD+SprD ) . In S . aureus strains RN4220 [19] and SH1000 [20] naturally lacking the sprD gene ( Figure 2A ) , SprD was expressed with the pCN38ΩsprD plasmid ( Figure 2A ) . In all three strains , levels of a ∼45kD protein decrease in the presence of SprD ( Figure 2B ) . In the RN4220 strain , proteins from that band were eluted , a tryptic digest was prepared and the fragments analyzed by MALDI-TOF mass spectrometry . Twenty-five peptides were identified , all matching the sequence of the Sbi protein ( Table S1 ) . A confirmation of the decrease of the Sbi levels by SprD within the extracellular proteins was obtained by monitoring the Sbi protein by Western blots ( Figure 2C ) . Interestingly , the SprD-dependent downregulation of the Sbi protein was also observable within the intracellular proteins ( Figure 2C ) , indicating that the regulation does not affect Sbi protein export but the overall Sbi protein expression levels . Complementation of ΔsprD with the pCN38ΩsprD vector reduces the Sbi protein levels in vivo ( Figure 2 , panels B and C ) , demonstrating that SprD by itself regulates the expression of Sbi protein . These results were also obtained in strain RN1 and its isogenic ΔsprD mutant ( data not shown ) . Wild-type N315 and ΔsprD strains growth curves are superimposable in rich broth ( Figure 2D ) , demonstrating that the SprD does not influence S . aureus proliferation . The complemented strain leads to lower Sbi protein levels compared to the wild-type N315 strain ( Figure 2 , panels B and C ) because the expression of SprD from pCN38ΩsprD is higher than its endogenous expression levels in wild-type N315 strain ( Figure 2A ) . In the N315 strain , the highest expression of the Sbi protein during growth is at mid-exponential phase and goes to zero at early stationary phase and beyond ( Figure 2D ) . In its isogenic ΔsprD mutant , the Sbi protein levels are higher during growth but the expression profile remains similar ( Figure 2D ) . Taken together , these data establish a functional link between the Sbi protein levels and the expression of SprD demonstrating that , in different S . aureus genetic backgrounds , SprD represses Sbi expression in vivo . To test whether the regulation of Sbi by SprD is at transcriptional and/or at translational level ( s ) , the sbi mRNA levels were monitored by Northern blots in wt N315 and ΔsprD strains . The sbi mRNA expression profiles are similar in both strains ( Figure 2E ) , with a gradual increase of the mRNA expression up to the early-exponential phase and a sharp decrease to basal levels later on . When the sbi mRNA strongly decreases at the stationary phase , SprD is not more expressed , indicating that the Sbi repression at the S phase is ‘SprD-independent’ . Also , the sbi mRNA expression profile does not follow the protein synthesis pattern , probably meaning that the Sbi protein is stable and accumulates during growth . Therefore , in strain N315 , the expression of the Sbi protein is dictated by its transcription profile during bacterial growth . In addition , SprD does not modify the steady state level of the sbi mRNA . Taken together , these results show that SprD downregulates Sbi expression at translational level . We focused our next investigations on Sbi to elucidate the mechanism of its regulation by SprD . A substantial fraction of bacterial regulatory RNAs for which a function was identified interacts with target mRNAs to regulate gene expression [21] . Putative interactions between SprD and the 5′-portion of sbi mRNA were detected in silico ( Figure 3A ) . We first determined the sbi mRNA transcriptional start site by RACE at position G2476039 from the N315 genomic sequence [16] . Therefore , sbi mRNA 5′-end is located 41 nts upstream of the AUG initiation codon ( Figure 3A ) . Duplex formation between SprD and a 179 nts-long sbi mRNA fragment containing its 5′ UTR sequence followed by 46 codons was analyzed by gel retardation assays . A ‘SprD-sbi mRNA’ duplex was detected at a 1∶4 molar ratio and nearly all sbi mRNA was in complex with SprD at a 1∶20 molar ratio ( Figure 3B ) . The binding is specific since a 100- to 2 , 000-fold molar excess of total tRNAs do not displace the sbi mRNA from a preformed ‘SprD-sbi mRNA’ complex . A sbi mRNA deletion mutant lacking 61 nts at its 5′-end ( sbiΔ61 , Figure 3A , brackets ) , predicted to be part of the interaction , does not bind SprD ( Figure 3B ) , demonstrating that these nucleotides are required to interact with SprD . Reciprocally , the deletion of 36 nucleotides ( U35 to U70 ) from SprD ( SprDΔ36 , Figure 3A , brackets ) abolishes complex formation ( Figure 3B ) , showing that these nucleotides are also required for the ‘SprD-sbi mRNA’ interaction . To provide a direct evidence in vivo of the interaction between SprD and the sbi mRNA , we have expressed the SprDΔ36 RNA in the ΔsprD strain . Western blots indicate that SprDΔ36 RNA is unable to dowregulate the Sbi protein levels in vivo , in contrast to full-length SprD ( Figure 3C ) . Northern blot indicates that the SprDΔ36 mutant RNA is expressed at similar levels than SprD wt , demonstrating that the absence of Sbi downregulation by the SprDΔ36 mutant RNA is not due to its instability in vivo . Therefore , this result is a strong evidence of a direct interaction between SprD and the sbi mRNA in vivo , as illustrated in Figure 3A . The interaction between the sbi mRNA and SprD forms in vitro without the contribution of a helper molecule ( Figure 3B ) , as the Sm-like Hfq protein . To test the contribution of the Hfq protein in vivo , we have monitored the SprD-mediated regulation of Sbi in an hfq deletion strain versus an isogenic wild-type strain . As shown in Figure 3D , the in vivo regulation of Sbi expression by SprD takes place independently of the presence or absence of Hfq . These results demonstrate that SprD forms a stable complex with the sbi mRNA in vitro and in vivo , as well as deletions altering the complementarities between the two RNAs impair complex formation . Next , we analyzed in detail complex formation between SprD and the sbi mRNA . As a prerequisite to this study , conformations of the free SprD ( nt 1–142 ) and of the 5′-sbi mRNA ( nt 1–179 ) were investigated using chemical and enzymatic probes . Both transcripts were end-labeled and their solution structures were probed by RNase V1 , which cleaves double-stranded ( ds ) RNAs or stacked nucleotides , and by nuclease S1 and lead , which both cleave accessible single-stranded ( ss ) RNAs . The reactivity toward these structural probes were monitored for each nucleotide ( Figure S2 for SprD and Figure S3 for the 5′-sbi mRNA ) . The data are summarized onto SprD and sbi mRNA 5′-end models that they support ( Figure 4A ) . Out of the 142 nts of SprD , 96 are involved in intramolecular pairings , implying structural stability . SprD has two folded ends ( H1 and H3–H4 ) flanking a 54 nt-long accessible domain made of an unstable stem ( H2 ) capped by a loop ( L2 ) , bordered by two ss ( H1/H2 and H2/H3 ) junctions . For the 5′-end of the sbi mRNA , the data support the existence of two folded stem-loops ( S1-B1 and S2-B2 ) flanking a 9 nt-long accessible domain ( S1/S2 junction ) that contains the predicted SD ( Shine-Dalgarno ) sequence . The AUG initiation codon is located in loop B2 . The pairing prediction and structural changes induced by complex formation between the two RNAs were examined by subjecting a ‘SprD-sbi mRNA’ complex to statistical nuclease S1 and RNases V1 cleavages . Binding of sbi mRNA induced structural changes in a restricted region of SprD ( from U21 to G76 ) , covering the H1/H2 junction , H2 , L2 , and the H2/H3 junction ( Figure 4B ) . The structural data that supports the interaction within each helix , as drawn in Figure 4D , are the following: in the presence of the sbi mRNA , S1 cleavages at U23–U30 ( H1/H2 junction and H2 ) , U37-G39 and U60–U61 ( H2 ) , U46–U47 ( loop L2 ) and A69-U70 ( H2/H3 junction ) disappeared within the SprD structure , whereas S1 cuts at G51-A54 ( H2 ) appear . Upon duplex formation , V1 cuts appeared at A32 ( H2 ) and at U45–U46 ( L2 ) . The binding of SprD led to correlated structural changes in 5′-end of sbi mRNA ( from G1 to G62 , Figure 4C ) , encompassing the predicted SD and AUG initiation codon . In the ‘sbi mRNA-SprD’ duplex , S1 cleavages appeared at positions A53-C59 ( S2 ) and disappeared at positions U43-A48 ( B2 ) within the sbi mRNA sequence . Also , RNAse V1 cuts at positions A16 and U50-A52 appeared , supporting the complex formation as drawn at Figure 4D . Therefore , these data are consistent with the deletion analysis of the ‘SprD-sbi mRNA’ complex and support a bipartite helical interaction between the two RNAs ( Figure 4D ) . Structural probing of the RNA duplex indicates that a discontinuous helical domain forms between the two RNAs ( 22–48SprD/28–53sbi mRNA involving the SD and AUG codon and 56–75SprD/1–19sbi mRNA ) . This helical domain is interrupted by an accessible ss RNA ( 49–55SprD/22–27sbi mRNA ) . Nucleotides from the sbi mRNA flanking the interaction domain ( 54–59sbi mRNA ) become heavily cleaved by nuclease S1 due to steric constraints from the neighbouring duplex . These data demonstrate that the interaction between SprD and the sbi mRNA involves it's predicted SD sequence and AUG initiation codon . Since the interaction of SprD with the sbi mRNA coincides with the region of mRNAs covered by the ribosomes during translation initiation [22] , SprD should prevent ribosome loading on the sbi mRNA . To test this , toeprint assays were performed on ternary initiation complexes including purified 70S ribosomes , initiator tRNAfMet and the sbi mRNA . Two ribosome toeprints were detected onto the sbi mRNA , at 15 and 17 nts downstream from the initiation codon respectively ( Figure 5A , lane 4 ) , supporting the location of the sbi mRNA start codon as drawn on Figure 4 . SprD reduced ribosome loading onto the sbi mRNA in a concentration-dependent manner ( Figure 5A , lanes 5–7 ) . Increasing amounts of SprDΔ36 , that cannot form a complex with the sbi mRNA ( Figure 5B ) , did not prevent ribosome loading onto the sbi mRNA ( Figure 5A , lanes 8–10 ) . It is concluded that SprD inhibits sbi mRNA translation by preventing ribosome binding by antisense pairings with the sbi mRNA 5′-end . These results are in agreement with data obtained in vivo , showing that SprD inhibits Sbi expression at translational level . Since one SprD target is the Sbi immune-evasion molecule that was proposed to be involved in S . aureus pathogenicity [4] , [5] , this RNA may play ( a ) role ( s ) during staphylococcal infections . This suggestion is in agreement with its co-location with virulence factors [12] . Therefore , we tested the importance of the SprD RNA during staphylococcal infections on an animal infection model . Using a murine i . v . sepsis model with an inoculum of 109 S . aureus per mouse , we showed that the virulence of the ΔsprD mutant is abolished ( 100% survival at day 21 of infection ) , whereas all animals infected with the parental wild-type strain die ( Figure 6A ) . The virulence of the trans-complemented ΔsprD+SprD strain is partially restored as compared to the wild type ( 50% survival at day 21 , P<0 . 02 ) . In a different i . v . infection experiment with a 5×108 CFU inoculum per mouse in which animals were sacrificed at day 6 , the kidneys of mice inoculated with the ΔsprD mutant are small and homogenous red-brown , whereas those of mice inoculated with the wild-type strain are substantially swollen and displayed mottled discoloration suggesting numerous abscesses ( Figure 6B ) . Kidneys of mice infected with the ΔsprD+SprD strain are slightly less swollen than the latter , but display homogenous discoloration with no distinct abscesses ( Figure 6B ) . Results of the macroscopic observation are confirmed in the same experiment by viable bacteria counts , as the mean kidney titres ( ± SD ) were 7 . 2±0 . 3 , 4 . 9±1 . 0 , 8 . 5±0 . 6 log10 CFU per pair of kidneys for the wild-type , ΔsprD , and ΔsprD+SprD strains , respectively ( Figure 6C ) . After 6 days of infection , the in vivo persistence of plasmid pCN38ΩsprD in the ΔsprD+SprD strain was verified in 160 randomly selected colonies obtained from kidney homogenates . All of them have retained resistance to chloramphenicol , a specific marker of pCN38 . The virulence defect of a SprD-deletion strain , compared to an isogenic wild-type strain , was also observed in the agr positive RN1 strain ( data not shown ) . Altogether , these results demonstrate the importance of SprD during bacterial infections triggered by S . aureus clinical isolates . Using the same murine i . v . sepsis model , we also tested the implications of Sbi in S . aureus virulence . For this purpose , a sbi deletion strain ( Δsbi ) and a strain overexpressing sbi under its endogenous promoter from the pCN35Ωsbi plasmid ( sbi+ ) were constructed ( Figure panels S6A and S6B ) . We showed that the virulence of the two Δsbi and sbi+ mutants is similar to that of the isogenic wild-type strains ( Figure S6C ) . These results indicate that only varying the expression levels of the Sbi protein is insufficient to account for the SprD virulence phenotype in our animal infection model and imply that SprD has additional target ( s ) involved in staphylococcal virulence . Taken together , our findings indicate that SprD plays a major role in the virulence of S . aureus .
In this report , we show that a small regulatory RNA expressed by S . aureus clinical strains plays an essential role in bacterial virulence during the infection of mice in a model of septicaemia . After RNAIII , SprD is the second regulatory RNA that plays a major role in S . aureus virulence . RNAs are emerging as regulators that enable bacterial pathogens to express virulence genes when required during infection , illustrating their essential roles in pathogenesis [23] . Numerous sRNAs are implicated in the infections caused by Gram-positive and negative bacteria [23] . Some sRNAs are expressed from pathogenicity islands [12] , and such horizontally acquired post-transcriptional regulators can regulate the expression of genes encoded by the core genome [24 and this report] . Some sRNAs regulate the expression of virulence factors [10] or are expressed when bacteria multiplies within mammalian cells [25] . Their implication in bacterial pathogenesis , however , was not demonstrated in animal models of infection . Recent studies have shown that several sRNAs expressed from various bacteria including V . cholerae , L . monocytogenes and S . typhimurium modulate or are involved in virulence on mice infection models [26]–[28] . In S . aureus , RNAIII is the paradigm for RNA-controlled expression of virulence genes , being the effector of the agr system . RNAIII was the first RNA shown to be involved in bacterial pathogenesis more than fifteen years ago [9] and is the only example in S . aureus until now . Compared to the 142 nt-long SprD , the RNAIII ( 514 nt-long ) is almost four times bigger , encodes a small protein , has a complex structure made of 14 stem-loops [29] and regulates the expression of several virulence genes [10] . The importance of agr for virulence in animal models has been reported [30]–[31] , but the exact contribution of RNAIII awaits the experimental testing of an RNAIII deletion strain . This report reveals that a small regulatory RNA expressed by S . aureus , SprD , enhances the virulence of the agr negative N315 clinical strain ( Figure 6 ) and of the agr positive RN1 strain ( data not shown ) . All the mice infected with the S . aureus strain that does not express SprD survive three weeks after the inoculation , whereas all mice challenged with the wild type strain expressing SprD die within 16 days following inoculation . The virulence of the trans-complemented strain is half restored , with the mice kidneys containing viable bacteria as for the wild type strain . The partial restoration of the virulence of the complemented strain could be due to partial plasmid loss after day 6 or , on the other hand , to a negative impact on bacterial virulence of the higher expression of SprD from the plasmid , compared to the wild type strain . The macroscopic aspect of kidneys from mice infected with bacteria expressing , or not , SprD as well as the lower amounts of bacteria detected in the infected kidneys when SprD is not expressed , indicate that this RNA plays a major role in the virulence of S . aureus ( Figure 6 ) . The effect of SprD on virulence might be linked to the lower amount of bacteria detected in the infected kidneys in the absence of the RNA . We tested the ability of SprD to modify gene expression in S . aureus cells and identified the immune evasion Sbi protein as one molecular target of the RNA . The Sbi protein is among the most abundant secreted proteins [32] produced by many S . aureus clinical isolates [4] , [33] . We have unravelled the mechanism by which SprD regulates Sbi expression . The action of SprD on the sbi mRNA proceeds by antisense pairings , blocking translation initiation . The pairing interaction between SprD and the sbi mRNA and its functional outcome is presented as a model in Figure 5B . A central domain of SprD pairs with the sbi mRNA 5′-end that includes its SD sequence and AUG initiation codon , blocking translation initiation . For SprD , all the structural changes induced by the formation of the duplex are located in stem-loop H2 and single-stranded flanking domains H1/H2 and H2/H3 . The pairings between SprD and the sbi mRNA could be divided into three interacting domains that include the very 5′-end of the sbi mRNA , its SD sequence and its AUG initiation codon . The interacting domains that are single-strand in each of the two RNA structures probably pair first ( the H2/H3 junctionSprD with B1sbi , L2SprD with the purine-rich S1/S2 junctionsbi and the H1/H2 junctionSprD with B2sbi ) , followed by spreading through their respective secondary structures . In vitro and in vivo , experimental evidences demonstrate that the regulation of Sbi expression by SprD takes place without the need of the Hfq protein , illustrating the facultative requirement of the Hfq protein for sRNA–mRNA duplex formation among bacteria . The ‘SprD-sbi mRNA’ interaction involves 41 base-pairs and , as suggested [34] , extended pairings probably overcome the requirement for the Hfq RNA chaperone . This strategy of gene expression inhibition is frequently used by bacterial regulatory RNAs [reviewed in 21] , including the downregulation of another IgG binding protein , SpA , by the RNAIII [35] . Translation inhibition by regulatory RNAs in bacteria is usually sufficient for gene silencing and can occur in the absence of mRNA destabilization [36] . If target mRNA degradation is triggered , as with the double-strand specific RNAse III in some RNA-mediated gene regulations in S . aureus [35] , the process of gene silencing becomes irreversible . SprD does not affect the sbi mRNA levels , indicating that this gene regulation could be reversible . In this report , we demonstrate that Sbi is directly regulated by SprD in vivo and in vitro and we also show that this SprD-mediated regulation is agr independent . Indeed , it was previously reported that the inactivation of the agr global virulence regulator increases the abundance of Sbi in vivo [32] , indicating that agr , as SprD , is a negative regulator of Sbi expression . We show that SprD regulates the expression of the Sbi protein in both agr positive ( SH1000 , RN4220 and RN1 ) and agr negative ( N315 ) strains , demonstrating that the SprD-mediated regulation of Sbi occurs independently of agr . In addition , in various clinical strains , Sbi expression is induced by human IgGs [37] although the mechanism of such a positive regulation is currently unknown , but is independent of the RNAIII ( data not shown ) . IgGs increase the levels of the Sbi protein in the presence and absence of SprD ( Figure S4 ) , indicating that the two regulations are independent . Hence , the Sbi expression is monitored by at least three regulatory pathways , suggesting that the amount of Sbi has to be precisely controlled in S . aureus cells . Such a sophisticated regulation network implies that this protein should be an important factor for staphylococcal physiology . The Sbi protein interferes with innate immune recognition by binding multiple host proteins including the complement factors H and C3 as well as IgG ( the Sbi protein traps human IgGs [4] ) and β2-glycoprotein I [5] , [38] , [39] , suggesting that Sbi has a role during staphylococcal infections . Analysis of the virulence of sbi deletion and overexpression strains suggests that Sbi does not appear to be a major virulence factor for staphylococcal infection in a model of septicaemia . Similarly to SpA , the first discovered staphylococcal immunoglobulin-binding protein which has properties comparable to those of Sbi [40] , its contribution to bacterial virulence was difficult to prove in vivo , demonstrating the variability of the results obtained depending on the animal model considered [41]–[43] . As for SpA , the effect of Sbi on virulence is probably hard to be identified , only visible in a few infection models . Since the Sbi protein is predicted to be implicated at early stages of the infection , its contribution is difficult to assess in our infection model . Moreover , the Sbi and SpA proteins could have overlapping functions in host immune evasion , deregulation of expression of either sbi or spa may be insufficient to induce virulence defects on animal models . The sbi deletion or the Sbi overproduction have no detectable virulence phenotypes in our infection model , indicating that the virulence defect of the sprD deletion mutant is not caused only by the deregulation of the Sbi expression levels . Thus , SprD is predicted to have other target ( s ) and/or more general functions implicated in staphylococcal virulence . We do not exclude , however , the implication of Sbi in S . aureus virulence . The expression profile of SprD during growth shows elevated expression levels at stationary phase ( Figure 1 ) when the sbi mRNA levels are sparse ( Figure 2E ) , implying that SprD functions are not restricted to the regulation of Sbi expression , also suggesting that SprD has additional target ( s ) that could be involved at various times during the infection . Indeed , RNAs often regulate the expression of more than a single target , as for several E . coli RNAs [reviewed in 44] and for the S . aureus RNAIII [11] . Also , it would not be so surprising that regulatory RNA ( s ) other than SprD act synergistically to regulate the expression of the sbi mRNA during cell growth , and a reasonable candidate could be the RNAIII . As for SprD that regulates the expression of Sbi and of other putative target ( s ) , the RNAIII represses , by antisense pairings , the expression of the Sbi-like SpA protein and also controls the expression of additional genes either directly or by limiting the expression of the Rot transcriptional regulator [11] . Preliminary data obtained in our laboratory indicate that SprD has at least one mRNA target in S . aureus cells . The identification of SprD additional target ( s ) and learning how they are regulated by SprD will be required to understand implication of this sRNA in S . aureus virulence . Identification of Sbi as the first target of SprD is an important step in elucidating the complete gene network regulated by this small RNA which has such a major role in virulence . Our work , in combination with what is known about RNAIII , suggests a major role for regulatory RNAs in S . aureus pathogenicity . This study also illustrates how sophisticated the regulations of virulence factors productions are during S . aureus infections . It reinforces the roles of RNAs in regulating numerous biological processes in this bacterium . Further studies will be necessary to identify the complete gene network regulated by SprD , its additional target ( s ) , why SprD has such an important role in staphylococcal virulence and the underlying mechanisms of regulations .
Strains and plasmids are listed in Table S2 . S . aureus trains were cultured at 37°C in brain heart infusion broth ( BHI , Oxoid ) . When necessary , chloramphenicol and erythromycin were used at a 10 µg/ml concentration . In pCN38ΩsprD and pCN35ΩsprD sprD is expressed from its own promoter . The sprD sequence with 40 nts upstream and 35 nts downstream was amplified from N315 genomic DNA as a 217-bp fragment , with flanking PstI and EcoRI sites . The PCR product was cloned in pCN38 [45] and pCN35 [45] . For producing the pCN38ΩsprDΔ36 , mutagenized oligonucleotides ‘T7sprD_delfor’ and ‘T7sprD_delre’ were used ( Table S3 ) . In pCN35Ωsbi , sbi is expressed from its endogenous promoter . The sbi sequence was PCR amplified from N315 genomic DNA as a 1700-bp fragment with flanking PstI and EcoRI restriction sites . To inactivate the sprD gene , DNA fragments of 1000 bp upstream and 800 bp downstream of sprD were amplified by PCR from genomic DNA and cloned together with the ermB from pCN51 [45] into XbaI-EcoRI sites of temperature-sensitive plasmid pBT2 [46] . Primers used for cloning are indicated in Table S3 . The resulting plasmid pBT2ΔsprD was transformed into S . aureus strain RN4220 and then into S . aureus N315 to achieve integration of the ermB gene into the genome by homologous recombination . Mutants were enriched by cultivation at 42°C . Cells from the stationary-phase culture were plated on TSA plates and incubated at 37°C . Colonies were imprinted on plates supplemented with 10 µg/mL chloramphenicol . Chloramphenicol-sensitive colonies were tested by PCR for replacement of sprD for the erythromycin cassette . The deletion of sprD was confirmed by Northern blot ( Figure 2A ) . Inactivations of the sbi and RNAIII genes were performed by the same method except that no resistance marker was inserted between their 5′ and 3′ DNA sequences . The primers used for constructing pBT2Δsbi and pBT2ΔRNAII are shown in Table S3 . Virulence levels of the SprD+ strain N315 , its isogenic mutant ΔsprD and complemented strain ΔsprD pCN38ΩsprD were compared using a murine intravenous sepsis model . Groups of 10 female Swiss mice , 6- to 8-weeks old ( Charles River Laboratories , L'Arbresle , France ) were inoculated i . v . with 300 µL of bacterial suspensions containing 109 S . aureus cells in 0 . 9% NaCl . The survival of the mice was monitored for 21 days , and the statistical significance of differences between groups was evaluated using the Mann-Whitney U test . A P value of <0 . 05 was considered significant . With the same three strains , 3 groups of 5 female Swiss mice , 6- to 8-weeks old ( Charles River Laboratories ) were then infected i . v . with 5×109 bacteria . Six days after inoculation , the mice were euthanized with CO2 and their kidneys excised . After photographs were taken , the organs were homogenized , diluted in 0 . 9% NaCl and plated on 5% blood agar for determination of bacterial titres , expressed as log10 CFU per pair of kidneys . Morphology observation included swelling , discoloration and presence of macroscopic abscesses . The stability of plasmid pCN38ΩsprD ( encoding chloramphenicol resistance ) in the complemented ΔsprD mutant was assessed by plating randomly selected colonies grown from kidney homogenates on nutrient agar with containing 20 µg/mL chloramphenicol . For the preparation of protein extracts , bacteria are grown until the exponential or stationary phases and the cells are pelleted for 10 min at 4°C ( 8 . 000g ) . For purifying the extracellular proteins , the supernatants are collected , filtered ( 0 . 45 µm sterilized filter ) and precipitated with 10% trichloroacetic acid . The precipitates are washed with ice-cold acetone and loaded onto SDS-PAGE according to [47] . For the total protein extractions , pellets of 2-ml cultures are washed with TE ( 50 mM EDTA , 50 mM Tris pH 7 . 5 ) , and suspended in 0 . 2 ml of the same buffer containing 0 . 1 mg/ml lysostaphin . Following incubation at 37°C for 10 min , samples are boiled for 5 min , analyzed by SDS-PAGE and stained by Coomassie blue R-250 . The proteins of interest are extracted from gel , trypsin digested and the peptides identified by MALDI MS/MS and RP-HPLC/NanoLC/ESI-MS-MS . For the immunoblots , proteins are transferred to PVDF membrane ( Immobilon-P , Millipore ) . Signals are visualized using a STORM 840 Phosphor-Imager ( Molecular Dynamics ) and quantified using Image-QuantNT 5 . 2 . Total RNAs are prepared as described [48] . For SprD and other sRNAs , Northerns are performed with 5 µg of total RNAs , as described [12] . For sbi mRNA , Northerns are performed as described [49] . RACE assays are carried out according to 49 with the primers from Table S3 . Wild-type and mutant RNAs for probing , gel-shift assays or toeprints are transcribed from PCR fragments generated from genomic DNA with the primers from Table S3 . For producing the template-encoding SprDΔ36 , mutagenized oligonucleotides ( Table S3 ) were used . The RNAs were produced by in vitro transcription using MEGAscript ( Ambion ) . Adding [α32-P]UTP within the transcription mix produces radioactive transcripts . 5′-RNA labeling is performed as described [49] . The RNAs are purified by 8% PAGE , eluted , ethanol precipitated and stored at −80°C . Gel retardation assays are performed as described [49] , 0 . 4 pmol of labeled wt or sbiΔ61 mRNAs are incubated with various concentrations ( from 1 . 6 to 20 pmols ) of unlabeled wt SprD or SprDΔ36 . For structural analysis duplexes between sbi mRNA and SprD are prepared by incubating 0 . 4 pmol of labeled RNA and 1 . 6 pmol of unlabeled RNA in a buffer containing 10mM Tris-HCl ( pH 7 , 5 ) , 60 mM NaCl , 10mM EDTA and 5 mM DTT for 15 min at 25°C . Structural assays are performed as described [49] . Digestions are at 25°C for 15 min with 2 . 5 µg of yeast tRNAs with 0 . 2 or 1 unit of S1 and 10−4 or 5 . 10−5 units of V1 . Lead ( II ) cleavages are performed with 0 . 2 or 0 . 4 mM PbAc in 25 mM Hepes ( pH 7 . 5 ) , 7 mM Mg acetate and 35 mM K acetate for 10 min at 25°C . The reactions are precipitated , the pellets dissolved in loading buffer ( Ambion ) . The samples are denatured for 5 min at 65°C prior to separation on 8% polyacrylamide/8M urea gels . Gels are dried and visualized ( STORM 840 Phosphor-Imager ) . The toeprints are as described [50] with modifications . Annealing mixtures contain 0 . 2 pmol of sbi mRNA and 1 pmol of labeled ‘SBIrevTR’ primer in a buffer containing 10 mM Tris-acetate ( pH 7 . 5 ) , 60 mM NH4Cl , and 1 mM DTT . For the assays in the presence of SprD , various concentrations of wt or SprDΔ36 are added prior to the purified E . coli 70S ribosomes . The ribosomes are reactivated for 15 min at 37°C and diluted in the reaction buffer in the presence of 1 mM MgCl2 . 4 pmols of 70S are added in each assay , incubated for 5 min and MgCl2 is adjusted to 10 mM . After 5 min , 10 pmols of uncharged tRNAfMet are added and incubated for 15 min . cDNA is synthesized with 2 UI of AMV RT ( Biolabs ) for 15 min . Reactions are ended by 10 µl of loading buffer II ( Ambion ) . The cDNAs are loaded and separated onto 8% PAGE . The toeprints are located on the sbi mRNA sequence by sequencing the DNA . All animal experiments were performed in accordance to European guidelines and recommendation of the French Agricultural Office for the care of animals subjects . Experiments were carried out in the accredited research animal facility of Institut Pasteur de Lille ( accreditation number , A59107 ) . All animal protocols were approved by the locally appointed investigational review board ( Institut Pasteur de Lille , accreditation number , A59107 ) . S . aureus Immunoglobulin G binding protein A : GenBank ID: BAB41326 . 1 S . aureus Sbi protein: Genbank ID: AF027155 S . aureus RNAIII ( nt 1260 to 1571 ) : GenBank accession number: X52543 S . aureus Hfq: PDB code 1Kq1A
|
Bacteria possess numerous and diverse means of gene regulation using RNA molecules , including small RNAs ( sRNAs ) . Here we show that one sRNA is essential for a major human bacterial pathogen , Staphylococcus aureus , to cause a disease in an animal model of infection . Our study provides evidence that this RNA regulates the expression of an immune evasion molecule secreted by the bacterium to impair the host immune responses , and we have solved the mechanism of the RNA-based regulation at molecular level . So far , the mechanism of bacterial virulence controlled by SprD is unrevealed , but that small RNA has a huge impact in the course of a bacterial infection . It implies possible new strategies in fighting against that major human and animal bacterial pathogen in preventing the expression of this regulatory RNA .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biochemistry/rna",
"structure",
"microbiology/immunity",
"to",
"infections",
"infectious",
"diseases/bacterial",
"infections",
"biochemistry/transcription",
"and",
"translation",
"microbiology/cellular",
"microbiology",
"and",
"pathogenesis"
] |
2010
|
A Staphylococcus aureus Small RNA Is Required for Bacterial Virulence and Regulates the Expression of an Immune-Evasion Molecule
|
Somitogenesis is a process common to all vertebrate embryos in which repeated blocks of cells arise from the presomitic mesoderm ( PSM ) to lay a foundational pattern for trunk and tail development . Somites form in the wake of passing waves of periodic gene expression that originate in the tailbud and sweep posteriorly across the PSM . Previous work has suggested that the waves result from a spatiotemporally graded control protein that affects the oscillation rate of clock-gene expression . With a minimally constructed mathematical model , we study the contribution of two control mechanisms to the initial formation of this gene-expression wave . We test four biologically motivated model scenarios with either one or two clock protein transcription binding sites , and with or without differential decay rates for clock protein monomers and dimers . We examine the sensitivity of wave formation with respect to multiple model parameters and robustness to heterogeneity in cell population . We find that only a model with both multiple binding sites and differential decay rates is able to reproduce experimentally observed waveforms . Our results show that the experimentally observed characteristics of somitogenesis wave initiation constrain the underlying genetic control mechanisms .
Somitogenesis is the process by which vertebrate embryos develop somites , which are transient , repeated blocks of cells arising from the presomitic mesoderm ( PSM ) that differentiate further into vertebrae , ribs , musculature , and dorsal dermis . The tailbud is a proliferative zone at the posterior end of the embryo where immature cells are continually added to the posterior-most PSM . As the tailbud grows away posteriorly , the oldest cells in the anterior PSM segment in groups to form lateral pairs of somites along the midline . The process stops when the anterior formation of somites has progressed posteriorly across the entire PSM , reaching the arresting growth in the tailbud [1]–[4] . Somitogenesis is an impressively robust mechanism of pattern formation in developmental biology that has received much experimental and theoretical attention . In 1976 , based on theoretical considerations , Cooke and Zeeman [5] postulated that somitogenesis proceeds by a “Clock and Wavefront” mechanism . In this model , the susceptibility of cells in the PSM to form somites oscillates between susceptible and insusceptible ( the clock ) , while a determination wavefront sweeps posteriorly across the PSM . The passing wavefront triggers cells to form somites , but does so only when cells are susceptible , i . e . , when their clocks are in the correct phase of oscillation . Since adjacent cells are in phase , cohorts of cells are recruited in succession to form somites . Initially , the clock was thought to be closely linked to the cell cycle [6] . In 1997 , Palmeirim et al . [7] discovered a gene with oscillatory expression in the PSM of the chick embryo , providing an alternative candidate for the clock . Experimental work has since identified multiple oscillatory genes in each of several model organisms , including mouse [8] and zebrafish [2] . In all of these organisms , the oscillatory gene expression in individual cells is coordinated throughout the PSM in order to produce spatiotemporal waves of mRNA and protein expression , which we call the clock-wave . Synchronized , periodic expression is observed in the tailbud with a frequency that matches the anterior formation of somites [1] , [2] . Broad waves of expression repeatedly initiate in the posterior-most PSM and narrow while traveling anteriorly [1]–[3] , [7] . The waves slow considerably as they reach the region of forming somites . Successive waves arriving at the anterior-most PSM help sequentially establish stable bands of high-low gene expression in several additional genes , indicating nascent somite boundaries and polarity [1]–[3] . Separate experiments have identified biochemical candidates for the wavefront [9] , [10] . These bio-molecules exhibit graded concentration profiles across the PSM that shift posteriorly in synchrony with tailbud growth . A changing gradient level triggers mesodermal cell differentiation and somite formation [1] , [2] , [11] , [12] . We call this the gradient-wavefront . The precise mechanism in which the clock-wave interacts with the gradient-wavefront , as well as their possible interactions with intercellular signaling mechanisms , remains unknown [1] , [2] , [12] , [13] . Many mathematical models of the dynamics of somitogenesis have been proposed , with reviews and comparisons of several prominent models available in the literature [12]–[18] . Zebrafish is a standard model organism in the study of somitogenesis , and we now describe in more detail the mathematical modeling work in zebrafish that is most closely related to our present work . Lewis [19] has studied a two-cell model of the clock , where the oscillations arise from delayed , intracellular negative feedback of a protein dimer on its own mRNA production . In two different versions of the model , the clock consisted of one or two genes ( her1 and/or her7 ) , and when both were included they interacted by protein heterodimerization . To synchronize gene expression between two neighboring cells , Lewis extended similar mammalian models [19]–[21] by introducing delayed , intercellular positive feedback via DeltaC/Notch protein signaling [22] . In 2006 , Horikawa et al . [23] extended Lewis's model to a lateral line of synchronized cells in the PSM . Neither Lewis nor Horikawa's group addressed the posterior-to-anterior slowing of the oscillation rate that leads to formation of the clock-wave . However , Giudicelli et al . [24] experimentally quantified the slowing of oscillations in PSM cells , and measured model parameters such as mRNA production and transport delays and decay rates . Özbudak and Lewis [25] used this information to refine Lewis's two cell model and concluded that DeltaC/Notch signaling is used for coordinating oscillations between cells in the zebrafish PSM , but not for generating oscillations or forming somite boundaries . Concurrently to the above work , a protein ( Her13 . 2 ) was discovered in zebrafish that interacts with at least one of the clock-gene proteins [26] , [27] and controls the rate of oscillatory expression in individual PSM cells , thereby inducing the formation of the clock-wave [26] . This protein is expressed in a graded fashion along the anteroposterior ( AP ) axis of the PSM . Based on this information , Cinquin [28] proposed a multicellular model for zebrafish somitogenesis that requires heterodimerization of two clock proteins ( Her1 and Her7 ) . In this model , formation of this heterodimer competed with the formation of other dimers , including heterodimers of each of the clock proteins with Her13 . 2 . This competitive dimerization , combined with different levels of repression by the various dimers , produced waves of gene expression . Her13 . 2 acts downstream of a morphogen gradient FGF , which is the presumed gradient-wavefront that controls somite formation in the anterior-most cells of the PSM [1] , [2] , [9] , [10] , [12] , [26] . This suggests distinguishing between two distinct phases of somitogenesis; the first is the generation of a clock-wave in the PSM that narrows and slows as it propagates anteriorly , while the second is the commitment of cells in different phases of oscillation to different developmental pathways and somite formation . In this paper we develop a biologically informed , yet minimally constructed , mathematical model that generates the initial narrowing and slowing of the clock-wave in the posterior PSM . Our model incorporates the delayed , intracellular negative feedback model of Lewis [19] for the clock and was motivated by the results of Cinquin [28] , in which competitive dimerization of clock proteins with a graded control protein contributes to the slowdown of clock oscillations . Our multicellular model retains much of the simplicity of Lewis's deterministic , single clock-gene model with intercellular coupling , incorporating a minimum of additional biological components to generate the experimentally observed posterior clock-wave in zebrafish . Our main goal is to determine if this experimentally observed aggregate behavior of the clock-wave is sufficient to constrain the genetic control mechanisms responsible for the oscillatory gene expression of the clock . We consider two different genetic control mechanisms , giving four different model scenarios with either one or two binding sites for the self-repressing clock protein homodimer , and where either only monomers of clock protein decay or where both monomers and dimers decay linearly with the same rate constant . The differential decay of monomers and dimers is an example of cooperative stability , which was found to have a significant impact on behavior of a bistable switch and the repressilator in [29] . We parametrize the model to the extent possible with experimentally determined parameters from zebrafish [19] , [24] , [25] , but in each model scenario there are a number of parameters with unknown values . We uniformly sampled 40 , 000 combinations of the unknown parameter values from a biologically realistic range and tested if each model scenario was able to reproduce the experimental data at each sampled parameter combination . Two main experimental observations that the model must match are tailbud clock period of 30 minutes to within 10% , and sufficient decrease in the oscillation rate along the axial PSM in order to generate the observed clock-wave . We find that only the model scenario that combines two binding sites for the clock protein repressor and different decay rates of the clock protein monomers and dimers is able to accurately reproduce both experimental observations . Sensitivity of clock-wave formation to each estimated parameter is investigated by analyzing the successful combinations of parameters . We find sensitivity with respect to clock mRNA transcriptional delay ( in agreement with [19] ) , clock protein homodimer binding affinity to DNA , its binding cooperativity and protein dimerization constants . To further confirm model validity , we test the optimal model's robustness to heterogeneity in the cell population . For the best choice of estimated parameters , random perturbation in each cell of 22 parameters around their nominal values produces a heterogeneous population of cells . We selected size of our perturbations so that , on average , the majority ( 99 . 7% ) of parameters lie within 1% or 2 . 5% of their nominal values . We test two spatial arrangements of heterogeneous cells: a line of fifty cells along the anterior-posterior axis and 250 cells arranged in five parallel rows along the AP axis . We find that oscillation and clock-wave formation in the PSM is robust to cell heterogeneity at these levels , although we observed a disorganization reminiscent of the salt-and-pepper patterns seen in many DeltaC/Notch knockout/knockdown experiments ( e . g . , Figure 3l in [22] ) at the 2 . 5% level of heterogeneity . Based upon our mathematical model , we conclude that the experimentally observed behavior of the clock-wave significantly constrains the genetic control mechanisms responsible for the clock behavior . The necessity of multiple binding sites for the self-repressive clock protein homodimer verifies an existing hypothesis for the genetic control mechanism of the clock [30]–[32] . Furthermore , very recently and after our paper had been submitted , Brend and Holley [30] experimentally identified two active dimer binding sites for the her1 clock-gene in zebrafish . This result is highly encouraging for our modeling work , even though we concentrate on the clock protein her7 . The necessity of differential decay rates for clock protein monomer and dimer represents further confirmation that the molecular dynamics can be significantly affected not only by the nonlinearities in the production of molecular species , but also the nonlinearities in the decay process [29] . The hypothesized nonlinear decay mechanism may be an important alternative and/or complement to rate-limited protein decay mechanisms studied in [33]–[35] and warrants experimental investigation .
We considered four model scenarios that differed in the clock-gene control and protein decay mechanisms . In scenario I , we assumed a single binding site for the self-repressing clock-protein homodimer and that only clock-protein monomers decay . In scenario II , we still considered a single binding site , but instead assumed that clock-protein monomers , homodimers , and heterodimers with the control protein all decay with the same rate constant . In scenario III , we assumed two binding sites for the self-repressing homodimer and monomer-only decay . Lastly , in scenario IV , we assumed two binding sites and decay of all forms of the clock protein . Table 1 gives the choice of model parameters corresponding to each scenario . Through numerical simulation of the mathematical model we assessed the ability of the above four model scenarios to: Wherever possible , we used experimentally determined parameter values in the model . However , for ten parameters , including minimal ( ) and maximal ( ) total control protein levels , dimerization dissociation constants , clock mRNA production delay , clock monomer decay rate , and clock homodimer binding affinities , only a feasible range of values was known . Tables 2 and 3 summarize the values and ranges of the model parameters and Text S1 includes details on the parameter selection process . We searched this space of parameters for those sets that reproduce experimental clock-wave . A parameter set was considered to produce a valid fit to experimental data if the corresponding model simulation satisfied criteria ( a ) – ( d ) . The following important observation allowed parameter estimation to proceed in two stages: formation of a realistic clock-wave in a large simulation of fifty cells along the AP axis of the embryo depends upon the key value of , defined as the maximum change in clock oscillation period observed over a range of total control protein . The parameters , , and were estimated by simulation of a smaller simulation of two identical , coupled cells by increasing in steps of ten from 0 to 2500 copies per nucleus and recording the period of the oscillation at each step . Parameter combinations giving minutes were observed to generate a biologically realistic posterior clock-wave in the large simulation of fifty cells ( also see [24] ) . Therefore , in the first stage , we took a random sample of size 40 , 000 from a joint distribution of the remaining eight estimated parameters ( Table 2 , see Model Simulation and Selection in the Models section for more details . ) For each parameter set we simulated two identical , coupled cells in each of the four model scenarios . By stepping through the values of from zero to 2500 , we determined if there were values and for which a given model scenario satisfied criteria ( a ) and ( b ) with minutes . In the second stage , an AP line of fifty coupled cells with a spatiotemporally graded control protein was simulated to verify that the selected parameter set from stage one indeed produced a realistic clock-wave in the absence ( c ) and presence ( d ) of cell heterogeneity . Details of the simulation procedure are described in the Models section . For model scenarios I–IV , the first two lines of Table 4 list: 1 ) the number of parameter sets out of 40 , 000 total selections that produced periodic solutions in two coupled cells for some level of in , and 2 ) the number of parameter sets for which the periodic solution also exhibited a period of minutes for some level of . Figure 1A presents the same data using percentages rather than raw counts . The last two lines of Table 4 list the number of parameter sets that support periodic solutions with period minutes for some level of while also producing the indicated differences in period over some interval . The vast majority of solutions that exhibited sustained oscillations with a period of minutes , and thus satisfied criterion ( a ) above , were for scenarios III and IV with two clock protein binding sites . For scenarios I and II , with a single clock protein binding site , the largest was 3 . 4 minutes , and so neither scenario satisfied criterion ( b ) above . For scenarios III and IV , Figure 1B shows the distribution of for those simulations that produced a period of minutes . The important observation is that even though scenario IV produced the required period of oscillation minutes for almost 40% of parameter sets ( as opposed to 8% in scenario III , see Figure 1A ) , the maximum period change for scenario IV was minutes . This was smaller than the minutes necessary for realistic clock-wave formation . In scenario IV , less than 1% ( 131 out of 15509 ) of the parameter sets that produced a period of also produced . In contrast , for scenario III , 10 . 6% ( 346 out of 3247 ) of the parameter sets that produced a period of minutes also produced . Eight out of 3247 parameter sets in scenario III produced , and the maximum period change was minutes , see Table 4 . We remark that since the choice of 40 , 000 parameter sets in 8 dimensional space , if spaced in a regular grid , only gives 3 to 4 different values for each parameter , so that when the parameter sets are chosen randomly each set represents a significant volume of the parameter space . Viewed in this light our success rate of parameter sets that produce is not disappointing . Figure 1C shows the oscillation period as a function of the total control protein in scenarios III and IV for the parameter sets that produced and , respectively . The parameter set selections that produced these optimal values are given in Table 5 . See Text S1 for a complete tabulation of results . Since no parameters for scenario I produced oscillations with the required period of minutes , we concluded that a single binding site with differential protein decay is not capable of producing the experimentally observed oscillations in the zebrafish tailbud . Scenarios II and IV , with equal monomer and dimer clock protein decay rates , did not produce sufficient over the given range of total control protein . Only scenario III , combining two binding sites and monomer-only clock protein decay , admitted a significant number of parameter sets that produced large enough to generate a biologically realistic clock-wave . The second validation stage verified proper clock-wave generation across a growing AP line of fifty coupled cells in the axial PSM . In these simulations , a sigmoidal spatiotemporal gradient of the total control protein was prescribed across the cells in the PSM , decreasing from in the tailbud to anteriorly . Figure 2A–B compares the simulated mRNA clock-waves for model scenarios III and IV using the parameter sets that produced and , respectively . Note that in Figure 2A , the spacing of the posterior-most bands of clock-gene mRNA expression is 14–15 cells ( which narrows to 13 cells toward the anterior PSM ) , comparable to the mean value of approximately 10 . 5 cells measured experimentally for her1 in zebrafish ( see Figure 3 in [43] ) , whose expression is synchronized with her7 [2] . In contrast , the spacing in Figure 2B is about 20 cells , which is considerably larger than the experimentally observed spacing . Videos S1 and S2 show movies of the simulations for scenarios III and IV , respectively . The differential decay of monomers and dimers ( cooperative stability ) and two binding sites for the repressor dimers combined to produce a significant change in oscillation rate between the tailbud and the intermediate PSM . The cooperative stability effect was similar to that discussed in [29]: since the proportion of dimers to monomers increases with the total concentration of protein , the marginal decay rate ( i . e . , decay per unit of total protein ) decreases with total concentration . This effect can be seen in Figure 5 of Text S1 where we compare the linear decay rate and the differential decay rate as a function of total protein concentration , and in Figure 3B where we graph the relative quantities of monomer and dimers during oscillations . The two binding sites primarily affected the production of the clock mRNA , because they increased the effective Hill coefficient of the nonlinearity . Figure 6 in Text S1 compares the nonlinear production curve of clock mRNA as a function of total clock protein level . The production curves for scenarios III and IV ( two binding sites ) were shifted toward low levels of total clock protein as compared to production curves for scenarios I and II ( single binding site ) . Note that significant production of clock mRNA occured only in a limited part of the oscillation cycle of two coupled cells ( red part of the curve in Figure 3D ) for the lowest levels of total clock protein . Since it took a longer time for the total clock protein to decay to this low value , the shifted production curve also enhanced the length of the period . These two effects combined to cause a slow decay of the total clock protein from its peak , compare Figure 3D to Figure 3E where we replaced differential decay by linear decay of total clock protein . So far we have discussed how cooperative stability increased the period of the oscillation in the PSM where total control protein . However , the key to clock-wave formation across many cells is the change in oscillation rate between the tailbud , where is high , and the intermediate PSM , where is low ( cf . , the value of computed in the two-cell simulations in stage one of model selection ) . For two coupled cells , Figure 3A shows the absolute levels of clock monomer , homodimer , heterodimer with control protein , and total clock protein as the level of total control protein was decreased dynamically from to . Figure 3B shows these monomer and dimer levels relative to the total clock protein level . In the tailbud , a significant proportion ( about 75% ) of clock protein was bound in the heterodimer and as a result of this buffering , the oscillations were small in amplitude and more symmetric , see Figure 3C . After the level of total control protein dropped , the oscillator was released from the buffering , the mRNA production curve shifted toward smaller values of total clock protein ( Figure 6 in Text S1 ) , and both the amplitude and the period of the oscillation rapidly increased . The transition from high to low caused a transition from gentle , faster oscillations to slower , burst-like oscillations . The results in Figure 3 were from a simulation of two coupled cells . We examined the effect of the coupling signal on the change in oscillation rate by repeating this simulation for a single cell and found negligible differences in oscillation rates . Therefore , we graphed the production curves in Figure 6 in Text S1 for the mean value of coupling signal ( ) in the respective regions ( tailbud or PSM ) . We examined the sensitivity of eight estimated parameters in each model scenario . We first selected nested collections of the 40 , 000 random parameter sets by imposing increasingly stringent requirements on the corresponding solution: 1 ) ( collection A ) parameter sets for which the solution was periodic for some level of total control protein in the range 0–2500 copies per nucleus , 2 ) ( collection B ) parameter sets for which the solution satisfied 1 ) and had a period of minutes for some level of total control protein in the range 0–2500 copies per nucleus , 3 ) ( collection C ) parameter sets for which the solution satisfied 1 ) and 2 ) and had a period change minutes over a range of total control protein , and 4 ) ( collection D ) parameter sets for which the solution satisfied 1 ) and 2 ) and had a period change minutes over a range of total control protein . A period change of at least minutes is sufficient for generating a biologically realistic posterior clock-wave for zebrafish . Inclusion of collection ( ) allowed direct comparison between scenarios III and IV . Figure 4 shows the coefficient of variation ( C . V . ) of each of the eight parameters in collections – . Text S1 contains histograms showing projections of each collection onto the individual parameters for scenarios III and IV . The C . V . 's for each parameter for each collection were computed from the corresponding distributions . Small values of C . V . show that the parameter value is tightly determined by the particular requirement 1 ) –4 ) and hence the wave formation is sensitive to this parameter . A decreasing C . V . value from left to right signifies increasing sensitivity as a function of more stringent requirements . As expected from the model selection discussion , the largest selective pressure on the parameter sets was imposed by the requirement for . Inspection of the C . V . data ( Figure 4 ) and the corresponding histograms ( Text S1 ) suggested the following . In both scenarios III and IV , attaining the proper period of oscillation ( ) was most sensitive to the clock mRNA production delay . Furthermore , attaining sufficiently large for clock-wave formation was most sensitive to the clock homodimer binding affinity and cooperativity , the clock monomer decay rate , and the dimer dissociation constants , , and . While both scenarios showed sensitivity of the dimer dissociation constants to increasing , scenario IV showed additional sensitivity to the heterodimer dissociation constant . Finally , we note that the parameter sets that belong to collection , but not to collection support oscillation with period of about 30 minutes , but do not produce sufficient which would lead to successful clock-wave formation . This suggest that there may be mutants where a change in certain parameter values will produce uniform oscillation throughout PSM and thus the clock wave initiation will fail . If the estimated parameters in this study can be experimentally measured , then our dataset can be mined for related parameter sets for which little or no change in oscillation rate occurs with changing levels of total control protein . We examined if the optimal scenario III solution was robust to cell heterogeneity . We examined if the optimal scenario III solution was able to reproduce several experiments reported in zebrafish .
The modeling and experimental work of Lewis and coworkers in zebrafish [19] , [24] , [25] was a major foundation for the present work . Compared to their coupled two-cell model , our multicellular model adds explicit tracking of monomer and dimer forms of protein , differential protein decay , and multiple transcription binding sites modeled using the approach of Shea and Ackers [45] . Multiple , active transcription binding sites for the clock-gene her1 have recently been reported in zebrafish [30] . Our model supports this finding , but also suggests the importance of the differential decay of clock protein monomer and dimer . In spite of the added complexity in our model , a fast dimerization assumption allows it to retain much of the simplicity of Lewis's original deterministic , single clock-gene model [19] . The idea of competitive dimerization of a control protein with clock protein was first introduced by Cinquin [28] . A major difference between our model and Cinquin's model is our inclusion of only a single clock-gene ( her7 ) . Whereas Cinquin's model suggests the importance of a Her1-Her7 clock protein heterodimer to clock-wave formation in zebrafish , our model reproduces the initiation of the posterior clock-wave with a single clock-gene . Furthermore , in our model the control protein ( Her13 . 2 ) never acts as a repressor , either as a homodimer or as a heterodimer . However , our parameter sensitivity analysis shows that competitive heterodimerization of clock protein with control protein ( Her7-Her13 . 2 ) is fundamentally important to the rate tuning mechanism of the model . While the the decay rates for protein monomer and dimer are very similar to each other in Cinquin's model [28] , we show that the difference between these rates is largely responsible for tunability of the oscillations . Buchler et al . [29] termed preferential decay of monomers to dimers as “cooperative stability” , and found that it increased the robustness of both a bistable switch and a synthetic oscillator via enlarged parameter regions . More recent work by Wong et . al [35] showed that rate-limited protein decay could also enlarge the viable parameter space for an oscillatory genetic circuit . A similar rate-limited protein decay mechanism was identified as potentially playing a positive role in the somitogenesis oscillator in mouse modeled by Zeiser et al . [34] . In relation to these results , we see that the effect of differential decay through cooperative stability of dimers is more intricate in our model of the somitogenesis oscillator . While the differential decay reduces the parameter region for sustained oscillations as compared to linear decay , it increases the rate-tuning of the oscillator with a changing level of control protein ( larger ) , which is crucial to proper clock-wave formation . We note that we only examined the two most extreme cases of differential decay of monomers and dimers , which is almost certainly not what happens in vivo . Experimental data on dimer dissociation constants , binding affinities , decay constants , and the quantitative shape and magnitude of the control protein gradient would be particularly useful in further validation , refinement , and application of the presented model . Finally , although Delta/Notch coupling was not the focus of the present study and no Delta/Notch parameters were estimated during the parameter selection , our robustness studies showed that the synchronization of heterogenous cells in the tailbud is crucial for the proper formation of the clock-wave . While in this paper we assumed relatively weak coupling and mainly explored the interaction between the control protein and the clock protein , a stronger effect of the signaling protein on clock mRNA production could add complexity to this interaction . Both the amplitude and timing of the Notch signal may be important . Because the decreasing level of total control protein along the PSM shifts the clock mRNA production curve , the relative influence of Notch signal on clock mRNA production also changes . It was noted in [19] that increasing the Notch delay can cause two coupled cells oscillating in synchrony to anti-synchronize . While in the present study the Notch delay is fixed , the underlying oscillation rate is changing as a function of the total control protein . This change in relative timing presents another potential mechanism for Notch coupling to act differently along the AP axis of the PSM . In the last decade , our understanding of somitogenesis benefited from great experimental advances which identified , in multiple organisms , candidates for both clock- and signaling-genes and various candidates for graded morphogens ( control proteins ) that may interact with these genes . However , there is still a vigorous discussion about which of the genes are driving the clock and which are driven by the clock , the role of multiple clock genes , how and which morphogens interact with the clock , and how the somite boundaries ultimately form . What can mathematical modeling bring to the table in face of such uncertainty and complexity ? One approach has been to radically simplify the underlying biology and concentrate on just the observed phenomena . As an example , one can model the clock as a phase oscillator and the wavefront as a prescribed decrease in oscillation frequency and see if a viable clock-wave is generated , see [46] for example . Results of these models highlight the essential features necessary for the clock-wave: slowing of the oscillation as the cell matures in the PSM and coordination of oscillations in cohorts of cells with the same fate . These models , however , do not draw conclusions about the biological mechanisms underlying the clock formation . Our results suggest that a mathematical model can incorporate the existing ( incomplete ) understanding of biology and still suggest concrete , experimentally refutable hypotheses about the biological mechanisms of somitogenesis clock-wave generation . Although our mathematical model was validated using zebrafish data , our model is readily adaptable to other organisms and we believe that its minimal construction makes it a good candidate for further investigation of the key biological questions .
As described in the Results section , model selection occurred in two stages . In each stage , the model was simulated numerically using one of Matlab's delay differential equation solvers ( dde23 or ddesd [56] ) . Because the system of differential equations contained algebraic constraints , each evaluation of the right hand side of the system of delay differential equations by the solver required that the nonlinear algebraic system ( 6 ) be solved for the monomer copy numbers and in terms of the prescribed value of and the state variable . As an alternative to Newton's method , a simple iterative technique for solution of this nonlinear algebraic system was developed ( see Text S1 for details ) . Computation of the dimer copy numbers , , and followed from the corresponding fast equilibrium equations . A matlab class ( Params . m ) was developed to handle the various model configurations and parameter perturbations , ensuring accuracy and reproduction of results . See Text S2 for listings of the Matlab codes employed .
|
The vertebral column is a characteristic structure of all vertebrates . Individual vertebrae , together with ribs and attached muscles , develop from repeated embryonic structures called somites . The somite pattern forms in the embryo during somitogenesis . We know that this process uses periodic gene expression ( a biomolecular “clock” ) to generate the pattern , but we do not know precisely how this expression is controlled within the cell and coordinated across multiple cells . We propose a mathematical model that incorporates experimentally confirmed features of somitogenesis . We then test four different mechanisms that may control the clock and ask if the comparison between model simulations and experimental observation can select the best model and thus suggest how the clock is controlled . We find that the model scenario with both multiple DNA binding sites and differential protein decay rates is best able to reproduce experimental observations . Because these findings can be tested experimentally , our results should help guide future experiments .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Models"
] |
[
"mathematics",
"developmental",
"biology/pattern",
"formation",
"biophysics/biomacromolecule-ligand",
"interactions",
"developmental",
"biology/developmental",
"molecular",
"mechanisms",
"computational",
"biology/systems",
"biology"
] |
2010
|
Somitogenesis Clock-Wave Initiation Requires Differential Decay and Multiple Binding Sites for Clock Protein
|
Gene silencing is a natural antiviral defense mechanism in plants . For effective infection , plant viruses encode viral silencing suppressors to counter this plant antiviral response . The geminivirus-encoded C4 protein has been identified as a gene silencing suppressor , but the underlying mechanism of action has not been characterized . Here , we report that Cotton Leaf Curl Multan virus ( CLCuMuV ) C4 protein interacts with S-adenosyl methionine synthetase ( SAMS ) , a core enzyme in the methyl cycle , and inhibits SAMS enzymatic activity . By contrast , an R13A mutation in C4 abolished its capacity to interact with SAMS and to suppress SAMS enzymatic activity . Overexpression of wild-type C4 , but not mutant C4R13A , suppresses both transcriptional gene silencing ( TGS ) and post-transcriptional gene silencing ( PTGS ) . Plants infected with CLCuMuV carrying C4R13A show decreased levels of symptoms and viral DNA accumulation associated with enhanced viral DNA methylation . Furthermore , silencing of NbSAMS2 reduces both TGS and PTGS , but enhanced plant susceptibility to two geminiviruses CLCuMuV and Tomato yellow leaf curl China virus . These data suggest that CLCuMuV C4 suppresses both TGS and PTGS by inhibiting SAMS activity to enhance CLCuMuV infection in plants .
In the course of plant-virus interactions , plants have evolved ingenious counter-attack mechanisms to diminish or eliminate invading viral pathogens . Of various plant antiviral defenses , gene silencing can target either viral RNAs for degradation through post-transcriptional gene silencing ( PTGS ) or DNA sequences of DNA viruses for epigenetic modification through transcriptional gene silencing ( TGS ) [1 , 2] . DNA cytosine methylation is an important epigenetic marker for gene silencing and controls plant development and gene expression . It also plays an important role in plant defense against invading DNA viruses [3–5] . Cytosine nucleotides are methylated at the 5′ position of the pyrimidine ring to generate 5-methyl cytosine ( 5-mC ) , which is catalyzed by cytosine methyl transferases . In addition , PTGS also requires small RNA methylation [6] . In PTGS , the methyltransferase HEN1 adds a 2’-O-methyl group to the 3’-terminal nucleotide of small RNAs to protect them from a 3’-end uridylation activity [7] . S-adenosylmethionine ( SAM ) is the universal methyl group donor in DNA or RNA methylation , and its metabolic regeneration through methyl cycle involves four reactions: ( 1 ) SAM demethylation to generate S-Adenosyl-L-homocysteine ( SAH ) by SAM methyltransferases , ( 2 ) SAH hydrolysis to produce homocysteine by SAH hydrolase ( SAHH ) , ( 3 ) methylation of homocysteine to yield methionine by methionine synthase , and ( 4 ) adenylation of methionine to form SAM by SAMS ( S-adenosylmethionine synthase , MetK , MAT ) [8] . Several publications have proposed that some DNA viruses can interfere with the proper function of the methyl cycle to reduce plant DNA methylation for effective infection [9–12] . Geminiviruses are a large family of plant viruses with small circular , single-stranded DNA genomes . They infect a broad range of plants and are classified into seven genera ( Becurtovirus , Begomovirus , Curtovirus , Eragrovirus , Mastrevirus , Topocuvirus and Turncurtovirus ) according to the viral genome organization , insect vector and host range [13] . Begomoviruses are transmitted by whiteflies and possess monopartite or bipartite genomes , the latter designated as DNA-A and DNA-B . Many monopartite begomoviruses are associated with satellites [14] . For example , Cotton Leaf Curl Multan virus ( CLCuMuV ) is an important monopartite begomovirus and infects many plant species including cotton and N . benthamiana . CLCuMuV can cause leaf curl disease , a devastating disease of cotton when it is associated with a betasatellite DNA . To effectively infect plants , geminiviruses often need to interfere with PTGS or TGS pathways . Viral proteins such as AC2/C2 , AC4/C4 , V2 , AC5/C5 encoded by different geminiviruses and βC1 protein encoded by viral satellite DNA are capable of inhibiting various steps in the PTGS pathway [15–20] . Other geminiviral proteins including C2/AC2 of Beet curly top virus , Beet severe curly top virus ( BSCTV ) , Tomato golden mosaic virus and Cabbage leaf curl virus [9 , 10 , 12 , 21] , βC1 of Tomato yellow leaf curl China virus betasatellite ( TYLCCNB ) , C4 of Tomato leaf curl Yunnan virus [11 , 22] , Rep of Tomato yellow leaf curl Sardinia virus ( TYLCSV ) and V2 of Tomato yellow leaf curl virus ( TYLCV ) [23 , 24] are identified as suppressors of TGS . Geminivirus AC4/C4 gene overlaps entirely within the replication initiation protein ( Rep ) coding region , but is in a different open reading frame . In the bipartite begomoviruses , African cassava mosaic virus and Sri Lankan cassava mosaic virus , AC4 suppresses PTGS by binding to miRNAs and siRNAs [15 , 25 , 26] . C4 affects the movement of some curtoviruses and monopartite begomoviruses [27–29] and acts as a cell cycle regulator [30] . Geminiviral C4 is one of viral symptom determinants [31–33] , and functions as viral suppressor of RNA silencing for several bipartite and monopartite geminiviruses . Transgenic expression of curtovirus C4 gene induces hyperplasia and alters plant development in N . benthamiana and Arabidopsis . It has been reported that geminiviral C4 interacts with host factors including multiple plant shaggy-like kinases [33–38] . The C4 protein from Tomato yellow leaf curl virus can inhibit the intercellular spread of RNAi [39] . These observations reveal the multifunctional nature of the C4 protein . CLCuMuV C4 also suppresses PTGS [40] , but the underlying mechanism of PTGS suppression by C4 is still unknown . In this study , we report that CLCuMuV C4 protein is able to suppress both TGS and PTGS by interacting with and inhibiting NbSAMS2 , a novel geminiviral target . Further , silencing of NbSAMS2 suppresses both TGS and PTGS and enhances CLCuMuV and TYLCCNV infection , providing direct evidence that interfering with a key enzyme in methyl cycle can promote geminiviral infection . These results provide compelling evidence of the requirement of a functional methyl cycle for both TGS-based and PTGS-based anti-geminiviral defense in plants . Disruption of this process by C4 contributes to successful establishment of geminivirus infection .
To identify CLCuMuV C4 interacting proteins , we employed GFP Trap coupled with mass spectrometry analysis . GFP-tagged CLCuMuV C4 ( C4-GFP ) was expressed transiently in N . benthamiana leaves . The total protein extracts were incubated with GFP-Trap_A beads ( ChromoTek ) . After washing , the purified GFP-tagged C4 proteins were denatured at 98°C , separated on a SDS-PAGE gel ( 12% ) , and visualized by silver staining ( S1A Fig ) . When the Liquid Chromatography with Mass Spectrometric ( LC-MS/MS ) polypeptide profiles were searched against N . benthamiana protein databases , a S-adenosyl methionine synthesase 2 ( NbSAMS2 ) partial sequence was found with high scores ( FVIGGPHGDAGLTGR ) ( S1B Fig ) . At the next stage , we searched the N . benthamiana genome database ( http://solgenomics . net ) using NbSAMS2 , and found eight SAMS homologs . However , we were only able to clone the cDNAs of NbSAMS1 , NbSAMS2 and NbSAMS3 , suggesting that the other five genes are not expressed or poorly expressed in leaf tissues . This finding is consistent with the RNA-seq database ( http://benthgenome . qut . edu . au/ ) , which shows the reads for NbSAMS1 , NbSAMS2 and NbSAMS3 are 147 , 858 and 87 , respectively , but 0 or below 40 reads for the other NbSAMS homologues ( We refer to these SAMS homologues collectively as NbSAMS ) . NbSAMS2 shares 90 . 7% and 91 . 5% amino acid identity with NbSAMS1 and NbSAMS3 , respectively ( S2 Fig ) . To verify the interaction of C4 with NbSAMS , we first employed Firefly Luciferase Complementation Imaging assay [41] . Three different NbSAMSs were fused to the C-terminal domain of luciferase ( cLUC ) to generate cLUC -NbSAMS1 , cLUC-NbSAMS2 and cLUC-NbSAMS3 , and C4 was fused to the N-terminal domain of luciferase ( nLUC ) to generate C4-nLUC . C4-nLUC was co-expressed with cLUC-NbSAMS1 , cLUC-NbSAMS2 or cLUC-NbSAMS3 in N . benthamiana . Chemical signals were detected when C4 was combined with all three NbSAMSs , due to the reconstitution of the luciferase activity by C4-NbSAMS interactions . However , no interaction was detected in negative controls ( S3A Fig ) . These data suggest that C4 interacts with all three NbSAMSs . Considering very high amino acid identities among the three NbSAMS homologues and its highest expression level in leaf tissues , we focused on NbSAMS2 for subsequent analyses . The interaction between C4 and NbSAMS2 was further confirmed by co-immunoprecipitation ( co-IP ) assays . We found that C4-GFP protein co-immunoprecipitated with HA-NbSAMS2 but not with the control protein , HA-cLUC ( Fig 1A ) . Further , C4 amino acids 1 to 20 are critical for interaction with NbSAMS2 ( S3B Fig ) . Moreover , co-IP assays , GST pull-down assays and BiFC assays revealed that mutation of residue 13 from arginine to alanine ( C4R13A ) eliminates the interaction between C4 and NbSAMS2 ( Fig 1 ) . Several other C4 mutants ( A3L , I5A , S6A , C8A , R13K , A14G ) still interact with SAMS . It is worthwhile mentioning that the R13A mutation in C4 does not change the subcellular localization of C4 ( S3D Fig ) . To examine whether C4 interacts directly with NbSAMS2 in vitro , we performed glutathione S-transferase ( GST ) pull-down assays . 6×His-tagged C4 ( C4-His ) or C4R13A ( C4R13A-His ) was expressed in E . coli BL21 ( DE3 ) and purified on a Ni-NTA agarose column . After elution , C4-His and C4R13A-His were incubated with GST-tagged NbSAMS2 ( GST-NbSAMS2 ) or GST-cLUC , respectively . C4-His , but not C4R13A-His , was pulled down with GST-NbSAMS2 , but not GST-cLUC , suggesting that C4 directly interact with NbSAMS2 ( Fig 1B ) . C4-NbSAMS2 interaction was also examined using citrine yellow fluorescent protein ( YFP ) -based bimolecular fluorescence complementation ( BiFC ) assays [42] . Only when C4-nYFP was coexpressed with cYFP-NbSAMS2 , strong YFP fluorescence was visible in cytoplasm in N . benthamiana leaf cells , whereas no YFP fluorescence was detected when C4R13A-nYFP was coexpressed with cYFP-NbSAMS2 ( Fig 1C ) or negative controls ( S4A Fig ) . Immunoblot assays showed that all constructs were successfully expressed ( Fig 1D ) . Taken together , our results demonstrate that C4 interacts with NbSAMSs , and the R13 residue is essential for the C4-NbSAMS2 interaction . Given that C4 interacts with SAMS , it may affect SAMS enzymatic activity . To address this hypothesis , we performed an in vitro assay to test the effect of C4 on NbSAMS2 enzymatic activity by detecting amount of 35S-SAM in the presence of C4 or its mutant . For this , we expressed the fusion proteins GST-C4 , GST-C4R13A and GST-NbSAMS2 in E . coli and purified the proteins using glutathione sepharose . NbSAMS2 was pre-incubated for 20 minutes with varying amounts of GST-C4 , GST-C4R13A or GST . The protein mixtures were then added to solutions containing 35S-Met , dATP and MgCl2 . After incubation at 30°C for 20 minutes , conversion of methionine to SAM was blocked by the addition of EDTA and monitored by thin layer chromatography ( TLC ) . As expected , GST-NbSAMS2 efficiently converted methionine to SAM , showing a strong signal on TLC , but GST alone did not give any signal . Intriguingly , GST-C4 inhibited NbSAMS2 activity and maximal inhibition with approximately an 80% reduction were detected at an 8:1 molar ratio of C4:NbSAMS2 . NbSAMS2 activity was not influenced by GST-C4R13A or free GST ( Fig 2 ) . These results showed that C4 inhibits the enzymatic activity of NbSAMS2 through physical interaction . C4 but not C4R13A inhibits the enzymatic activity of NbSAMS2 . Thus , we asked whether infection of CLCuMuV or CLCuMuV mutant virus ( CLCuMuV-C4R13A , replacing C4 with its mutant counterpart C4R13A ) affect the enzymatic activity of NbSAMS2 and influence the synthesis of SAM . N . benthamiana leaves were inoculated with CLCuMuV or CLCuMuV-C4R13A . At 21 days post inoculation ( dpi ) , SAM was extracted from systemically infected leaves with 5% ( w/v ) trichloroacetic acid , and its level was analysed by LC-MS/MS [43] . CLCuMuV infected plants showed a reduced level of SAM ( 2 . 4 nmol/g ) compared to CLCuMuV-C4R13A ( 4 . 9 nmol/g ) and healthy control plants ( 5 . 6 nmol/g ) ( Fig 2C ) . Taken together , these results showed that CLCuMuV C4 protein inhibits NbSAMS2 enzyme activity by interacting with NbSAMS2 . Given that SAMS is a key enzyme that converts methionine to SAM , a methyl donor for methylation of RNA , proteins and DNA [44] , it may be involved in gene silencing in plants . To test the role of SAMS in TGS pathway , we cloned two 345-bp NbSAMS2 fragment ( at 3′ or 5′ UTR ) and a 345-bp luciferase fragment ( LUC ) into CLCuMuB-based VIGS vector βM2 [45] to generate βM2-SAMS2F1 , βM2-SAMS2F2 and βM2-LUC ( a negative control ) , respectively . We performed a heterologous VIGS assay using TYLCCNV and βM2 to silence NbSAMS2 in 16c-TGS plants , because CLCuMuV encodes a TGS suppressor ( e . g . C4 in this study ) and TYLCCNV does not suppress TGS [11] . 16c-TGS plants contain a transcriptionally silenced GFP transgene flanked by the 35S promoter [2 , 21] . TYLCCNV supports the replication of βM2 ( S4 Fig ) , indicating that the βM2 vector can be used to silence a plant gene when co-agroinfiltrated with TYLCCNV . Indeed , NbSAMS2 mRNA levels were reduced ~ 50% in NbSAMS2-silenced 16c-TGS plants compared to control plants ( S5 Fig ) . As expected , no GFP fluorescence was seen in control plants . However , GFP fluorescence was visible in NbSAMS2 silenced 16c-TGS plants ( Fig 3A ) . GFP expression was confirmed by immunoblot assays in NbSAMS2-silenced but not in control 16c-TGS plants ( Fig 3B ) . The relative level of GFP mRNA in NbSAMS2-silenced plants was 7 to 9 times more than in the controls ( Fig 3C ) . Further , silencing of NbSAMS2 was associated with earlier symptom appearance and enhanced TYLCCNV DNA accumulation ( 4 to 6 times ) ( S5D Fig ) . To test the role of SAMS in the PTGS pathway , we silenced NbSAMS2 in N . benthamiana using Tobacco rattle virus ( TRV ) -based virus-induced gene silencing ( VIGS ) [46] . SAMS2-silenced plants showed albino and dwarf phenotypes 3–4 weeks post inoculation ( S6 Fig ) , perhaps because TRV induced stronger VIGS of SAMS2 than CLCuMuB-based VIGS of SAMS2 ( approximately 90% vs 55% , also see below ) . However , we were able to perform PTGS assays 2 weeks post inoculation . NbSAMS2-silenced plants were co-agroinfiltrated with a binary construct expressing GFP ( 35S-GFP ) [47] and a binary construct expressing the HA-tagged C-terminal domain of luciferase ( cLUC-HA ) or P19-silencing suppressor from Tomato bushy stunt virus ( P19 ) [48] . Five days after inoculation , co-expression of 35S-GFP and TBSV P19 gave GFP fluorescence in both non-silenced and NbSAMS2-silenced plants ( Fig 3D ) . However , co-expression of 35S-GFP and cLUC-HA gave GFP fluorescence in NbSAMS2-silenced plants , but not in non-silenced plants . GFP expression was confirmed by immunoblot analysis ( Fig 3E ) . The level of NbSAMS2 mRNA was reduced in NbSAMS2-silenced N . benthamiana plants ( S7 Fig ) . These results suggest that NbSAMS2 is required for both TGS and PTGS in plants , and involved in antiviral defense against TYLCCNV . To assess how C4-NbSAMS2 interaction contributes to CLCuMuV infection , we silenced NbSAMS2 in N . benthamiana using CLCuMuB-based VIGS [45] , because the transgenic knockdown of SAMS [49] and TRV-mediated silencing of SAMS2 ( S6 Fig ) cause a strong abnormal developmental phenotype . For this purpose , N . benthamiana plants were infected with CLCuMuV plus βM2-SAMS2F1 , βM2-SAMS2F2 or negative control βM2-GFPF containing 345-bp GFP fragment [45] . Both control and SAMS-silenced plants showed a leaf curl symptom . However , viral symptom is much more severe in SAMS2-silenced plants . We observed some whitening veins in SAMS-silenced plants , suggesting a possible involvement of SAMS in pigment synthesis ( Fig 4A ) . Total DNA was extracted from systemic leaves and analyzed DNA gel blots with biotin-labeled probes specific for CLCuMuV . Results indicated that CLCuMuV DNA accumulation significantly increased in SAMS-silenced plants compared to control plants ( Fig 4B ) . CLCuMuV genomic DNA levels were measured by real time PCR using eIF4a as an internal control . The relative CLCuMuV DNA level in SAMS2-silenced plants was 9 to 12-fold compared to the non-silenced controls ( Fig 4C ) . Real-time RT-PCR showed that silencing of SAMS using either βM2-SAMS2F1 or βM2-SAMS2F2 significantly reduced mRNA level of NbSAMS2 ( 50% or 60% , respectively ) , but not of NbSAMS3 ( S8 Fig ) . In addition , silencing of SAMS using βM2-SAMS2F2 , but not βM2-SAMS2F1 , also reduced mRNA level of NbSAMS1 ( 40% ) ( S8 Fig ) . Silencing of SAMS2 show weak or no silencing of SAMS1 and SAMS3 because of the use of 3′ or 5′ UTR fragment of NbSAMS2 to specifically silence SAMS2 . RNA-seq data ( https://solgenomics . net/ ) indicated that the expression levels of SAMS1 and SAMS3 are very low compared to SAMS2 . We examined the effect of silencing of SAMS2 on the methylation status of CLCuMuV genomic DNA by bisulfite sequencing of the viral 5’-intergenic region ( IR ) . This region contains 52 potential methylation sites , including nine CG sites , seven CNG sites , and 36 CHH sites . Cytosine methylation levels of the CLCuMuMV IR in N . benthamiana plants infected with CLCuMuV plus βM2-SAMS2F1 , βM2-SAMS2F2 or the negative control βM2-GFPF are summarized in Fig 4A . The methylation level of the CLCuMuV IR is much lower in SAMS-silenced plants ( 20% ) than the control plants ( 80% ) ( Fig 4D ) . Silencing of SAMS2 caused strong suppression of DNA methylation . Taken together , these results suggest that NbSAMS2 takes part in plant antiviral defenses against CLCuMuV by positively contributing to both TGS and PTGS . Given that C4 interacts with NbSAMS2 and inhibits its enzymatic activity , we assumed that C4 may suppress NbSAMS2-mediated TGS . For this , we tested whether Potato virus X ( PVX ) -based expression of C4 ( PVX-C4-HA ) can reverse GFP expression in 16c-TGS plants . PVX-C4-HA infected 16c-TGS plants showed chlorosis and up-curling symptoms in systemic leaves while PVX-cLUC-HA and PVX-C4R13A-HA control plants only showed mild viral symptoms . At 14 dpi , GFP fluorescence became visible in leaves , especially in phloem tissue of the PVX-C4-HA infected 16c-TGS plants , but not in PVX-C4R13A-HA and PVX-cLUC-HA plants ( Fig 5A ) . The relative GFP mRNA level in PVX-C4-HA infected 16c-TGS plants was 15-fold of that in the PVX-cLUC-HA infected 16c-TGS plants . We verified this result at the GFP protein level inn immunoblotting assays ( Fig 5 ) . Bisulfite sequencing was used to assess the level of TGS suppression in the PVX-C4-HA infected N . benthamiana 16c-TGS plants [3 , 9 , 11 , 50] . We analyzed cytosine methylation at seven CG , four CNG , and 26 CHH sites in the 35S promoter of the GFP transgene ( Fig 5E ) . PVX-based expression of C4 reduced cytosine methylation at CG ( 23% ) , CNG ( 15% ) and CHH ( 8% ) sites compared to PVX-based expression of C4R13A or control protein cLUC ( Fig 5G ) . We also measured the DNA methylation level of Tnt1 retrotransposon . Here , Tnt1 was served as an endogenous epigenetic marker in N . benthamiana . C4 reduced Tnt1 cytosine methylation at CG ( 19% ) , CHH ( 26% ) and CNG ( 20% ) sites compared to C4R13A or control protein cLUC ( Fig 5F and 5H ) . These results indicated that CLCuMuV C4 , but not its mutant C4R13A , is able to reduce cytosine methylation of both an endogenous gene and a transgene in plants . TYLCCNV co-infection with its betasatellite Y10β or the Y10β-encoded protein βC1 alone can reverse TGS [11] . To further confirm TGS suppression activity of CLCuMuV C4 , we replaced βC1 of Y10β with CLCuMuV C4 or 345 bp of luciferase fragment ( LUC ) to generate Y10mβ-C4 or Y10mβ-LUC , that were co-inoculated with TYLCCNV onto N . benthamiana 16c-TGS plants . At 15 dpi , TYCLCCNV plus Y10β caused severe symptoms and TYCLCCNV plus Y10mβ-LUC only caused very mild symptoms . Y10mβ-C4 induced stronger viral symptoms than Y10mβ-LUC , but weaker symptoms than Y10β , when co-infected with TYLCCNV . Further , GFP fluorescence was visible in phloem tissue and young leaves of 16c-TGS line infected with TYLCCNV plus either Y10mβ-C4 or Y10β , but not when infected with TYLCCNV plus Y10mβ-LUC ( S9A Fig ) . GFP expression was confirmed on immunoblots ( S9B Fig ) . Interestingly , CLCuMuV C4 enhanced TYLCCNV DNA accumulation ( S9C Fig ) . These data suggest that CLCuMuV C4 can reverse TGS and impair plant antiviral defense against TYLCCNV . NbSAMS2 is involved in both TGS and PTGS in plants and C4 is a PTGS suppressor . To test whether the interaction of CLCuMuV C4 with NbSAMS2 is required for C4 PTGS suppressor activity , we coexpressed 35S-GFP with cLUC-HA , C4-HA , C4R13A-HA or P19 in N . benthamiana leaves using agroinfiltration to trigger PTGS . Strong GFP fluorescence were observed in tissues co-expressing 35S-GFP with C4-HA or P19 , but not in tissues co-expressing 35S-GFP with C4R13A-HA or cLUC-HA at 5 dpi ( S10A Fig ) . The relative GFP mRNA level in presence of C4 was 9 times more than the controls ( S10B Fig ) . We verified this result GFP mRNA and protein by real-time RT-PCR and immunoblotting ( S10C Fig ) . RNA gel blot analysis detected GFP-specific siRNAs in leaves infiltrated with GFP plus cLUC-HA or C4R13A-HA , but not in those infiltrated with GFP plus P19 or C4-HA and non-infiltrated plants ( S10D Fig ) . These data suggest that CLCuMuV C4 functions as a PTGS suppressor dependent on its interaction with SAMS . Taken together , our results suggest that CLCuMuV C4 is able to inhibit methylation mediated TGS and PTGS . C4-NbSAMS interaction is essential for C4 suppressor activity of gene silencing . Because the C4 gene entirely overlaps the Rep coding region , and the R13A mutation in the C4 protein resulted in a E65G mutation in the Rep protein ( Fig 6A ) . To examine whether the E65G mutation in Rep ( RepE65G ) impacts viral replication , we quantified DNA accumulation of a null mutant virus for the Rep gene , CLCuMuVΔRep ( CLCuMuV with ATG-to-TGA mutation in start codon of Rep ) . For this , we isolated protoplasts from the leaves of N . benthamiana plants , and then transfected them with CLCuMuVΔRep and an expression construct for either Rep or RepE65G . We extracted total DNA at 24 , 48 , 72 hours post transfection , and measured CLCuMuV DNA levels by real time PCR using eIF4a as an internal control . Real-time PCR analysis showed that viral DNA accumulation is similar between cells expressing Rep and RepE65G ( Fig 6B ) . These results suggest that E65G mutation in Rep does not affect the function of Rep protein in viral replication , consistent with the fact that Rep protein of a natural CLCuMuV associated with Okra leaf curl disease has a G at its 65th position ( accession number ADD70021 ) . To further explore biological significance of the C4-NbSAMS2 interaction in CLCuMuV infection , we generated a CLCuMuV mutant virus ( CLCuMuV-C4R13A ) by replacing C4 with its mutant counterpart C4R13A . N . benthamiana leaves were inoculated with CLCuMuV and CLCuMuV-C4R13A . At 15 dpi , viral DNA was detected in systemic leaves in both CLCuMuV and CLCuMuV-C4R13A infected plants . At 21 dpi , CLCuMuV-C4R13A caused an attenuated leaf curl symptom compared to wild-type CLCuMuV ( Fig 6C ) , and viral DNA accumulation was reduced 3-fold in plants infected with CLCuMuV-C4R13A compared to CLCuMuV ( Fig 6D ) . This result was verified on DNA gel blots ( Fig 6E ) . Because the R13A mutation in C4 abolished its interaction with SAMS and the E65G mutation in Rep had no effect on CLCuMuV replication , these results suggest that C4-NbSAMS2 interaction is important for viral infection . We hypothesize that C4 affects methylation of the viral DNA genome to enhance viral accumulation by interacting with and inhibiting SAMS2 . To test this hypothesis , we examined the methylation status of the IR of CLCuMuV and CLCuMuV-C4R13A by bisulfite sequencing . The results indicated that an R13A mutation in C4 enhanced IR methylation at CG ( 13% ) , CNG ( 11% ) and CHH ( 8% ) sites in plants ( Fig 6F ) . Taken together , our data suggest that CLCuMuV C4 contributes to viral infection by inhibiting SAMS-dependent methylation-mediated TGS and PTGS through its interaction with SAMS to inhibit SAMS enzyme activity in plants .
In this study , we showed that CLCuMuV C4 suppresses both TGS and PTGS by inhibiting SAMS activity through its interaction with SAMS in plants . Furthermore , a point mutation from arginine to alanine at position 13 in C4 abolishes the C4-NbSAMS2 interaction and C4-mediated TGS/PTGS suppression , resulting in impaired viral infection and enhanced viral DNA methylation . In addition , silencing of NbSAMS2 reversed methylation-mediated TGS and PTGS and enhanced geminiviral infection . Our work provides direct evidence that a geminivirus is able to interfere with the methyl cycle to promote the effective infection , and reinforces the importance of methylation as an epigenetic defense against geminiviruses . It is unclear how CLCuMuV C4 suppresses SAMS2 activity by interacting with SAMS . CLCuMuV C4 shares 78% identity and 85% similarity with Cotton leaf curl Kokhran virus ( CLCuKoV ) C4 at the protein level . CLCuMuV C4 could also have an ATPase activity like CLCuKoV C4 [51] . We found that CLCuMuV C4 R13 is required for its interaction with SAMS and its suppression of SAMS activity . Interestingly , CLCuKoV C4 R13 is involved in its ATPase activity [51] . Thus , CLCuMuV C4 may use its ATPase activity to suppress SAMS activity . Consistent with this idea , SAMS requires ATP for converting methionine to SAM [8] . We provided several lines of evidence to demonstrate the ability of the CLCuMuV C4 protein to inhibit DNA methylation-mediated TGS . First , both PVX-based and TYLCCNV-based expression of C4 reversed the transcriptionally silenced GFP transgene in N . benthamiana 16c-TGS plants . Second , expression of C4 reduced cytosine methylation of CaMV 35S promoter of the GFP transgene in 16c-TGS lines ( Fig 5 ) , suggesting that the demethylation in CaMV 35S promotor in 16c-TGS plants reactivates expression of the TGS-silenced GFP transgene . Third , expression of C4 reduced cytosine methylation of the Tnt1 retrotransposon in N . benthamiana . Fourth , an R13A mutation in C4 enhanced geminiviral DNA methylation . DNA methylation in plants can occur in different sequence contexts including symmetric ( CG and CHG , where H is A , C , or T ) sites and asymmetric ( CHH ) sites . In this study , we observed that silencing of NbSAMS2 can reverse TGS of GFP transgene , and CLCuMuV C4 reduced cytosine methylation of CG and CNG in the transgene promoter region and CG , CHG , CHH in the retrotransposon Tnt1 and the CLCuMuV genome . The C4-reduced methylation levels of Tnt1 and 35S promoter were different . The reason may be due to the difference in the methylation level of different loci . Indeed , The CHH methylation level of Tnt1 are higher than the methylation of 35S promoter , the reduction of methylation level of Tnt1 ( CHH ) was higher compared to the 35S promoter in the presence of C4 . Further , CLCuMuV carrying C4R13A showed increased viral DNA methylation in plants . Consistent with our observations , knockdown of SAMS effectively represses symmetric cytosine methylation in some key flowering genes [49] . In plants , the addition of a methyl group to DNA or RNA is thought to be one of the major host defense mechanisms against viruses [6 , 9] . Potyviral HCPro from Potato virus A ( PVA ) was reported to bind two methyl cycle-related proteins SAMS and SAHH . Further , knockdown of SAMS and SAHH partially rescues the HCPro-deficient PVA viral phenotype [52] , suggesting that HC-Pro may suppress PTGS through disruption of the methyl cycle . However , the role of HC-Pro-SAMS interaction in PVA-mediated suppression of SAMS activity and PTGS has not been investigated . In addition , the symptoms of CLCuMuV in N . benthamiana are mild ( Fig 6C ) . However , PVX-C4 exhibited severe symptoms in Fig 5A . The severe phenotype caused by PVX may be caused by the methyl cycle inhibition itself or synergistic effect of C4 on PVX , similar to effect of HC-Pro on PVX [52] . Plant DNA viruses have evolved different proteins to interfere with the plant methylation pathway . Adenosine kinase ( ADK ) is a cytoplasmic enzyme involved in the adenine and adenosine salvage pathways , and may have a role in sustaining the methyl cycle in both yeast and plants [53–55] . Geminiviral AC2/AL2 proteins negatively regulate TGS pathway by interacting with and inactivating ADK [9] . S-adenosyl-methionine decarboxylase 1 ( SAMDC1 ) is a key enzyme for the synthesis of polyamines in mammals and plants [56] . BSCTV C2/L2 protein inhibits SAMDC1 activity [10] . TYLCCNV βC1 protein also represses cytosine methylation by interacting with S-adenosyl homocysteine hydrolase ( SAHH ) , a methyl cycle enzyme required for SAM production and methylation-mediated TGS although it is not reported whether inhibiting SAHH affects geminivirus infection in plants [11] . These studies suggest that geminiviruses may disturb the proper function of the cellular methyl cycle to affect TGS . In our study , we demonstrate that CLCuMuV C4 impairs host methylation by directly binding to SAMS and inhibiting its enzyme activity in SAM production . The R13A mutation abolished ability of C4 to inhibit SAMS enzyme activity and to suppress TGS and PTGS . Further , CLCuMuV-C4R13A has less infectious ( Fig 6 ) , suggesting that CLCuMuV C4 suppresses SAMS enzyme activity and gene silencing to enhance geminivirus infection . More importantly , silencing of NbSAMS2 reversed TGS and PTGS and reduced plant resistance against two geminiviruses CLCuMuV and TYLCCNV ( Figs 3 and 4 and S5 Fig ) , providing direct evidence that a geminivirus-encoded protein is able to promote virus infection by interfering with SAMS-mediated methylation-dependent TGS and PTGS . Based on previous research and our work , we propose a working model for C4-SAMS interaction in the regulation of gene silencing . Host methylation modifies the geminivirus genome and protects viral RNA-targeting siRNAs from degradation , promoting TGS and PTGS defense against geminiviruses . It seems that methylation interference via methyl cycle inhibition is a common approach against geminiviruses to counter host methylation dependent silencing defense ( Fig 7 ) . In this model , CLCuMuV-encoded C4 directly targets the core methyl cycle enzyme SAMS and inhibits SAMS activity to generate SAM by the adenylation of methionine . SAM is the methyl donor for most transmethylation reactions . CLCuMuV C4-mediated reduction of SAM levels further decrease viral DNA methylation , promote the stability of geminivirus RNA and enhance virus infection . In this scenario , we cannot exclude the possible involvement of polyamines and viral protein methylation in CLCuMuV infection , because SAM is also the methyl donor for various methylation . Recently , a viral protein was shown to promote ethylene production for the benefit of virus infection by enhancing host SAMS1 activity [43] . It will need further investigation whether CLCuMuV C4 contributes to viral infection by inhibiting the biosynthesis of polyamines and ethylene or viral proteins .
Wild-type N . benthamiana , N . benthamiana line 16c containing a GFP transgene [57] ( provided by xueping Zhou lab ) and 16c-TGS a transcriptionally silenced Green Fluorescent Protein ( GFP ) transgenic line , was generated as described [21] . N . benthamiana plants were grown at 25 °C under a 16-h-light/8-h-dark cycle . Four to five weeks old plants were used in experiments . All experiments were conducted at least three times , with at least 7 plants per construct each time . The infectious clones of CLCuMuV and its betasatellite CLCuMuB , CLCuMuB-based VIGS vector βM2 and its derivative βM2-GFPF ( βM2 containing GFP fragment ) were used as described previously[45] . The βM2-LUC was generated by replacing GFP sequence of βM2-GFPF with 345 bp of luciferase fragment LUC sequence . CLCuMuV-C4R13A was generated by replacing CLCuMuV C4 with C4R13A by overlapping PCR . The infectious clones of TYLCCNV and its betasatellite Y10β were previously described [58] . Y10mβ-LUC and Y10mβ-C4 were generated respectively by replacing the βC1 of Y10β with 345bp of ( LUC ) and CLCuMuV C4 sequences . DNA fragment of NbSAMS2 , C4 and C4R13A were PCR amplified respectively , and then cloned into NdeI-XhoI-digested pGEX4T-1 vector to express GST-tagged fusion proteins GST-NbSAMS2 , GST-C4 and GST-C4R13A in E . coli . Full-length C4 and C4R13A was individually cloned into BamHI-XhoI-digested pET28a to express C4-His and C4 R13A-His in E . coli . For generating T-DNA based fusion protein expression constructs by ligation-independent cloning ( LIC ) -based method , various LIC-based expression vectors were generated as below: pcLUC-LIC was described , and generated by cloning LIC cassette containing ccdB gene flanking LIC adaptors with ApaI site ( PCR amplified using pYL436 ( GenBank Accession: AY737283 ) as a template ) into pCAMBIA-cLUC [41] . pcLUC-LIC is used to generate cLUC fusion protein constructs . NbSAMS1 , NbSAMS2 and NbSAMS3 ORFs were PCR amplified , and then cloned into pcLUC-LIC to cLUC-NbSAMS1 , cLUC-NbSAMS2 and cLUC-NbSAMS3 expression constructs by LIC method as described . LIC-pnLUC was generated by cloning LIC cassette containing ccdB gene flanking LIC adaptors with ApaI site ( PCR amplified using pYL436 as a template ) into pCAMBIA-nLUC [41] . C4 and C4R13A were PCR amplified , and then cloned into LIC-pnLUC to C4-nLUC , C4R13A-nLUC expression constructs by LIC method as described [59] . LIC-pGFP , LIC-pnYFP , pcYFP-LIC and LIC-Pha were generated by replacing cLUC sequence of pcLUC-LIC with DNA fragments ( without stop codon ) of GFP , nYFP , cYFP and 3×HA sequence respectively . DNA fragments of C4 , C4R13A , an N-terminal part of C4 ( N50 , amino acids 1–50 ) , M-terminal part of C4 ( M60 , amino acids 21–80 ) and C-terminal part of C4 ( C50 , amino acids 51–100 ) were PCR amplified respectively , and cloned into LIC-pGFP vector to generate C4-GFP , C4R13A-GFP , C4-N50-GFP , C4-M60-GFP , C4-C50-GFP expression constructs . DNA fragments of C4 and C4R13A were cloned into LIC-pnYFP to C4-nYFP , C4R13A-nYFP expression constructs . cLUC DNA fragment was cloned into LIC-pHA to generate cLUC-HA expression construct . NbSAMS2 ORF was PCR amplified , and cloned into pcYFP-LIC and pHA-LIC respectively to generate cYFP-NbSAMS2 and HA-NbSAMS2 expression constructs . DNA fragments of C4-HA , C4R13A-HA and cLUC-HA were PCR amplified , and then cloned into PVX-LIC vector [60] by LIC method to generate PVX-based expression constructs PVX-C4-HA , PVX-C4R13A-HA and PVX-cLUC-HA . Before LIC cloning , all LIC-based expression vectors except PVX-LIC were digested with ApaI , while PVX-LIC vector was digested with SmaI . LIC cloning was performed as described . For LIC cloning , digested LIC-based expression vector was further treated with T4 DNA polymerase in the presence of dTTP ( 0 . 5 mM ) and DTT ( 1 mM ) for 30 min at 37 °C to produce the 14 nt sticky 5’-end . After the polymerase was inactivated at 75 °C for 20 min , the sticky-ended LIC-based expression vector was purified by phenol extraction and ethanol precipitation . PCR product was treated with T4 DNA polymerase in the presence of dATP ( 0 . 5 mM ) and DTT ( 1 mM ) for 30 min at 37 °C , followed by a 75 °C inactivation step and purified by ethanol precipitation . The equal volumes of LIC-based expression vector and PCR product treated with T4 DNA polymerase were mixed and incubated at 37 °C for 30 min , and then transformed into E . coli . strain DH5α [60] . The resulting constructs were verified by sequencing . Primers sequences and information used for plasmid construction in this study are listed in S1 Table . Total proteins were extracted as described previously [61] from N . benthamiana leaves infected with C4-GFP or GFP constructs respectively . The total protein extracts were incubated with GFP-Trap_A beads ( ChromoTek ) . After washing , the purified GFP-tagged C4 proteins were denatured at 98 °C and separated by SDS-PAGE gel ( 12% ) and visualized by silver staining . Protein bands were excised and in-gel digested with trypsin ( Promega ) and the peptides were extracted twice with 1% ( v/v ) trifluoroacetic acid in 50% ( v/v ) acetonitrile aqueous solution for 30 min . All the peptides were subjected to LC-MS/MS analysis as described previously [61] . For protein analyses , proteins were transiently expressed by agroinfiltration in N . benthamiana leaves and harvested at 2 or 3 dpi . Total proteins were extracted with a ratio of 1:1 of 2 × Laemmli buffer . After boiling for 10 min , protein extracts were separated by SDS-PAGE for immunoblot analysis using indicated antibodies . HA-tagged SAMS ( HA-NbSAMS2 ) was transiently coexpressed with GFP or GFP-tagged C4 or C4R13A- GFP ( C4-GFP ) , in N . benthamiana . Leaf tissues were then collected at 60 hours of post inoculation ( hpi ) . Co-IP experiments was performed as described previously[62] . Total protein extracts were immunoprecipitated using anti-GFP antibody coupled to agarose beads , and the resulting precipitates were analyzed by immunoblot using anti-HA antibodies . For GST pull-down assay , GST-NbSAMS2 and C4-6×His or C4R13A-6×His fusion proteins were produced in BL21 ( DE3 ) cells ( Stratagene ) . GST- NbSAMS2 was purified using glutathione-Sepharose beads ( GE Healthcare ) according to the manufacturer’ s instructions . The GST pull-down assays were performed as described previously[63] . For BiFC assay , proteins were transiently expressed by agroinfiltration in N . benthamiana leaves . The experimental group and corresponding control group were inoculated in a same leaf to reduce the difference of expression condition . The leaves were detached 60 hpi , and confocal imaging was acquired by Zeiss LSM 710 laser scanning microscope ( Carl-Zeiss ) . LCI assays were performed as described [64] . All combinations tested were agroinfiltrated into leaves of N . benthamiana . The leaves were detached 60 hpi , sprayed with 1 mM luciferin , and observed under a low-light cooled CCD imaging apparatus ( iXon; Andor Technology ) . The photos were taken 5 min after exposure . Direct assays of SAMS activity were performed according to the scheme presented in Fig 2A , using SAM production by SAMS as a measure of SAMS-catalyzed MET that yields SAM and free phosphor . The fusion proteins GST-C4 , GST-C4R13A and GST-NbSAMS2 were produced in BL21 ( DE3 ) codon plus RIL cells and purified using Glutathione Sepharose 4B ( GE , USA ) . Target proteins were collected by elution buffer ( 300 mM NaCl , 50 mM Tric-HCl , 10 mM Reduced Glutathione ) at 4°C . Mixtures containing , in a total volume of 15 μl , 10 ng SAMS , and various amounts of C4 were pre-incubated at 30 °C for 20 min . Mixtures were then added to reactions containing ( final concentrations ) 50 mM Tris-HCl , pH 7 . 6 , 5 μCi 35S-Met ( 3000 Ci/mmol ) , 50 mM ATP and 10 mM MgCl2 . Reactions were incubated at 30 °C for 20 min , when SAMS activity was terminated by addition of 1 μl of 1 M EDTA . SAM production was analyzed by thin layer chromatography on poly ethylene iminecellulose plates developed with 1M acetic acid [65] . After chromatography , radioactive signals on plates were quantitated using a phosphor imager ( Bio-Rad Molecular Imager FX ) . GFP recovery assays were performed as described [9 , 21] . Briefly , 16c-TGS plants were agroinoculated with PVX vectors . After the primary harvest , plants were allowed to continue growing under the same conditions and symptom development was observed in new secondary shoots after additional 2–3 weeks . GFP expression was evaluated under long-wavelength UV light and photographed with a Nikon 5000 digital camera ( Tokyo , Japan ) . Genomic DNA was extracted from plant leaf samples using the DNeasy Plant Mini kit ( Qiagen , Valencia , CA ) . To improve the efficiency of bisulfite treatment , DNA ( 1 mg ) was digested with a restriction enzyme that cuts outside the region of interest to decrease the size of DNA , followed by overnight treatment with proteinase K . Bisulfite modification was carried out using the EZ DNA Methylation Gold kit ( Zymo Research , Irvine , CA ) in a PCR machine . Bisulfite-modified DNA was purified using a Zymo-Spin IC column and dissolved in 10 μl of Elution Buffer according to the manufacturer’s instructions . PCR amplification was then carried out using ZymoTaq and products were cloned into a pGEM-T easy vector ( Promega ) . Individual clones were sequenced . Primers were designed against templates and are listed in S1 Table . The statistical analysis was performed using the OriginPro 8 program . The bars denote the SE of the means , asterisks are representing significantly different from each group ( one-way analysis of variance , *p<0 . 05 ) . Total RNA was extracted from apical developing leaves using the Trizol reagent ( TIANGEN , China ) and treated with RNase-free DNase I ( Sigma-Aldrich ) . First strand cDNA was synthesized using 2–5 μg of total RNA with oligo-d ( T ) primer and M-MLV reverse transcriptase ( TIANGEN , China ) . Real time RT-PCR was performed using SYBR Green-based real-time PCR , a 10 μL reaction mixture containing 5 μl Power SYBR Green PCR Master Mix ( 2× ) ( Life , USA ) , 0 . 1 μL of each 20 μM primer and 0 . 3 μL , 60 ng/μL templet were chosen to amplify target sequences . PCR amplification procedures: the first step: 95 °C to 5 minutes . The second step: 95 °C for 5 seconds , 60 °C for 30 seconds , 40 cycles . Third step: from the beginning of 65 degrees , every 5 seconds to increase the temperature of 0 . 5 degrees , until the end of the reaction at 95°C . eIF4a was used as internal control for N . benthamiana for normalization . The values were calculated using the comparative normalized Ct method [66] and all the experiments were repeated at least three times . Data were analyzed and plotted with Origin 8 . 1 . Geminivirus-based VIGS assay and viral replication determination were performed as described previously[45] . Total DNA was extracted from apical developing leaves using the DNAsecure Plant Kit ( TIAN-GEN , China ) . A single copy of CLCuMuV genome was amplified by PCR and then was ligased into pMD19-T ( TaKaRa , Japan ) to generate a CLCuMuV-positive plasmid . A 10-fold serial dilution of the plasmid DNA from 2×108 to 200 copy was prepared and used as the standard . A CLCuMuV-specific primer set ( qCLCuMuV V1-F and qCLCuMuV V1-R ) was used to amplify a 198-bp amplicon . Because the standard curves generated were linear in the whole range tested with a coefficient of regression R2:0 . 99 and calculated slope around -3 . 5 for SYBR Green assay . The copy number of viral DNA can be calculated via Ct value of each sample and the standard curve . To obtain the ratio of viral DNA: plant genome DNA , Plant genome DNA can also be calculated via internal reference method . The genome DNA of healthy N . benthamiana was extracted and a 2-fold serial dilution of the genome DNA from 145ng to 1 . 13ng was prepared and used as the standard . An eIF4a-specific primer set ( qeIF4a-F and qeIF4a-R ) was used to amplify a 60-bp amplicon . The plant genome DNA can be calculated via Ct value of each sample and the standard curve . Total DNA was extracted from apical developing leaves using the DNAsecure Plant Kit ( TIAN- GEN , China ) . Total DNAs ( 100 ng DNA ) separated electrophoretically in 1% agarose gels containing chloroquine ( 20 μg/ml ) , and analyzed by Southern blot hybridization with biotin-labeled probes specific for CLCuMuV . The DNA agarose gel was stained with ethidium bromide as a loading control . After denaturation and neutralization , total DNA was transferred to Hybond N+nylon membranes ( GE Healthcare , Pittsburgh , PA ) . Membranes were hybridized at 55°C to specific probes . To analyze the production siRNAs , low-molecular-mass RNAs were enriched from total RNA as described previously [67] . The enriched small RNAs ( 15 mg ) were fractionated on a 15% denaturing polyacrylamide–7 M urea gel in 0 . 5 × Tris–borate EDTA ( TBE ) buffer . The RNA was transferred to Hybond N+ membranes ( GE Healthcare ) by electroblotting in 0 . 5 × TBE at 400 mA for 1 h . The transferred RNAs was UV crosslinked to the membrane 4 times at 1200 mJ in a UV Stratalinker 1800 ( Stratagene , La Jolla , CA ) . Membranes were stored at 4 °C until probing . One DNA oligonucleotides complementary to N . benthamiana U6 RNA and a mixture of oligonucleotides corresponding to G , F and P regions of GFP mRNA sequences were synthesized and used as probes for siRNA hybridization . The oligos were end-labelled with [γ-32P] ATP in 50 mL reactions containing 1 mM DNA oligo and 7 U T4 polynucleotide kinase . Hybridizations were performed overnight at 42 °C and the membranes were subsequently washed three times ( 10 min each ) at 40 °C with 1 × SSC ( 0 . 15 M NaCl and 0 . 015 M sodium citrate ) supplemented with 0 . 1% SDS . Hybridization signals were detected as described above for Northern blot analysis . Protoplast isolation assay was performed as described previously [68] . We isolated protoplasts from the leaves of N . benthamiana plants , and then transfected them with CLCuMuVΔRep and expression construct of either Rep or RepE65G . Sequence data from this article can be found in the GenBank data libraries under accession numbers: CLCuMuV ( EF465535 ) ; CLCuMuV isolate Okra ( GU574208 . 1 ) ; CLCuMuB ( EF465536 ) ; TYLCCNV ( AJ319675 ) ; TYLCCNB ( AJ781300 . 1 ) ; NbSAMS1 ( KX452091 ) ; NbSAMS2 ( KX452092 ) ; NbSAMS3 ( KX452093 ) ; NbTnt1 ( AJ228076 . 1 ) ; eIF4a ( KX247369 ) ; GFP ( U87973 ) .
|
Geminiviruses are single-stranded DNA ( ssDNA ) viruses that infect a wide range of plant species and are responsible for substantial crop damage worldwide . However , how geminiviruses inhibit plant antiviral gene silencing defense is unclear . Here , we report that a single geminiviral protein CLCuMuV C4 inhibits both plant transcriptional gene silencing ( TGS ) and post-transcriptional gene silencing ( PTGS ) to promote an effective viral infection . We show that CLCuMuV C4 protein interacts with SAMS , a core enzyme in methyl cycle , and inhibits SAMS activity . Overexpression of CLCuMuV C4 reduces the DNA methylation levels of both a transgene and an endogenous locus . Further , silencing of SAMS reduced both TGS and PTGS , and enhanced viral infection while CLCuMuV virus carrying a mutation in C4 that fails to interact with SAMS showed decreased infection . These findings reveal a novel mechanism by which the CLCuMuV C4 protein suppress SAMS mediated TGS and PTGS , leading to enhanced viral infection in plant .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"plant",
"anatomy",
"chemical",
"compounds",
"gene",
"regulation",
"nucleotides",
"organic",
"compounds",
"mutation",
"plant",
"science",
"pyrimidines",
"amino",
"acids",
"epigenetics",
"dna",
"molecular",
"biology",
"techniques",
"dna",
"methylation",
"chromatin",
"research",
"and",
"analysis",
"methods",
"gene",
"silencing",
"small",
"interfering",
"rnas",
"artificial",
"gene",
"amplification",
"and",
"extension",
"chromosome",
"biology",
"proteins",
"gene",
"expression",
"chemistry",
"chromatin",
"modification",
"dna",
"modification",
"leaves",
"methionine",
"molecular",
"biology",
"sulfur",
"containing",
"amino",
"acids",
"biochemistry",
"rna",
"point",
"mutation",
"cytosine",
"cell",
"biology",
"nucleic",
"acids",
"organic",
"chemistry",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"non-coding",
"rna",
"polymerase",
"chain",
"reaction"
] |
2018
|
Cotton Leaf Curl Multan virus C4 protein suppresses both transcriptional and post-transcriptional gene silencing by interacting with SAM synthetase
|
PML nuclear bodies ( PML-NBs ) are enigmatic structures of the cell nucleus that act as key mediators of intrinsic immunity against viral pathogens . PML itself is a member of the E3-ligase TRIM family of proteins that regulates a variety of innate immune signaling pathways . Consequently , viruses have evolved effector proteins to modify PML-NBs; however , little is known concerning structure-function relationships of viral antagonists . The herpesvirus human cytomegalovirus ( HCMV ) expresses the abundant immediate-early protein IE1 that colocalizes with PML-NBs and induces their dispersal , which correlates with the antagonization of NB-mediated intrinsic immunity . Here , we delineate the molecular basis for this antagonization by presenting the first crystal structure for the evolutionary conserved primate cytomegalovirus IE1 proteins . We show that IE1 consists of a globular core ( IE1CORE ) flanked by intrinsically disordered regions . The 2 . 3 Å crystal structure of IE1CORE displays an all α-helical , femur-shaped fold , which lacks overall fold similarity with known protein structures , but shares secondary structure features recently observed in the coiled-coil domain of TRIM proteins . Yeast two-hybrid and coimmunoprecipitation experiments demonstrate that IE1CORE binds efficiently to the TRIM family member PML , and is able to induce PML deSUMOylation . Intriguingly , this results in the release of NB-associated proteins into the nucleoplasm , but not of PML itself . Importantly , we show that PML deSUMOylation by IE1CORE is sufficient to antagonize PML-NB-instituted intrinsic immunity . Moreover , co-immunoprecipitation experiments demonstrate that IE1CORE binds via the coiled-coil domain to PML and also interacts with TRIM5α We propose that IE1CORE sequesters PML and possibly other TRIM family members via structural mimicry using an extended binding surface formed by the coiled-coil region . This mode of interaction might render the antagonizing activity less susceptible to mutational escape .
Promyelocytic leukemia protein PML is the organizer of small nuclear matrix structures termed nuclear bodies ( NBs ) or nuclear domain 10 ( ND10 ) [1] . PML , also named TRIM19 , is a member of the tripartite motif ( TRIM ) family of proteins , which are characterized by the presence of RING , B-box and coiled-coil domains [2] . Recent studies showed that an unprecedented large number of TRIMs positively regulate innate immune signaling pathways by acting as E3-Ub ligases [3] , [4] . Additionally , a subgroup of TRIMs , including PML , was demonstrated to exhibit small ubiquitin related modifier ( SUMO ) E3 activity and PML itself is covalently conjugated to SUMO on three lysine residues [5] , [6] . This modification , which affects PML localization , stability and interaction with other partners , is critical for NB functions [7] . In response to stimuli , PML-NBs recruit a number of proteins implicated in different cellular processes such as DNA damage response , apoptosis , senescence and protein degradation [8] , [9] . Accumulating evidence implicates this subnuclear structure as an important component of intrinsic immunity against viruses from different families including herpes- , adeno- , polyoma , rhabdo- and retroviruses [10]–[12] . Unlike the innate and adaptive immunity , the intrinsic immune response is mediated by cellular restriction factors that are constitutively expressed and permanently active , even before a pathogen enters the cell . Other characteristics of intrinsic immune mechanisms are that they are saturable and subject to viral countermeasures [13] . Besides PML , a number of NB components , such as Sp100 , hDaxx and ATRX , function as cellular restriction factors . Recent evidence suggests that NB proteins independently contribute to the repression of herpesvirus replication , raising the concept that individual NB components , rather than the PML-NB structure as a whole , restrict viral infections [14]–[20] . Consequently , various viruses have been shown to antagonize the intrinsic cellular defense via the modification of NB proteins . For instance , the herpes simplex virus type I immediate-early protein ICP0 has been described as a viral ubiquitin ligase with preferential substrate specificity for SUMO-modified isoforms of PML thus promoting the degradation of PML [21] . However , no structural information on this and many other NB-antagonistic proteins is available , yet . Human cytomegalovirus ( HCMV ) , a ubiquitous beta-herpesvirus causing serious disease in immunocompromised individuals , encodes an abundant immediate-early protein termed IE1 that modulates innate immune mechanisms as well as other cellular processes ( reviewed in [22] ) . Although IE1 is a major player in the initiation of lytic HCMV infection and has been subject to extensive studies over the last decades , structural data on this protein are still limited . Four distinct regions have been identified within the 491 amino acid IE1 protein: a short N-terminal region that is required for nuclear import , a large core domain , an acidic region near the C-terminus that harbors a SUMOylation site and a 16 amino acid chromatin-tethering domain ( CTD ) at the extreme C-terminus [23]–[26] . Recent results have suggested that the acidic C-terminal region of IE1 is characterized by a lack of well-defined three-dimensional structure , but contains a binding motif for signal transducer and activator of transcription ( STAT ) proteins . This interaction site enables IE1 to compromise STAT-mediated interferon signaling , thereby interfering with a crucial branch of the innate immune system and promoting viral replication [27]–[29] . In addition to its effects on the innate immune system , IE1 is required to overcome the PML-NB-mediated intrinsic immunity that targets HCMV immediately upon infection . IE1 transiently co-localizes with PML-NBs during the first 2–4 hours after infection but subsequently induces disruption of these structures [30]–[32] . NB dispersal correlates with the functional activities of IE1 during infection and a PML knock-down efficiently compensates for IE1 in promoting replication of an IE1-deficient virus , establishing IE1 as an important antagonist of PML-mediated cellular repression of viral replication [15] , [16] , [33] . Studies on the mechanism of NB dispersal have demonstrated that IE1 induces the loss of the SUMOylated forms of PML , and also influences the SUMOylation state of Sp100 [34] , [35] . However , in contrast to ICP0 , this neither requires proteasomal activity nor does IE1 affect the abundance of unmodified PML [24] , [35] . In further studies , a physical interaction between IE1 and PML , which requires the N-terminal TRIM region of PML , has been detected as prerequisite for the transient co-localization and subsequent disruption of PML-NB integrity . The interaction site for PML has been mapped to the large core region of IE1 , since deletions or mutations affecting this domain abrogate PML binding and NB disruption [23] , [35] , [36] . However , it was noted in several reports that mutations in the core region often result in unstable IE1 proteins , so that the molecular basis for the IE1-PML interaction remains uncharacterized [37] , [38] . Here we report the crystal structure of the evolutionary conserved globular core domain of primate cytomegalovirus IE1 proteins , determined to 2 . 3 Å resolution . Unexpectedly , the overall structure does not resemble any known protein fold , but exhibits an unusual all α-helical , femur-like shape which shares secondary structure features recently observed in the coiled-coil domain of TRIM proteins . We show that this IE1CORE domain binds with high affinity to PML via the coiled-coil domain . This induces PML de-SUMOylation thus releasing the PML-associated factors hDaxx , Sp100 and ATRX , while PML accumulations itself are not dispersed . Since IE1CORE efficiently complements lytic replication of an IE1-deleted HCMV , we conclude that sequestration of PML via IE1CORE is sufficient for antagonization of NB-mediated intrinsic immunity . Thus , cytomegaloviruses may have evolved a distinct structural fold to effectively bind and neutralize an important cellular hub protein that exerts critical roles during the regulation of innate immune responses as well as the control of programmed cell death [8]–[10] , [12] .
In order to further clarify the mechanism of IE1-mediated PML antagonization , we investigated the molecular architecture of the IE1 proteins from human , chimpanzee and rhesus cytomegalovirus ( h- , c- and rhIE1 ) ( Figure S1 ) . As previously proposed by Krauss et al . [28] , in silico predictions using the web server IUPred [39] suggested that the N- and C-terminal regions of all IE1 proteins display consistently high intrinsic disorder propensities ( Figure 1A ) . Based on these predictions as well as on sequence conservation and the characterization of protease-resistant IE1 subdomains , we generated truncated IE1 constructs covering the folded core ( Figure 1B ) . Limited proteolysis of recombinant full-length hIE1 as well as of C- or N/C-terminally truncated hIE1 proteins confirmed the in silico predictions ( Figure 1C ) . These studies revealed the existence of a stably folded IE1CORE domain of about 360 residues ( Figure 1C , hIE1 20-382 ) that is flanked at the N- and C-termini by intrinsically disordered regions ( IE1N-IDR and IE1C-IDR ) . Circular dichroism ( CD ) spectroscopy [40] , [41] was applied to investigate the secondary structure composition of the IE1 variants . All hIE1 proteins produced typical α-helical spectra with negative ellipticity above 200 nm and two distinct minima at 208 nm and 222 nm ( Figure 1D and S2 ) . The spectra of full-length hIE1 and hIE1CORE differed in the region below 210 nm . Calculation of the difference spectrum revealed strong random coil characteristics thereby confirming the predicted predominantly disordered nature of the terminal regions ( Figure S2 ) . Sequence identities between 24 and 73% between h- , c- and rhIE1CORE domains suggest that the core domains share identical folds . Indeed , the CD spectra of the recombinant proteins hIE1CORE , cIE1CORE and rhIE1CORE match extremely well and indicate that all core domains consist mainly of α-helical segments ( Figure 1D ) . Crystallization trials with full-length hIE1 remained unsuccessful and , in case of the IE1CORE variants , yielded suitable crystals only for rhIE1CORE after chemical methylation of surface exposed lysine residues . The structure of rhIE1CORE was solved using experimental phases and refined to 2 . 3 Å resolution ( Rwork = 19 . 73% , Rfree = 24 . 96% ) ( Figures 2 and S3; Methods and Table S1 ) . The main chain of the model was traced between amino acids 41 and 393 . Since no well-defined electron density was visible for residues preceding residue 41 or following residue 393 , we conclude that the core domain spans at least amino acids 42 to 392 of rhIE1 corresponding to residues 27 to 379 of hIE1 . RhIE1CORE consists of a total of 11 α-helices ( Figure 2 and S1 ) . Helices H3 and H9 are unusually long and contain as many as 16 and 17 helical turns , respectively . RhIE1CORE adopts an elongated , femur-like shape with dimensions of 130×25×25 Å3 . The structure can be divided into three distinct regions , namely an N-terminal head region ( rhIE1: residues 62–118 , 236–283; hIE1 residues 46–103 , 221–267 ) and a C-terminal head region ( rhIE1 residues 151–207 , 315–393; hIE1 residues 136–192 , 300–380 ) interconnected by a stalk region ( rhIE1 residues 41–61 , 119–150 , 208–235 , 284–314; hIE1 residues 27–45 , 104–135 , 193–220 , 268–299 ) ( Figure 2 ) . The stalk consists of an uncommon right-handed three-helix coiled-coil ( α-helices H3 , H6 and H9 ) with the N-terminal helix H1 added to one side of the three-helix bundle ( Figure 2 ) . The right-handed pairing of the helices goes in hand with the presence of hendecad repeats in the sequences of these helices [42] . In these repeats of 11 residues ( numbered alphabetically abcdefghijk ) the hydrophobic amino acids at positions a , d , and h are interspaced by 2 ( bc ) , 3 ( efg ) and 3 ( ijk ) amino acids of predominantly polar nature ( Figure S1 ) . In contrast , the patterning of hydrophobic residues within the head regions of rhIE1CORE frequently resembles that observed in heptad repeats ( residue labeling abcdefg ) . Here the hydrophobic amino acids at positions a and d are interspaced by 2 ( bc ) and 3 ( efg ) polar residues , and the interdigitation of the a and d residues from neighboring helices gives rise to more commonly observed left handed coiled-coil supersecondary structure elements [42] . Hence , both head regions display left-handed coiled-coils . Whereas the N-terminal head region consists of a three-helix bundle ( α-helices H3 , H7 and H8 ) with an additional helix H2 added onto one side of this bundle , the C-terminal head-region comprises 5-helical segments ( H4 , H5 , H9 , H10 and H11 ) in total . These can be grouped into two pairs of antiparallel left-handed coiled coils ( H4–H5 and H9–H10 ) and an additional C-terminal helix ( H11 ) . The coiled-coil helix pairs pack against each other with crossing angles of approximately 50° , and thus the interactions between these coiled-coils resemble the ridges into grooves side chain packing observed in globins [43] . RhIE1CORE also displays an extended loop region between helices H1 and H2 ( residues 62 to 82 ) devoid of secondary structure elements . The conformation of this loop region is stabilized through extensive crystal packing contacts and differs between the two monomers of the dimeric unit as observed within the crystal ( see below ) . Overall , the structure of rhIE1CORE is in full agreement with the observed CD spectrum . Since , with the exception of the very terminal helices ( H1 , H2 , H10 and H11 ) , all intervening helical segments span the entire length of the molecule , IE1CORE is described best as consisting of a single contiguous domain ( Figure 2 ) . A search for structurally similar proteins revealed only partial hits that cover less than 50% of the total rhIE1CORE length . The top-scoring hits belong to a considerable variety of domain folds which either contain α-helical orthogonal bundles or up-down bundles that partly resemble the IE1 head or stalk region , respectively ( Table S2 ) . This indicates that IE1 cannot readily be assigned to any known topology and suggests that the overall fold of IE1 is so far unique . An extended search for local structural similarities , which also considered multimeric proteins , revealed a similarity between IE1 and the recently described coiled-coil region of homodimeric TRIM25 ( Figure 3 ) [44] . The coiled-coil region of TRIM25 is composed of three helices in which the long helices H1 and H1' ( from the second monomer ) align in an antiparallel fashion to form the TRIM25 dimer ( Figure 3B ) . The H1/H1' helix pair can be superimposed onto rhIE1CORE such that the helices superimpose with the H5/H6 and H8/H9 helices of IE1 . The crystal structure of TRIM25 is also highly similar to that of two further TRIM family members , namely TRIM69 [45] and TRIM5α [46] . These structures suggest that the coiled-coil topology may be conserved among the entire family of TRIM proteins , possibly extending to PML ( TRIM19 ) , the target protein of IE1 . This is further corroborated by sequence comparisons , demonstrating that TRIM family members , including PML , exhibit a distinct pattern of heptad and hendecad repeats for helix H1 ( Figure S4 ) [44] . Interestingly , the topological arrangement of helices H1 to H3 of rhIE1 closely resemble the topology of helices H1 to H3 in TRIM25 when allowing for an inversion of the sequential order of the helices ( Figures 3C and D ) . This also extends to the joint presence of heptad and hendecad repeats in H1 in TRIM25 and H3 in rhIE1 . Whereas in TRIM25 , the hendecad repeats occur in the central segment of helix H1 and are flanked on both sides by heptad repeats , H3 in rhIE1 displays a number of heptad repeats towards its N-terminus and switches to a segment of hendecad repeats that covers the second half of helix H3 ( Figures S1 and S4 ) . Taken together , sequence comparisons and three available coiled-coil structures demonstrate that the pattern of heptad and hendecad repeats is highly conserved across the TRIM protein family , and is also present in the viral IE1 protein . These observations suggest a common architecture of TRIM coiled-coils and provide evidence for structural similarities between IE1 and TRIM proteins . RhIE1CORE forms a dimer with C2 point group symmetry in the crystal ( Figure 4A ) , and oligomerization of rhIE1 and hIE1 was confirmed both by gel filtration experiments and co-immunoprecipitation analyses ( hIE1 ) ( Figure 4B and C ) . RhIE1CORE dimerizes with both stalk regions juxtaposed in an antiparallel fashion ( Figure 4A ) . The main-chain conformation differs in the two monomers ( Cα-RMSD for all helical segments = 2 . 13 Å , Figure S5 ) . This deviation originates from a pronounced kink that is observed in one of the two monomers and that causes a repositioning of the loop that interconnects helices H8 to H9 with a concomitant displacement of the N-terminal half of helix H9 ( Figure S6 ) . This kink solely occurs in the tetragonal crystal form whereas in the monoclinic space group all four monomers in the asymmetric unit are highly similar . Since this space group transition is triggered by the experimental dehydration of the crystals , we propose that this in situ molecular shaping reflects an inherent flexibility of the IE1CORE fold that allows for small readjustments in the packing of the helices . Intermolecular contacts are formed along the entire rhIE1 length resulting in an extraordinary large interface area ( Ø 3173 Å2 per molecule ) . In the stalk region , these contacts are predominantly hydrophilic , whereas several sparse hydrophobic patches are formed between head regions . Because of its predominantly hydrophilic nature , the dimer interface does not resemble interfaces typically observed in permanent oligomers , suggesting that the dimer could become disrupted upon interaction with binding partners . This idea is further corroborated by the analysis of IE1CORE surface conservation indicating that the dimer interface is not higher conserved than the solvent exposed regions ( Figure 4D ) . Evolutionary conserved surface patches are distributed almost over the entire surface of IE1CORE ( Figure 4D , blue ) , whereas non-conserved patches are mainly restricted to loop regions ( Figure 4D , red ) . This indicates that the overall biophysical properties are conserved within the IE1 family of proteins despite the rather low degree of sequence identity . This is in line with the results of a molecular model generated for hIE1 based on the rhIE1 crystal structure ( Figure 5A ) . This model exhibits a good global and local quality further indicating that hIE1CORE and rhIE1CORE adopt highly similar folds ( Figure S7 ) . Due to the observed structural conservation we asked whether rhIE1 is likewise capable of disrupting human PML-NBs . Infection of primary human fibroblasts with rhesus macaque cytomegalovirus ( RhCMV ) revealed an initial accumulation of rhIE1 at PML-NBs followed by a dispersal of NBs ( Figure 5B ) . Furthermore , RhCMV infection resulted in a depletion of polySUMOylated PML species ( Figure 5C ) and the isolated rhIE1 expression was sufficient to redistribute PML ( Figure 5D , lower panel ) indicating that hIE1 and rhIE1 do not only share biophysical properties but also functional activities . In order to study whether the function of hIE1CORE differs from that of full-length hIE1 , we investigated the subcellular localization of the hIE1CORE ( Figure 6A ) . For this purpose , primary human fibroblast cells ( HFFs ) were transfected with eukaryotic expression plasmids encoding full-length hIE1 or truncated hIE1CORE proteins . Surprisingly , while full-length hIE1 exhibited a dispersed nuclear localization and induced a loss of PML-aggregates , all truncated proteins showed a punctate staining pattern colocalizing with PML foci . Based on these data , we conclude that a region within the C-terminal hIE1IDR is necessary for PML-dispersal . In contrast , hIE1CORE alone was sufficient to induce deSUMOylation of PML in transient expression experiments using 293T cells ( Figure 6B ) . Consistent results were obtained with a whole cell population of HFF cells stably expressing the hIE1CORE variant 1–382 ( Figure 6C ) . Given that SUMO modification of PML is a prerequisite for the recruitment of other NB components like Sp100 , hDaxx and ATRX , it was important to explore the subcellular localization of these factors after expression of hIE1CORE . Interestingly , while PML was detected in a dot-like pattern , Sp100 , hDaxx and ATRX were released from NBs in the presence of hIE1CORE ( Figure 6C ) . Taken together , these data demonstrate that hIE1CORE is sufficient to sequester and deSUMOylate PML resulting in the dissociation of other NB components . Due to the accumulation of hIE1CORE at PML foci , it was attractive to speculate that the two proteins might strongly interact with each other , which was investigated by co-immunoprecipitation ( Figure 7A ) . Intriguingly , while only a trace amount of PML was associated with full-length hIE1 , PML was efficiently coprecipitated with hIE1CORE variants . An increased affinity of hIE1CORE for PML was also confirmed by yeast two-hybrid experiments ( Figure 7B ) , which is in line with previous results by Lee et al . ( 2004 ) that show an enhanced interaction of PML with an IE1 variant lacking the acidic C-terminus ( IE1 1–420 ) [35] . Having observed a structural similarity between IE1 and coiled-coil regions of TRIM proteins , we asked whether this domain of PML is required for binding of IE1 . In a yeast two-hybrid analysis utilizing a series of C-terminal PML deletion mutants we observed that a truncation of the coiled-coil domain abrogated the interaction with IE1 ( Figure 7C ) . To further confirm this finding , coimmunoprecipitation analyses were performed with additional N- and/or C-terminal PML deletions . Importantly , this experiment revealed that the coiled-coil domain of PML was sufficient to mediate an interaction with IE1 ( Figure 7D , lower panel , lane 4 ) . Furthermore , we observed that IE1 was also able to bind to TRIM5αsuggesting that IE1 targets additional TRIM factors via coiled-coil interactions ( Figure 7E ) . Having shown that hIE1CORE binds with high affinity to and deSUMOylates PML , but fails to disrupt PML accumulations , it was important to investigate whether this is sufficient to antagonize PML-NB mediated repression of viral infection . We constructed a recombinant HCMV expressing hIE1 lacking the C-terminal IE1IDR ( Figure 8A ) and could observe that this virus exhibited a severe defect to disperse PML after infection of HFFs ( Figure 8B ) . Consistent with our results obtained after isolated expression of hIE1CORE , deSUMOylation of both PML and Sp100 was fully preserved ( Figure 8C ) . Most importantly , however , the hIE1CORE-expressing virus replicated nearly as efficient as wild-type virus while an hIE1-deleted virus exhibited a severe growth defect ( Figure 8D ) . The approximately 10fold growth reduction observed for the hIE1CORE-expressing virus at 4 and 6 dpi is in line with previously published results on viruses lacking the C-terminal acidic domain which binds STAT2 thus antagonizing the interferon response [28] , [29] . This was also confirmed in a complementation experiment after infection of either hIE1- or hIE1CORE-expressing HFFs with an hIE1-deleted HCMV , finally demonstrating that hIE1CORE can efficiently substitute for full-length hIE1 during lytic HCMV infection ( Figure 8E ) .
The immediate-early protein IE1 of human cytomegalovirus that directly binds to PML is known as an important herpesviral antagonist of PML-NB-mediated intrinsic immunity [15] , [22] , [33] , [36] . However , the structural basis for its function has remained elusive due to the paucity of high-resolution structural information . Here , we present the first crystal structure for the evolutionary conserved primate cytomegalovirus IE1 proteins and demonstrate that a structurally conserved IE1CORE domain is sufficient to antagonize PML-mediated intrinsic immunity . The structure of IE1CORE consists of a femur-shaped bundle of helices , which surprisingly does not share any overall fold similarity with known protein structures . IE1CORE binds with high affinity to PML and efficiently abrogates PML SUMOylation , but fails to disrupt PML accumulations itself . Only upon inclusion of the C-terminal , intrinsically disordered region ( IE1C-IDR ) PML dispersal is observed . Thus , our study demonstrates that PML deSUMOylation can be discriminated from PML dispersal . Whereas the first activity is achieved by a distinctly folded IE1CORE domain , the second activity requires inclusion of C-terminal sequences of the IE1C-IDR region that is highly susceptible to proteolytic degradation and for which we did not observe any stable secondary structure formation . As it has been observed for many intrinsically disordered proteins , folding of the natively disordered IE1C-IDR region may occur upon binding to a specific interaction partner . First evidence for this comes from a recent study predicting that the chromatin-tethering domain ( CTD ) at the extreme C-terminus of IE1 forms a β-hairpin when bound to histone proteins [47] . Importantly , IE1CORE is able to release other NB-components like Sp100 , hDaxx and ATRX into the nucleoplasm and this correlates with antagonization of NB-mediated repression . This shows that PML dispersal as observed during infection with herpes simplex virus type I and HCMV is not a prerequisite to antagonize the repressive effects of this cellular multiprotein complex on viral gene expression [24] , [48] . As also suggested by recent findings on the γ-herpesviruses Herpesvirus saimiri and Kaposi sarcoma herpesvirus as well as on the polyomavirus BKV more subtle modifications like the release or degradation of individual NB-components appear to be sufficient [49]–[51] . IE1CORE resembles the action of the SUMO-specific protease SENP-1 which also abrogates the SUMOylation of PML but leaves most PML aggregated [52] . It was previously speculated that IE1 could harbor an intrinsic SUMO protease activity itself or recruit SUMO-specific proteases to the NBs [53] . However , our structural analysis of IE1CORE provides no evidence for the presence of a potential active site with hydrolase activity . Furthermore , in earlier studies no interaction of IE1 with SENPs could be detected , but it was reported that full-length hIE1 could still disassemble foci formed by a PML protein with all SUMOylation sites mutated [53] . Based on this observation , the authors raised the idea that SUMO-independent interference with PML oligomerization followed by exposure of SUMOylated PML to cellular SUMO proteases may account for NB disruption . However , the results of our study argue against such a scenario , since abrogation of PML SUMOylation by IE1CORE was detected while PML aggregates were still present . Thus , IE1 affects its targets via direct , SUMO-independent substrate interaction and this suggests that IE1 does not directly or indirectly act as a hydrolase that specifically targets SUMOylated PML . Importantly , our study revealed structural similarities between IE1CORE and the crystal structure of the tripartite motif coiled-coil that appears to act as a critical scaffold organizing the biochemical activities of TRIM proteins [44] , [45] . Moreover , we were able to confirm that the coiled-coil of PML is sufficient for strong binding to IE1 . Increasing evidence suggests that TRIM proteins function as E3-ubiquitin ligases in agreement with the family-wide presence of several conserved domains , namely a RING domain followed by two B-boxes and a coiled-coil region [3] . Based on the recently solved crystal structure of the TRIM25 coiled-coil it was shown that TRIM proteins dimerize by forming interdigitating antiparallel helical hairpins that position the N-terminal catalytic RING domain at opposite ends of the dimer and the C-terminal substrate-binding domains at the center [44] . For some of the TRIM members , and among these PML , E3-SUMO instead of E3-ubiquitin ligase activity has been reported [5] . Thus , we would like to propose that IE1CORE , via its strong interaction with the PML coiled-coil , may inhibit an E3-SUMO ligase activity of PML that is required for auto-SUMOylation . Alternatively , IE1 binding to the coiled-coil might block the accessibility of PML for other components of the cellular SUMOylation machinery . Thus , the results of our study favor a model whereby IE1 primarily affects the on-rate of SUMO modification which is also supported by the slow kinetics of IE1-mediated loss of PML SUMOylation [54] . This is different from the ICP0 protein of herpes simplex virus type I which induces the rapid degradation of SUMO-conjugated proteins by acting as a SUMO-targeted ubiquitin ligase ( STUbL ) [21] . Similar to IE1 , the adenoviral E4-ORF3 protein which has been shown to form a multivalent matrix via extensive self-interactions , appears to inactivate PML via tight binding [55] . This specific assembly of E4-ORF3 creates avidity-driven interactions that capture PML as well as other tumor suppressors thus disrupting PML bodies . However , in contrast to IE1 , the recently solved crystal structure of E4-ORF3 revealed the molecular mechanism of multimerization , but not the exact mode of PML recognition [55] . In this context it should be noted that the nonstructural NS1 protein of influenza A virus has also been shown to target a TRIM protein , TRIM25 , via interaction with the coiled-coil domain to inhibit its E3 ligase function [56] . Since TRIM25 catalyzes a critical ubiquitination of the viral RNA sensor RIG-I this constitutes a mechanism by which influenza virus inhibits the host IFN response . Interestingly , we detected that IE1CORE not only binds to PML but also to TRIM5α and a recent publication reported an interaction with TRIM33 [57] . Thus , the unique structure of IE1core may have developed during evolution to target an extended spectrum of members of the TRIM family via the conserved coiled-coil domain of these factors [44] , [45] . This is also supported by our analysis of evolutionary conserved surface patches of IE1CORE . When assuming that sites of protein-protein interaction are reflected by conserved surface patches , our observation that conserved residues are distributed evenly over the entire IE1CORE protein surface suggests that rather large parts of the IE1 surface are involved in recognition of the PML coiled-coil . Consequently , the helical structure of IE1CORE might have evolved as a decoy that , by means of extensive helix-helix interactions might either pair up with the coiled-coil region of PML or substitute for one of the PML monomers within the PML dimer interface . In this respect , the similarity between the topological arrangement of helices H1 to H3 of IE1 and of predicted helices H3 to H1 of PML in combination with the joint occurrence of regions with extended hendecad repeats might facilitate the formation of heteromeric assemblies . The formation of extended coiled-coil interactions would also readily offer an explanation for the finding that single mutations within the conserved surface patches of IE1 only moderately affect its interaction properties with PML . In contrast , mutations affecting the overall tertiary structure ( e . g . IE1 L174P ) abrogate the functionality . Furthermore , it agrees with our observation that rhIE1 can substitute for hIE1 during infection of human cells despite low overall sequence identity . Thus , the size and unique elongated fold of the IE1CORE could have developed during evolution to accommodate efficient binding of PML and possibly other TRIM factors via an extended surface involving coiled-coil interactions . This feature might render the interaction less amenable to mutational escape .
All variants of h- , c- and rhIE1 were recombinantly produced in E . coli strain BL21 ( DE3 ) ( Novagen ) as GST-tagged fusion proteins for in vitro experiments and crystallization . LB media ( Carl Roth GmbH + Co . KG , Karlsruhe , Germany ) were inoculated with freshly transformed E . coli colonies , and cell cultures grown at either 30° or 37°C . Seleno-methionine labeling of rhIE1 ( residues 36–395 ) was achieved by incubation of the cells with non-inducing PAG medium ( pre-culture ) and auto-inducing PASM-5052 medium ( main culture ) . Cell pellets were resuspended in phosphate buffer and lysed by sonication . Protein purification was achieved by the following steps: a first affinity chromatography ( Glutathione sepharose , GE Healthcare , Freiburg ) , proteolytic cleavage with PreScission protease , a second affinity chromatography and a final size exclusion chromatography ( Superdex 200 prepgrade , GE Healthcare ) . The gel filtration column was pre-equilibrated in 25 mM Tris , 150 mM NaCl , 10 mM DTT , pH 7 . 5 . The samples were separated with an isocratic gradient of 1 . 2 column volumes ( CV ) of the same buffer at a flow rate of 1 . 5 mL/min . The column was calibrated utilizing the elution peaks of thyroglobulin ( 670 kDa ) , bovine γ-globulin ( 158 kDa ) , chicken ovalbumin ( 44 kDa ) and equine myoglobin ( 17 kDa ) of the Bio-Rad gel filtration standard ( Bio-Rad Laboratories , Munich , Germany ) . The molecular weight of the samples was determined by linear regression . The Kav coefficients of the standard proteins were plotted vs the logarithm of their molecular weights to obtain the calibration curve , with Kav = ( Ve-V0 ) / ( Vc-V0 ) , where V0 is the column void volume , Ve is the elution volume and Vc is the geometric column volume . All purification steps were performed in the presence of 10 mM DTT . For the crystallization of variant rhIE1 ( 36–395 ) the protein was chemically modified by lysine methylation prior to the final size exclusion chromatography step . A 1 mg/mL IE1 protein solution was incubated on ice with 20 µL of 1 M dimethylamine borane ( DMAB ) and 40 µL of 1 M formalin per mL of IE1 solution . After two hours the addition of DMAB and formalin was repeated and , following an additional two-hour incubation , 10 µL of 1 M DMAB per mL of IE1 solution were added , and the solution was incubated at 4°C overnight . The reaction was quenched by adding 125 µL 1 M Tris/HCl , pH 7 . 5 per mL of IE1 solution , and the protein was stabilized by addition of 10 mM DTT [58] . Following the final chromatography step , the protein samples were concentrated to 20 mg/ml and stored at −20°C in 25 mM Tris/HCl , 1 . 5 mM NaCl , 15 mM DTT , 1 mM EDTA , pH 7 . 4 before further usage . Limited proteolysis was performed in order to probe the conformational architecture of the protein [59] . The assay was conducted at 21°C with protein concentrations between 0 . 2 and 0 . 5 mg/mL and 0 . 014 mU subtilisin ( Sigma-Aldrich ) per µg IE1 protein . Aliquots of 10 µL were taken at different time points , for example at 1 min , 10 min , 30 min , 60 min , 120 min , 180 min , 240 min and 300 min , mixed with 3 . 3 µL 4× SDS loading buffer and boiled at 95°C for 5 min to stop the cleavage reaction . Circular dichroism spectra were recorded between 185 and 260 nm from protein samples containing 1 . 5 µM or 2 µM protein for full-length or truncated IE1 , respectively . The measurements were performed in 10 mM potassium phosphate buffer , pH 7 . 5 with a Jasco J-815 spectropolarimeter ( Jasco , Tokyo , Japan ) at 20°C with standard sensitivity . The cuvette had a path length of 0 . 1 cm , the band width was 1 . 0 nm , the scan speed 20 nm*sec−1 , data integration time 1 sec and the data pitch 0 . 1 nm . All measurements were accumulated ten times and corrected for the sample buffer . Conversion of the data to concentration-independent mean residual weight ( MRW ) ellipticities [θ]MRW was done as described previously [40] . Initial crystallization conditions were identified with a sparse matrix screening approach ( Index Screen , Hampton Research , Aliso Viejo , USA ) and a Phoenix protein crystallization robot ( Art Robbins Instruments , Sunnyvale , USA ) [60] . The crystallization conditions were optimized in a hanging drop vapor diffusion setup and involved microseeding . Crystals of diffraction quality were obtained by mixing 1 µL of protein solution with 1 µL of reservoir solution and equilibrating the droplet of 2 µL against 700 µL reservoir solution [0 . 4 M magnesium formate , 15% ( w/v ) PEG 3350] . The crystals were soaked in cryo-solution [0 . 4 M magnesium formate , 15% ( w/v ) PEG 3350 , 15% ( v/v ) ethylenglycol or 20% ( v/v ) dimethyl sulfoxide ( DMSO ) ] prior to flash-cooling in liquid nitrogen . The crystal structure of rhIE1 ( 36–395 ) was initially solved in space group P21 using the following diffraction datasets collected at 100 K at beamline BL14 . 1 at BESSY II synchrotron ( Helmholtz Zentrum Berlin ) : a native dataset 1 extending to 2 . 85 Å resolution , a MAD dataset ( peak , inflection point , remote ) recorded from a gold-soaked crystal diffracting to 3 . 5 Å and a peak dataset from a seleno-methionine derivatized protein crystal diffracting to 3 . 1 Å resolution ( Table S1 ) . The gold-soaked crystal was prepared by incubating crystals for three days in cryo-solution containing DMSO and 2 . 5 mM KAu ( CN ) 2 . Before flash cooling , the crystals were back-soaked in cryo-solution for several minutes to remove unspecifically bound heavy atoms . All monoclinic datasets are highly isomorphous . The Matthews coefficient is 2 . 96 ( 58 . 42% solvent ) when assuming the presence of four rhIE1 ( 36–395 ) molecules in the asymmetric unit [61] . The final refinement of rhIE1 ( 36–395 ) was performed against a dataset with space group symmetry P43 extending to 2 . 3 Å ( Table S1 ) . The increase in resolution and concomitant space group change were obtained upon controlled dehydration of the monoclinic crystals from above with the HC1c crystal humidifier device at beamline BL14 . 3 at BESSY II ( Helmholtz Zentrum Berlin ) . Crystals were first equilibrated against 98% relative humidity before decreasing the humidity in steps of 4% and 10 min equilibration time to a final value of 86% humidity . Upon observation of an increase in diffraction power , the crystals were flash cooled and transferred to beamline BL14 . 1 for the recording of a complete dataset extending to 2 . 3 Å resolution ( Table S1 ) . All diffraction datasets were processed with XDS and scaled with XSCALE [62] . Initial protein phases were derived for the monoclinic crystal form using the MAD dataset collected from a gold-soaked crystal ( Table S1 ) . The positions of the gold atoms could be readily located with program SHELXD [63] . The non-crystallographic symmetry ( NCS ) relationship between the 4 monomers , i . e . the presence of two IE1 dimers with C2 point group symmetry , became apparent upon visualization of the gold positions in program COOT and the inspection of the initial electron density maps calculated with program SHELXE [64] , [65] . The NCS relationship was corroborated by the self-rotation function , calculated with program POLARRFN from the CCP4 program suite [66] . The quality of these initial electron density maps could be significantly improved upon phase calculation with program SHARP/autoSHARP [67] , [68] and density averaging with program DM [69] . The improved phases also allowed for the identification of the selenium positions in the peak dataset of the seleno-methionine-labeled protein crystal and the inclusion of this dataset in the calculation of the experimental protein phases . An initial atomic model covering a single monomer was then manually built starting from protein fragments derived with program autoSHARP and following the lead of electron density maps calculated with either program autoSHARP or MLPHARE/DM [66] , [68] . The registration of the protein sequence was obtained from the shape of the local electron density and the positions of the selenium atoms as visualized by an anomalous difference map . These considerations also showed that one gold cation is bound via a free cysteine side chain in each IE1 monomer chain . The model was then stepwise completed , extended to four molecules in the asymmetric unit and refined with program PHENIX [70] . Convergence of the refinement at 2 . 8 Å in space group P21 was facilitated upon inclusion of NCS weights and secondary structure restraints in program PHENIX [70] . The final model of rhIE1 ( 36–395 ) was obtained after transferring the monoclinic model into the tetragonal unit cell with program PHASER and upon refinement against the 2 . 3 Å dataset in space group P43 ( Table S1 ) [71] . Refinement converged at crystallographic R-factors of 19 . 73% ( Rwork ) and 24 . 96% ( Rfree ) . The space group change that took place upon dehydration can be easily explained by small readjustments in the packing of the IE1 dimers in the crystals . Coordinates and structure factors for the IE1CORE structure have been deposited in the Protein Data Bank under accession code 4WID . The size of the dimerization interface between the two rhIE1CORE monomers was calculated with the program PISA [72] . The reported value is the average of the buried surface of both chains . The rhIE1CORE monomers shown in Figure S5 were superposed with the program LSQKAB [66] . Only the α-helical segments of the protein as defined in Figure S1 were superimposed . Comparative modeling of hIE1 was performed with MODELLER 9 . 9 [73] and the resulting model was validated using ProSA [74] , [75] . Searches for structurally similar proteins were performed with PDBeFold [76] . Since standard parameters did not result in any hits , the following search options were set . ( i ) The threshold of 70% for the lowest acceptable match in query and target was reduced to 30% and 60% , respectively . ( ii ) The search was extended to proteins with a different connectivity of their secondary structure elements . Searches were performed independently for chain A and chain B of rhIE1 , and the list of hits was merged . For reasons of clarity , duplicate hits and hits related closely in sequence ( >90% identity ) were removed from the list . The normRMSD was calculated according to the following equation [77]: normRMSD = [RMSD • max ( L1 , L2 ) ]/Naln . Where RMSD is the root mean square deviation of the superposition of query and target , max ( L1 , L2 ) is the number of amino acids of the largest chain in the superposition , and Naln is defined by the number of structurally equivalent residue pairs . Sequence conservation was calculated based on a Blosum30 matrix using the MultiSeq Plugin [78] of VMD [79] . The IE1 sequences from the following viruses served as input: human CMV ( strain AD169 ) , Rhesus-CMV , Baboon-CMV , Simian-CMV , and Panine-HV2/Chimpanzee CMV ( Uniprot-accession-numbers P13202 , Q2FAE9 , D0UZW7 , Q98682 , Q8QRY6 ) . Prediction of intrinsically disordered regions in hIE1 , cIE1 and rhIE1 was performed with IUPred [39] using the prediction type “short disorder” . The disorder tendencies in the three IE1 homologs were plotted in one diagram using the rhIE1 amino acid numbering . Multiple sequence alignment was performed with TCoffee ( http://www . ebi . ac . uk/Tools/services/web/toolform . ebi ? tool=tcoffee ) . The oligonucleotide primers used for this study were purchased from Biomers GmbH ( Ulm , Germany ) and are listed in Table S3 . All prokaryotic expression plasmids were generated by PCR amplification of the respective codon-optimized IE1 sequences and subsequent cloning into pGEX-6P-1 ( GE Healthcare Bio-Sciences AB , Uppsala , Sweden ) . The synthetic , codon-optimized hIE1 cDNA ( strain AD169 ) was obtained from Mr . Gene GmbH ( Regensburg , Germany ) . The codon-optimized cDNAs of cIE1 ( NP_612746 . 1 ) and rhIE1 ( Q2FAE9 ) were synthesized by GENEART gene synthesis service ( Regensburg , Germany ) . The eukaryotic expression plasmids encoding full length or truncated hIE1 were generated via PCR amplification of the respective fragments using pHM494 [80] as template , followed by insertion into pHM971 ( pcDNA3 . 1-FLAG ) [80] , pHM1580 ( pcDNA3 . 1-Myc ) [80] , or into the yeast expression vectors pGBT9 and pGAD424 ( Clontech , Mountain View , CA ) . The synthetic gene coding for rhIE1 ( Q2FAE9 ) was obtained from GENEART gene synthesis service ( Regensburg , Germany ) . The rhIE1 coding sequence was subcloned into pHM971 ( pcDNA3 . 1-FLAG ) [80] using BamHI and XhoI . Full length PML , isoform VI , and truncated PML variants were amplified from pAS-PML ( a gift from G . G . Maul , Philadelphia , USA ) and inserted into pHM1580 ( pcDNA3 . 1-Myc ) [81] , pHM971 ( pcDNA3 . 1-FLAG ) [80] , pHM972 ( pcDNA3 . 1-FLAG-NLS ) , or yeast expression vectors pGBT9 and pGAD424 ( Clontech , Mountain View , CA ) . The eukaryotic expression plasmid encoding rhesus TRIM5α was a gift from T . Gramberg ( Erlangen , Germany ) . For transduction experiments , hIE1 variants were amplified utilizing pHM494 [80] as template and inserted into a pLKO-based lentiviral vector ( a gift of R . Everett , Glasgow , UK ) . HEK293T cells and primary human foreskin fibroblast ( HFF ) cells ( obtained from Life Technologies ) or telomerase-immortalized HFFs ( HFFi ) were cultured as described previously [54] , [82] . HFFs were infected with either the HCMV laboratory strain AD169 , a recombinant HCMV expressing hIE1 1–382 ( AD169/hIE1 1–382 ) , an IE1-deficient virus ( AD169ΔhIE1 ) , or rhesus macaque CMV ( RhCMV ) at specified multiplicities of infection ( MOI ) . Titers of wild-type ( wt ) AD169 and the AD169 recombinants were determined by UL112/113 fluorescence . For this purpose , HFFs were infected with various dilutions of virus stocks . After 72 h of incubation , cells were fixed and stained with a monoclonal antibody directed against UL112/113 . Subsequently , the number of UL112/113-positive cells was determined and was used to calculate viral titers . RhCMV was titrated via rhIE1 fluorescence , which was analyzed 24 h postinfection . The AD169-based HCMV bacterial artificial chromosome ( BAC ) HB15 was used for recombination-based genetic engineering of AD169/hIE1 1–382 and AD169ΔhIE1 . AD169/hIE1 1–382 was constructed by introducing a stop codon into the hIE1 gene replacing residue 383 . For this purpose , the two-step red-mediated recombination technique was utilized [83] , which uses the kanamycin gene as a first selection marker . The linear recombination fragment was generated by PCR using primers 5′BAC_short and 3'BAC_hIE1_382 ( Table S3 ) , and pEPkan-S ( kindly provided by K . Osterrieder , Berlin ) as template DNA . The PCR product was treated with DpnI , gel purified , and subjected to a second round of PCR amplification using primers 5′BAC_hIE1_382 and 3′BAC_hIE1_382_short ( Table S3 ) . For homologous recombination , the PCR fragment was transformed into Escherichia coli strain GS1783 ( a gift of M . Mach , Erlangen ) already harboring HB15 , and bacteriophage λ red-mediated recombination was conducted as described elsewhere [83] . To identify positive transformants , the bacteria were plated on agar plates containing 30 µg/mL kanamycin ( first recombination ) or 30 µg/mL chloramphenicol and 1% arabinose ( second recombination ) and incubated at 32°C for 2 days . BAC DNA was purified from bacterial colonies growing on these plates and was further analyzed by PCR , restriction enzyme digestion and direct sequencing . For construction of the AD169ΔhIE1 BAC by homologous recombination , a linear recombination fragment , comprising a kanamycin resistance marker along with 5′ and 3′ genomic sequences , was generated by PCR amplification using pKD13 as template and primers 5′Intron3/pKD13 and 3′Exon 4/pkd13 ( Table S3 ) . This fragment was used for electroporation of competent Escherichia coli strain DH10B harboring HB15 and recombination was performed as described previously in order to delete exon 4 of the IE1 gene [84] . The integrity of the resulting recombinant BAC was confirmed by PCR , restriction enzyme digestion and direct sequencing . For reconstitution of recombinant AD169 , HFFs seeded in six-well dishes ( 3×105 cells/well ) were cotransfected with 1 µg of purified BAC DNA , 0 . 5 µg of the pp71 expression plasmid pCB6-pp71 , and 0 . 5 µg of a vector encoding the Cre recombinase using FuGENE6 transfection reagent ( Promega , Mannheim , Germany ) . Transfected HFFs were propagated until viral plaques appeared , and the supernatants from these cultures were used for further virus propagation . For the generation of HFF cells stably expressing full length hIE1 or hIE1 1–382 , replication-deficient lentiviruses were generated using pLKO-based expression vectors . For this purpose , HEK293T cells seeded in 10 cm dishes ( 5×106 cells/dish ) were cotransfected with a pLKO vector encoding either full length hIE1 or hIE1 1–382 together with packaging plasmids pLP1 , pLP2 , and pLP/VSV-G using the Lipofectamine 2000 reagent ( Invitrogen , Karlsruhe , Germany ) . Viral supernatants were harvested 48 h after transfection , cleared by centrifugation , filtered , and stored at −80°C . Primary HFFs or telomerase-immortalized HFFs were incubated for 24 h with lentivirus supernatants in the presence of 7 . 5 µg/mL polybrene ( Sigma-Aldrich , Deisenhofen , Germany ) . Stably transduced cell populations were selected by adding 500 µg/mL geneticin to the cell culture medium . HFF cells were transfected with the DNA transfection reagent FuGENE6 ( Promega , Mannheim , Germany ) . One day before transfection , 3×105 cells were seeded into six-well dishes . DNA content and transfection procedure were according to the instructions of the manufacturer . 48 hours after transfection , cells were harvested for further analyses . HEK293T cells were transfected by applying the standard calcium phosphate precipitation method . For this , 5×105 to 5×106 HEK293T cells were seeded into six-well dishes or 10 cm dishes one day before transfection . For Western blot analyses and coimmunoprecipitations , 1 to 10 µg of plasmid DNA were used for each transfection reaction . At about 16 hours later , the cells were washed two times with PBSo and provided with fresh medium . 48 hours after transfection , cells were harvested for further analyses . Monoclonal antibodies used for immunofluorescence and Western blot analyses were: α-IE1 CH443 ( Santa Cruz Biotechnology , Santa Cruz , CA , USA ) , α-UL112/113 M23 , α-UL44 BS510 ( kindly provided by B . Plachter , Mainz , Germany ) , α-UL69 69–66 , α-FLAG M2 ( Sigma-Aldrich , Deisenhofen , Germany ) , α-Myc 9E10 , α-β-actin AC-15 ( Sigma-Aldrich ) , α-PML PG-M3 ( Santa Cruz ) . Polyclonal antibodies used for immunofluorescence and Western blot analyses were: α-rhesus IE1 ( a kind gift from M . Mach , Erlangen , Germany ) , α-PML #4 ( a kind gift from P . Hemmerich , Jena , Germany ) , α-PML H238 ( Santa Cruz ) , α-PML A301–167A ( Bethyl Laboratories , Montgomery , TX , USA ) , α-PML A301–168A ( Bethyl Laboratories ) , α-Sp100 #2 ( a kind gift from P . Hemmerich , Jena , Germany ) , α-Sp100 GH3 ( kindly provided by H . Will , Hamburg , Germany ) , α-hDaxx C-20 ( Santa Cruz ) , α-ATRX H-300 ( Santa Cruz ) . Secondary antibodies used for immunofluorescence and Western blot analyses were: Alexa Fluor 488-/555-/647-conjugated secondary antibodies for indirect immunofluorescence experiments were purchased from Molecular Probes ( Karlsruhe , Germany ) , horseradish peroxidase-conjugated anti-mouse/-rabbit secondary antibodies for Western blot analyses were obtained from Dianova ( Hamburg , Germany ) . HFF cells grown on coverslips in six-well dishes ( 3×105 cells/well ) were washed twice with PBSo at 48 hours after transfection or at various times after virus infection . Cells were fixed with a 4% paraformaldehyde solution for 10 min at room temperature ( RT ) and then washed for two times . Permeabilization of the cells was achieved by incubation with 0 . 2% Triton X-100 in PBSo on ice for 20 min . Cells were washed again with PBSo over a time period of 5 min and incubated with the appropriate primary antibody diluted in PBSo-1% FCS for 30 min at 37°C . Excessive antibodies were removed by washing four times with PBSo , followed by incubation with the corresponding fluorescence-coupled secondary antibody diluted in PBSo-1% FCS for 30 min at 37°C . The cells were mounted using the DAPI-containing Vectashield mounting medium ( Alexis , Grünberg , Germany ) and analyzed using a Leica TCS SP5 confocal microscope , with 488 nm , 543 nm , and 633 nm laser lines , scanning each channel separately under image capture conditions that eliminated channel overlap . The images were exported , processed with Adobe Photoshop CS5 and assembled using CorelDraw ×5 . In order to quantify PML-NB disruption in infected HFFs , 150 cells were analyzed for the presence of PML dots . Lysates from transfected or infected cells were prepared in a sodium dodecyl sulfate-polyacrylamide gel electrophoresis ( SDS-PAGE ) loading buffer , separated on sodium dodecyl sulfate-containing 8 to 15% polyacrylamide gels , and transferred to nitrocellulose membranes . Chemiluminescence was detected according to the manufacturer's protocol ( ECL Western blot detection kit; Amersham Pharmacia Biotech ) . Transfected HEK293T cells ( 1×106 or 5×106 were lysed for 20 to 40 min at 4°C in 800 µL of CoIP buffer ( 50 mM Tris-HCl [pH 8 . 0] , 150 mM NaCl , 5 mM EDTA , 0 . 5% NP-40 , 1 mM PMSF , 2 µg/mL of aprotinin , 2 µg/mL of leupeptin , and 2 µg/mL of pepstatin ) . After centrifugation , aliquots of each sample were taken as input controls and the remaining supernatant was incubated with anti-FLAG antibody M2 coupled to protein-A-sepharose beads for 2 h at 4°C . The sepharose beads were collected by centrifugation and washed five times in 1 mL CoIP buffer . Finally , the immunoprecipitated proteins were recovered by boiling in 4× SDS sample buffer and protein complexes were analyzed by SDS-PAGE and Western blotting . Saccharomyces cerevisiae Y153 was used in a two-hybrid system . Both the plasmid pGBT9 ( Clontech , Mountain View , CA ) encoding the GAL4-DB ( Trp+ ) fusion and the plasmid pGAD424 ( Clontech , Mountain View , CA ) encoding the GAL4-A fusion ( Leu+ ) were introduced into Y153 cells using a modified lithium acetate ( LiAc ) method . For this , cells were grown overnight in YAPD medium , pelleted and treated with LP-mix ( 40% w/v PEG 4000 , 0 . 15 M LiAc , 10 mM Tris/HCL pH 7 . 5 , 1 mM EDTA pH 8 . 0 ) and DMSO . Single-stranded carrier-DNA as well as both plasmids were added to the yeast cells . This step was followed by incubation at room temperature and subsequent incubation at 42°C . Thereafter , the cells were plated on minimal selection agar lacking Trp and Leu . For rapid in situ assays of lacZ expression from yeast colonies , an XGal filter assay was used . Nitrocellulose filters were laid onto the plate and allowed to wet completely , then lifted off the plate and placed in liquid nitrogen to permeabilize the cells . The filters were removed and placed cell side up in a petri dish containing Whatman Paper soaked with Z buffer containing β-Mercaptoethanol and XGal . The filters were incubated at 30°C and constantly analyzed for the development of a positive blue color . For quantitation of the ß-galactosidase activity in the yeast cells three colonies were picked and grown in medium also lacking Trp and Leu . The next day , the optical density was measured at 600 nm . After pelleting the culture , the cells were resuspended in Z buffer and permeabilized using chloroform and 0 . 1% SDS . The ß-galactosidase activity within the cells was assayed by the standard method using o-nitrophenyl-ß-D-galactopyranoside ( ONPG ) as substrate . The reaction was stopped by adding Na2CO3 and the absorbance was measured at OD405 . The unit of ß-galactosidase was defined as ( 1 . 000×OD405 ) / ( t×v×OD600 ) ( t , reaction time [min]; v , culture volume [mL] ) . The ß-galactosidase activity for each sample was corrected for background by subtracting the signal of the empty vectors . HFF cells were seeded into six-well dishes at a density of 3×105 cells/well and infected the following day with wt AD169 at an MOI of 0 . 01 and equivalent genome copies of AD169/hIE1 1–382 and AD169ΔhIE1 . Triplicate samples of the infected cell supernatants were harvested at 2 , 4 , 6 , 8 and 11 days after inoculation and subjected to lysis by proteinase K treatment . Thereafter , all samples were analyzed for the amount of genome copy numbers by quantitative real-time PCR ( TaqMan-PCR ) using an Applied Biosystems 7500 Real-Time PCR System ( Applied Biosystems , Foster City , CA , USA ) together with the corresponding software SDS ( sequence detection system ) [85] . For quantification of the viral DNA load , a sequence region within the gB gene locus was amplified using primers 5'gB_forw and 3'gB_rev along with the fluorescence labeled probe CMV gB FAM/TAMRA also directed against the gB gene region . In parallel , the cellular DNA amount was quantified using primers 5′Alb and 3′ Alb together with a fluorescence labeled probe , Alb FAM/TAMRA , specific for the cellular albumin gene . Real-time PCR was performed in 96-well plates being compatible with the ABI Prism sequence detector . For the determination of reference CT values ( cycle threshold ) , serial dilutions of the respective standards ( 107–101 DNA molecules of gB or albumin ) were examined by PCR reactions in parallel . The 20 µL reaction mix contained 5 µL sample or standard DNA solution together with 10 µL 2× TaqMan PCR Mastermix ( Applied Biosystems , Foster City , CA , USA ) , 1 . 5 µL of each primer ( 5 µM stock solution ) , 0 . 4 µL of probe ( 10 µM stock solution ) , and 1 . 6 µL of H2O . The thermal cycling conditions consisted of two initial steps of 2 min at 50°C and 10 min at 95°C followed by 40 amplification cycles ( 15 sec at 95°C , 1 min 60°C ) . The viral genome copy numbers and albumin copy numbers were subsequently calculated using the sample-specific CT value when set into relation to the standard serial dilutions . For analysis of complementation by immunofluorescence staining , control HFFi cells as well as HFFi cells expressing hIE1 or hIE1 1–382 were infected with AD169ΔhIE1 at an MOI of 0 . 01 . Triplicate samples of infected cells were fixed at 48 h post infection and subjected to immunostaining of UL44 . Images of approximately 500 cells per sample were taken and the number of UL44-positive cells was determined via measuring of mean gray values using the ImageJ software ( version 1 . 47 ) . For analysis of complementation by Western blotting , normal HFFi cells as well as HFFi cells expressing hIE1 or hIE1 1–382 were infected with AD169ΔhIE1 at an MOI of 0 . 05 in triplicate , and were harvested 72 h later for detection of UL44 , UL69 , IE1 and β-actin .
|
Research of the last few years has revealed that microbial infections are not only controlled by innate and adaptive immune mechanisms , but also by cellular restriction factors , which give cells the capacity to resist pathogens . PML nuclear bodies ( PML-NBs ) are dot-like nuclear structures representing multiprotein complexes that consist of the PML protein , a member of the TRIM family of proteins , as well as a multitude of additional regulatory factors . PML-NB components act as a barrier against many viral infections; however , viral antagonistic proteins have evolved to modify PML-NBs , thus abrogating this cellular defense . Here , we delineate the molecular basis for antagonization by the immediate-early protein IE1 of the herpesvirus human cytomegalovirus . We present the first crystal structure for the evolutionary conserved core domain ( IE1CORE ) of primate cytomegalovirus IE1 , which exhibits a novel , unusual fold . IE1CORE modifies PML-NBs by releasing other PML-NB proteins into the nucleoplasm which is sufficient to antagonize intrinsic immunity . Importantly , IE1CORE shares secondary structure features with the coiled-coil domain ( CC ) of TRIM factors , and we demonstrate strong binding of IE1 to the PML-CC . We propose that IE1CORE sequesters PML and possibly other TRIM family members via an extended binding surface formed by the coiled-coil domain .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"immediate-early",
"promoter",
"medicine",
"and",
"health",
"sciences",
"immunology",
"microbiology",
"infectious",
"disease",
"immunology",
"protein",
"structure",
"nuclear",
"bodies",
"cellular",
"structures",
"and",
"organelles",
"immune",
"system",
"proteins",
"protein",
"structure",
"determination",
"viral",
"immune",
"evasion",
"proteins",
"viral",
"replication",
"molecular",
"biology",
"viral",
"genes",
"biochemistry",
"immune",
"evasion",
"cell",
"biology",
"clinical",
"immunology",
"virology",
"biology",
"and",
"life",
"sciences",
"molecular",
"cell",
"biology",
"macromolecular",
"structure",
"analysis"
] |
2014
|
Crystal Structure of Cytomegalovirus IE1 Protein Reveals Targeting of TRIM Family Member PML via Coiled-Coil Interactions
|
Plasmodium undergoes one round of multiplication in the liver prior to invading erythrocytes and initiating the symptomatic blood phase of the malaria infection . Productive hepatocyte infection by sporozoites leads to the generation of thousands of merozoites capable of erythrocyte invasion . Merozoites are released from infected hepatocytes as merosomes , packets of hundreds of parasites surrounded by host cell membrane . Intravital microscopy of green fluorescent protein–expressing P . yoelii parasites showed that the majority of merosomes exit the liver intact , adapt a relatively uniform size of 12–18 μm , and contain 100–200 merozoites . Merosomes survived the subsequent passage through the right heart undamaged and accumulated in the lungs . Merosomes were absent from blood harvested from the left ventricle and from tail vein blood , indicating that the lungs effectively cleared the blood from all large parasite aggregates . Accordingly , merosomes were not detectable in major organs such as brain , kidney , and spleen . The failure of annexin V to label merosomes collected from hepatic effluent indicates that phosphatidylserine is not exposed on the surface of the merosome membrane suggesting the infected hepatocyte did not undergo apoptosis prior to merosome release . Merosomal merozoites continued to express green fluorescent protein and did not incorporate propidium iodide or YO-PRO-1 indicating parasite viability and an intact merosome membrane . Evidence of merosomal merozoite infectivity was provided by hepatic effluent containing merosomes being significantly more infective than blood with an identical low-level parasitemia . Ex vivo analysis showed that merosomes eventually disintegrate inside pulmonary capillaries , thus liberating merozoites into the bloodstream . We conclude that merosome packaging protects hepatic merozoites from phagocytic attack by sinusoidal Kupffer cells , and that release into the lung microvasculature enhances the chance of successful erythrocyte invasion . We believe this previously unknown part of the plasmodial life cycle ensures an effective transition from the liver to the blood phase of the malaria infection .
Two billion people , more than one third of the world's population , live at risk for malaria and about 1 billion are infected . Each year there are 300 million to 500 million new cases with 2–3 million deaths , the vast majority young children in Africa . We are now forty years past the discovery that radiation-attenuated sporozoites protect against malaria [1] , but we still lack an efficient malaria vaccine to combat this deadly parasitic disease , and drug resistance is wide-spread [2] . The malaria infection begins with the introduction of sporozoites from the bite of an infected Anopheles mosquito [3 , 4] . The sporozoites travel to the liver and develop in hepatocytes to large exoerythrocytic forms ( EEFs ) [5 , 6] . Schizogonic division of the EEF then results in the formation of thousands of first-generation merozoites , which are responsible for the initiation of clinical malaria . Merozoites have a short life span and must infect erythrocytes immediately after release into the bloodstream [7] . Merozoites are also highly susceptible to phagocytosis and must therefore avoid contact with macrophages [8] . Acute danger of phagocytic elimination is presented in the form of Kupffer cells [8] , the resident phagocytes of the liver that comprise by far the largest population of tissue macrophages of the body [9] . Kupffer cells are predominantly located at sinusoidal bifurcations , largely within and often spanning the sinusoidal lumen [9–11] , thereby presenting significant obstacles for non-self particulate material . This strategic position of Kupffer cells makes it difficult for free merozoites to exit the liver without being trapped by these surveillance cells of the innate immune system . The first evidence suggesting that merozoites can be released from hepatocytes as clusters , held together by host cell cytoplasm , was presented several decades ago in Garnham's ultrastructural examination of Plasmodium yoelii–infected murine livers and described in more detail in Meis' extensive electron microscopic studies on P . berghei infection of the mouse [5 , 12 , 13] . More recently , we and others reported that merozoites are released as “extrusomes” or “merosomes” that contain hundreds to thousands of parasites [14 , 15] ( reviewed in [16 , 17] ) . Our initial intravital observations using green fluorescent P . yoelii and BALB/c mice revealed extensive movement within EEFs nearing completion of merozoite maturation culminating in budding and release of merosomes into the hepatic bloodstream [14] . An elegant series of in vitro studies described the differentiation of P . berghei merozoites in the human hepatoma cell line HepG2 [15] . While developing into hepatic schizonts , the intracellular parasites prevent the initiation of a death program in their host cells , but leave them to die once merozoite formation is complete . Underlying molecular details remain to be determined , but the data suggest that host cell death in this in vitro model shares more features with autophagy than apoptosis or necrosis [18] . However , information on the viability of hepatocytes releasing merozoites into the sinusoidal blood is lacking to date . Because P . yoelii infection of the mouse represents an accepted model closely reflecting human malaria [19] , we used a variety of microscopic techniques to study the dynamics of merosome budding from infected hepatocytes and the fate of hepatic merozoites in the body . Confocal images provided measurements of merosome volume , merosomal merozoite content , and EEF volume , and appropriate mathematic processing of these data allowed us to calculate the number of hepatic merozoites produced by P . yoelii sporozoites in the murine host . Using intravital and ex vivo microscopy , we found that the vast majority of hepatic P . yoelii merozoites leave the liver camouflaged as merosomes , disseminate within the cardiovascular system , and arrest in the lungs . Molecular markers revealed that merosomal merozoites remain viable and infectious until being released into the pulmonary microcirculation . In contrast , various in vivo and ex vivo assays suggest that unreleased merozoites and the exhausted host cell eventually succumb to necrosis . The resulting inflammatory stimulus attracts neutrophils , and mononuclear phagocytes thus give rise to the formation of microgranulomata . Overall , this systematic temporal and quantitative analysis indicates that merosome formation and release by host hepatocytes , merosome transport to and sequestration in the lungs , and release of merozoites into the pulmonary microvasculature are parts of a previously unrecognized phase of the Plasmodium life cycle .
P . yoelii–infected mice have been suggested to represent a suitable model for human malaria [20] . We also consider P . yoelii an appropriate rodent model for liver stage analysis because it induces less inflammation in murine livers than P . berghei and produces more EEFs [21] , which in addition are generally larger and contain more merozoites [12 , 22 , 23] ( Table S1 ) . While available for other species such as P . berghei , information is scarce regarding ultrastructural changes during P . yoelii EEF maturation in the liver and the subsequent release of first generation merozoites [12 , 24] . To help fill this gap and to expand our previous investigation of Plasmodium merosomes in live mice [14] , we used several light and electron microscopy techniques to examine this process . Mature Plasmodium EEFs contained thousands of merozoites enclosed in a parasitophorous vacuole ( PV ) . Up to the final developmental stage and onset of merozoite release , infected hepatocytes remained in close contact with neighboring uninfected parenchymal ( Figure 1A and 1B ) and sinusoidal cells ( Figure 1C ) . Shortly before merosome formation , the PV membrane ( PVM ) disintegrated so that host cytoplasm contained a mixture of mature merozoites , morphologically intact hepatocyte organelles ( Figure 1D ) , parasite remnant bodies ( or pseudocytomeres [5] ) , and parasite stroma left over from schizogonic merozoite formation ( Figure 1E ) . Some of the sinusoids adjacent to infected hepatocytes remained filled with erythrocytes indicating preservation of function , but others were compressed by the expanding parasite and lacked erythrocytes suggesting local obstruction of blood flow ( Figure 1C ) . To calculate the merozoite content of mature EEFs ( see below ) , we needed accurate measurements of the EEF size . Compared to tissue sections , intravital microscopy of green fluorescent protein ( GFP ) Plasmodium yoelii parasite ( PyGFP ) –infected mouse livers ( Figure 1F ) offered the advantage of examining live tissue within an intact animal , thus avoiding artifacts associated with both fresh and fixed sections . Mature EEFs within the liver typically have a slightly ellipsoid shape with the minimum and maximum diameters ranging from 40 to 75 μm ( with averages of 49 . 2 ± 10 . 3 μm to 55 . 6 ± 9 . 0 μm ) , respectively ( n = 16 ) . Detailed intravital examination of 30 mice at times ranging from 30 to 74 h after intravenous infection with PyGFP sporozoites allowed us to follow the complex series of events involved in merozoite liberation from hepatocytes . We monitored more than 60 EEFs over this period and observed the earliest merosome budding at 46 h ( Figure 2A , Videos S1–S3 ) , a time in general agreement with earlier work reporting the first appearance of P . yoelii in the blood at 45 . 5 h [12] . Of these 60 EEFs , 20 reached maturity during the observation period and released merozoites , while the rest remained immature . The majority ( 13 ) of these 20 EEFs released merozoites by merosome formation . Merosome formation continued until 56 h after infection , thus confirming the asynchronous nature of P . yoelii EEF maturation , a common observation in Plasmodium-infected livers [5 , 25] . Because we infected by intravenous sporozoite injection , the well-known slow release of sporozoites from the mosquito bite site [3 , 26] alone cannot account for the asynchronicity observed here . For individual EEFs , the process of merosome budding and release lasted several hours during which time the host cell gradually decreased in size and separated from neighboring cells ( Figure 2B ) . In addition to fully formed green fluorescent merozoites , released merosomes contained non-fluorescent remnant bodies and host cell organelles , thus providing further evidence that merosome budding occurs after rupture of the PVM . Eventually , the host cell membrane appeared to lose its integrity and allowed some leftover merozoites to enter the bloodstream singly and without protection by a merosomal membrane ( Figure 2C and 2D , Video S4 ) . GFP radiated out from the disintegrating EEF into the surrounding tissue , implying that parasite antigens and host cell cytoplasm were set free as well . Indeed , electron microscopic examination showed free mitochondria in the sinusoidal lumen ( Figure 2E ) . Size and shape of these organelles revealed hepatocyte origin . Eventually , inflammatory cells were attracted to the site of the disintegrating EEF . During phagocytic removal of debris from dead merozoites and host cells , neutrophil granulocytes and mononuclear phagocytes transformed the site of the former EEF into a small granuloma ( Figure S1 ) , a structure commonly reported at late stages of Plasmodium liver infection [5 , 8 , 27–31] . Thus , merosome formation in the liver occurs over a period of about 10 h and is followed by disintegration of the host cell and some leftover parasites , clearance of the remains by infiltrating phagocytes , and production of a small granuloma . Merozoites were also liberated by a less frequent mechanism . Starting earlier than merosome formation ( 42 h post inoculation ) , some infected hepatocytes rapidly discharged their content of merozoites and cell organelles by a mechanism appearing to involve rupture of the cell membrane ( Figure 3A–3E , Video S5 ) . In some cases , the process was complete in as little as 5 min; in others it lasted as long as 60 min . Of the EEFs rupturing in this manner , 80% harbored mature merozoites , but 20% had a homogeneous cytoplasm; thus , schizogony had not even begun ( Figure S2 ) . Occasionally , electron micrographs showed immature merozoites incompletely separated from remnant bodies yet released into the sinusoidal bloodstream ( Figure 3F ) . This apparent rupture-release left large faintly fluorescent EEF ghosts at the site of the former host cell . Because our intravital observations were based on confocal microscopy , we considered the possibility of phototoxicity playing a role in this rupture-release mechanism . However , since EEF ghosts identical to those resulting from observed rupture were detectable at the very beginning of intravital examination , we could reject that possibility . Because the EEFs did not decrease in volume prior to transformation into a ghost , and we did not find erythrocytes associated with these ghosts , we suspect that the remains of the host cell cytoskeleton , the surrounding extracellular matrix , and/or the sinusoidal cell layer resealed the ghost after merozoite release; thus , preventing the formation of hemorrhages . Similar to the end of the merosome release mechanism ( see above ) , EEF ghosts were infiltrated by inflammatory cells that gave rise to small granulomata . When we combine results from intravital microscopy , showing that both mature and immature EEFs undergo this rapid decay , with our electron microscopy data , showing that some of the rupturing EEFs were immature , we conclude that this rapid release process is a result of abortive EEF development that , in the absence of host cell membrane protection , exposes the parasites to Kupffer cell phagocytosis . To demonstrate that merozoites within merosomes are alive and to help exclude the possibility that merosome release represents an abnormal development , we injected infected mice with markers that reveal cell viability in vivo . At points ranging from 51 to 74 h post inoculation , mice were injected with a mix of the membrane-permeable DNA stain Hoechst 33342 and the dead cell marker propidium iodide ( PI ) . Subsequent intravital confocal microscopy revealed that PI does not enter merosomes or intact EEFs ( Figure 4A and 4B ) , but does stain some of the merozoites left behind in EEF ghosts and also in EEFs that had disintegrated after merosome budding ( Figure 4C and 4D ) . These findings support the interpretations above in that they suggest that merozoites that fail to exit the host cell eventually succumb to necrosis . Efforts to determine the mode by which merosomes breach the sinusoidal cell layer failed so far due to insufficient numbers of suitable events for analysis . We suspect that budding occurs through the endothelial fenestration rather by a paracellular route , because of the extreme natural variability of the diameter of the fenestrae in response to changes in blood pressure and other physiologic stimuli . Interestingly , mature EEFs were frequently surrounded by a layer of flattened cells that had incorporated the dead cell stain PI ( Figure 4A and 4B ) . Perhaps the death of these cells is due to extreme compression by the extensive expansion of the EEF during the final stage of development . Occasionally , merosomes were found budding into such dead cells , but the immobility of the parasites indicated that they were trapped ( Figure 4A ) . When hepatic merosomes initially bud from infected hepatocytes ( Figure 5 ) , they are highly variable in size and contain hundreds to thousands of mature merozoites , while merosomes in blood draining from the liver were smaller and more uniform in size . Intravital microscopy showed very large merosomes moving far more slowly than small ones , which leave the liver lobules at a velocity close to that of blood cells ( unpublished data ) . We frequently observed merosomes hindering the free flow of the blood as they moved along a sinusoid ( Figure 5C–5G ) as well as being hindered by the vascular architecture . The speed of merosome transport at any instant depended on the diameter and local structure of the sinusoid as well as the size of the merosome . We recorded large merosomes being arrested at sinusoidal bifurcations where they occasionally even reversed direction of movement ( Figure 5A and 5B and Videos S6 and S7 ) . Because morphological measurements taken in vitro are subject to artifact and do not reveal in vivo dynamics , we sought a better understanding of sinusoidal architecture using intravital analysis of uninfected transgenic Tie2-GFP mice that have fluorescent vascular endothelia [32] . We found sinusoidal diameters to range from 3 . 4 μm to 14 . 1 μm ( 6 . 7 ± 1 . 9 μm; n = 94 ) under normal blood pressure conditions . Although large merosomes greatly exceed this size range , their considerable deformability allowed them to gradually wind their way towards the central vein and exit the liver without rupture and release of merozoites , a process aided by resizing ( Figure 5A and 5B ) . We occasionally observed large merosomes subdividing into smaller ones while traveling through sinusoids ( Videos S8 and S9 ) , but we suspect that shear forces associated with the faster blood velocity in larger vessels caused merosomes in the hepatic effluent and inside lung capillaries to be generally smaller and uniform in size compared to those in the liver . The importance of mechanical forces for resizing is demonstrated by another set of experiments in which PyGFP-infected mouse livers were removed from the animals and analyzed ex vivo by confocal microscopy , i . e . , in the absence of blood flow . The sinusoids of such livers contained merosomes of an unusually large size ( Figure S3A ) . When livers were perfused with medium prior to ex vivo confocal microscopy , the sinusoids contained even larger merosomes ( Figure S3B ) . We contend that lack of blood flow prevents subdivision of large merosomes into smaller ones and that liver perfusion hastened merosome budding and liberation from the host cell . Merosome formation results in packaging a mixture of parasites , remnant bodies , and host cell cytoplasm within host cell membrane for release into the sinusoidal lumen ( Figures 5H and 6A ) . Ultrastructurally , the merosomal matrix contained well-preserved merozoites and morphologically intact host cell mitochondria ( Figure 6B ) suggesting that these organelles are viable at the time of merosome budding . Merosomes also typically contained remnant bodies ( Figure 6A ) suggesting that these leftovers from EEF schizogony represent a natural component of the merosomal cytoplasm . In the absence of better viability markers , we interpret the presence of MSP-1 on the surface of merozoites in both mature EEFs and merosomes ( Figure 6D and 6E ) to indicate intactness and complete differentiation of the parasites , and propose that merosomes are linked to productive infection of erythrocytes . Disintegration of the PVM prior to merosome formation indicates the merosome membrane is derived from hepatocyte cell membrane . Asialoglycoprotein receptor 1 ( ASGR1 ) , a protein expressed only on parenchymal liver cells [33–37] , was detectable by immunofluorescence lining the basal hepatocyte surface within the space of Disse ( Figure 6C ) . ASGR1 clearly surrounded mature EEFs ( Figure 6D ) , but it was absent from the merosomal membrane ( Figure 6E ) . The lack of this hepatocyte surface protein could be due to dedifferentiation of the infected host cell or modification by the intracellular parasite at late stages of EEF maturation . However , because the ASGR1 label was located predominantly in the space of Disse rather than on the hepatocyte membrane , more work is needed to define the composition of the merosome membrane . Because intravital observations showed that merosomes remain intact during transport towards the central vein , we examined the hepatic venous effluent for membrane-enveloped parasites . To do this , we opened the inferior Vena cava at its point of entry into the diaphragm and collected blood from the peritoneal cavity . Thick smears were prepared from 5 μl blood , and the concentrations of venous merosomes from three separate experiments were measured . In hepatic venous blood collected 52 h post infection with 2 . 5 × 106 sporozoites , we found 28 . 7 ± 4 . 3 merosomes per μl , and 69% of these merosomes contained between 100 and 200 merozoites ( unpublished data ) . While much information is available on P . berghei–infected HepG2 cells [15] , in vivo data on the molecular composition of the merosome membrane , for example phosphatidylserine ( PS ) exposure , are lacking to date . To obtain more detail on merosome structure , another set of experiments was performed in which the parasite material available for examination was enhanced by liver perfusion . Beginning 52 h after infection with PyGFP or wild-type ( wt ) P . yoelii , livers were perfused with culture medium and the perfusate collected . Cells were immobilized by attachment to Alcian blue–treated glass-bottom dishes and immediately examined by confocal microscopy using conditions that maintain viability . Perfusate merosomes typically adapted a spherical shape in vitro ( Figure 7A ) and 3-D images from confocal stacks demonstrated a relatively uniform size containing several hundred merozoites . Labeling by the phospholipid marker FM 4–64 FX verified that the parasites were held together by a membrane ( Figure 7B ) . Immediately after harvesting , the majority of the parasites appeared viable and merosome membranes were negative for annexin V labeling ( Figure 7C ) ; thus they do not display PS that targets cells for phagocytosis . However , with increased time in vitro , the presence of PS gradually became apparent ( Figure 7D ) . Merozoites in freshly isolated merosomes did not stain with the dead cell marker PI , but those that became positive for PS also lost the ability to exclude PI ( Figure 7D ) . A further viability assessment utilized YO-PRO-1 , a DNA stain that selectively passes through the ( intact ) plasma membrane of apoptotic cells . Again , merozoites in freshly isolated merosomes did not label with YO-PRO-1 , but as time in vitro progressed , they began to incorporate YO-PRO-1 along with PI ( Figure 7E and 7F ) . Within roughly 60 min of in vitro examination , all merosomes were positive for annexin V , YO-PRO-1 , and PI . Attempts to quantify the time course of these processes more precisely were prevented by the sensitivity of the merosomes to the various steps of isolation from the liver and concentration by centrifugation . Taken together , these results suggest that P . yoelii merosomes leaving the liver contain viable merozoites , and , similar to P . berghei–infected HepG2 cells [15] , lack PS as a membrane marker that signals “eat-me” to phagocytes . Considering that Kupffer cells are located largely within and often spanning the sinusoidal lumen , thus presenting a significant obstacle for non-self particulate material and damaged host cells [9–11] , the lack of PS on the merosome membrane is likely critical for merozoite escape from this defense mechanism of the host . Because the entire hepatic effluent must pass through the right ventricle and the pulmonary microcirculation before reaching any other capillary bed , we suspected merosomes might sequester in the lungs . To address this , we used ex vivo confocal microscopy to examine the alveolar microvasculature immediately after lung removal while the tissues were intact and the cells alive . At time points from 46 to 58 h after inoculation with PyGFP sporozoites , we found numerous intact merosomes as well as individual parasites ( Figure 8A–8C ) . We did not find pulmonary merosomes earlier than 46 h post inoculation nor later than 65 h , timing consistent with our observation that merosome release begins and ends at roughly 46 h and 56 h . MSP-1 labeling confirmed the maturity of merozoites , both those within merosomes and those already released into pulmonary capillaries ( Figure 8D–8G and Video S10 ) . The small liver stage protein UIS-4 , which localizes to the PVM [38] , was not detected ( unpublished data ) ; thus providing more evidence that the PVM is not involved in merosome formation [5] . As for merosomes in the liver , pulmonary merosomes were negative by immunofluorescence for the hepatocyte receptor ASGR1 ( Figure 8F and 8G ) . In confocal images we often observed an asymmetric arrangement of individual merozoites in relation to lung merosomes and the pattern suggested some of the merosomes were in the process of disintegrating and releasing merozoites into the pulmonary microvasculature just as blood circulation was stopped by lung removal ( Figure 8B ) . Electron microscopy supports the notion of merozoite release by merosomal membrane degradation . Pulmonary merosomes typically contained morphologically well-preserved merozoites , but the cytoplasmic matrix was swollen , and host cell organelles were clearly degenerating . The membrane of lung merosomes was frequently disrupted or barely detectable ( Figure 8H ) suggesting that free merozoites found in nearby pulmonary microvasculature had just been released before fixation ( Figure 8I ) . The presence of erythrocytes containing newly invaded merozoites ( Figure 8J ) supports the notion that blood infection occurred in the lungs . To determine whether merosomes leaving the liver can pass through the lungs and disseminate throughout the body , we analyzed thick smears of blood collected from the aorta and tail vein for merosomes . We also used intravital microscopy , ex vivo imaging , and immunofluorescence microscopy to examine capillary beds of spleen , kidney , and brain of the same mice . While individual small parasites were occasionally detectable , merosomes were completely absent from aorta and tail vein blood and the microcirculation of these organs ( unpublished data ) . These results demonstrate effective retention of hepatic merosomes in the lungs . At 52 h after infection with PyGFP , mice were injected with Hoechst 33342 and PI , and the lungs were removed and analyzed ex vivo . Confocal microscopy revealed that pulmonary merosomes and free merozoites excluded PI ( Figure 9 ) and were also TUNEL-negative ( unpublished data ) ; thus providing evidence of viability . Because infectivity is the ultimate criterion for viability , we tested merosomal merozoites for their ability to induce a parasitemia in naïve mice . However , interpretation of results from inoculation with blood containing merosomes is complicated by the presence of infected erythrocytes . To circumvent this , we initially attempted to eliminate parasitized erythrocytes using selective hypotonic lysis , but this also affected the integrity of the merosomes . Our solution was to control for infected erythrocytes by comparing the infectivity of two types of blood taken from the same mouse: hepatic effluent ( with merosomes and some infected erythrocytes ) and tail vein blood ( without merosomes , but with the same number of infected erythrocytes ) . At 52 h after intravenous infection with 2 . 5 × 106 wt P . yoelii sporozoites , hepatic effluent and tail vein blood samples were collected for inoculation into other mice . Parasitemia and merosome concentration were determined by analysis of thin and thick blood smears , respectively . Preliminary studies showed that the parasitemia in recipient mice injected with hepatic effluent blood rose significantly faster compared to control mice injected with tail vein blood ( unpublished data ) suggesting that merosomes exiting the liver are infectious . This conclusion can be confirmed once a method for merosome purification is available . Because of the large number of parasites and their high packing density , counting merozoites in EEFs is not feasible , although estimates have been published [6 , 22 , 23] . We therefore used measurements obtained from merosomes to calculate the number of merozoites contained in mature P . yoelii EEFs . To do this , we first quantified the merozoite content of a subset of smaller merosomes . We found that merosomes with a diameter of 13 . 4 ± 2 . 0 μm contained 134 . 7 ± 51 . 6 merozoites . Then , we determined the average effective volume merozoites take up inside merosomes . P . berghei merozoites measure 1 . 0–1 . 2 × 1 . 5–1 . 7 μm [39–41] , but because the parasites are embedded in cytoplasm that also contains parasite remnant bodies and host cell organelles , the effective volume the parasites occupy is larger than their actual volume of 0 . 78–1 . 23 μm3 ( based on the ellipsoid volume v = 4/3 π r1 r2 r3 ) . Using a mathematical algorithm for optimal packing of small spheres in a large sphere [42] , we found that P . yoelii merozoites have an effective diameter of roughly 2 . 2 μm and occupy an effective volume of 5 . 56 μm3 in merosomes . Because intravital and electron microscopy showed that the merozoite packing density and the composition of the cytoplasm was basically identical in merosomes compared to mature EEFs ( after rupture of the PVM and mixing of parasites and host cell organelles ) , we then used the sphere packing algorithm to determine the merozoite content of mature P . yoelii EEFs . Based on the measured EEF diameter of 40–75 μm ( see above ) , the effective merozoite diameter of 2 . 2 μm , and assuming a round EEF shape , we calculated that P . yoelii sporozoites produce 4 , 200–29 , 000 hepatic merozoites ( Table S1 ) .
We present here a new model for the transition from the liver to the blood phase of the malaria life cycle ( Figure 10 ) : large merosomes of various sizes bud from infected hepatocytes , enter the hepatic circulation , exit the liver intact , subdivide into smaller more uniform sizes , but otherwise withstand bloodstream shear forces during passage through the right ventricle , and accumulate in the lungs where the merosomes disintegrate and release merozoites to initiate the erythrocytic phase of the malaria cycle . While EEF of avian and reptilian malaria parasites develop in the reticulo-endothelial or hematopoietic systems [43–45] , a major evolutionary change occurred with the mammalian malaria parasites , whose EEF mature in hepatocytes . Perhaps the nutritionally rich and immunologically privileged hepatic environment offers advantages , but it also presents a problem for merozoites released from EEFs into hepatic sinusoids: unless they invade an erythrocyte very quickly they face a gauntlet of highly phagocytic Kupffer cells . The location of most EEFs in the periportal area of the liver lobule [46] means they must travel almost the full length of the sinusoid and pass by a large complement of Kupffer cells before escaping into relative safety outside the liver . As proposed previously by us and others [14 , 15] , our premise is that evolution produced a countermeasure to this threat: release of merozoites within large packets that are initially hidden from the host's innate immune system by envelopment with a hepatocyte-derived membrane . Here we show that merosomes are delivered to the pulmonary microcirculation where they are released . We propose that release of merozoites into the lung microvasculature rather than into larger blood vessels is advantageous , because the low macrophage density and the reduced blood velocity with reduced shear forces will enhance the ability of merozoites to invade erythrocytes . Merosome disintegration in the lungs appears to be the predominant mechanism of merozoite liberation into the bloodstream for the following reasons: ( 1 ) In confirmation of previous reports on the asynchronous nature of EEF maturation [5 , 25] , we observed P . yoelii merosome formation in the liver from 46 h to 56 h after sporozoite infection . Assuming a 10-h window of merosome release , roughly 3 ml total blood volume in a 40 g mouse , and a 100% rate of sporozoite infection and EEF development , 2 . 5 million sporozoites would generate 4 , 167 maturing EEFs per minute , corresponding to 1 . 4 merosome-releasing EEFs per μl blood . ( 2 ) Assuming that extrahepatic merosomes contain on average 150 merozoites , the roughly 29 merosomes we found per μl venous liver blood should have contained 4 , 350 merozoites . Since P . yoelii EEFs contain 4 , 200–29 , 000 merozoites ( Table S1 ) , up to 74% of the total number of merozoites released by 1 . 4 EEFs per min and μl would have been enclosed in merosomes . ( 3 ) A large number of merosomes was arrested in alveolar capillaries suggesting that many merosomes withstand the shear forces inside the central cardiovascular system . Together , these data indicate that a major proportion of the merosome population arrives intact in the lungs and then gradually disintegrates , thus liberating merozoites into the microvasculature . Pulmonary merosomes were detectable in the lungs at least up to 58 h after infection , i . e . , beyond the period of release from the liver ( 46–56 h ) , suggesting that they remained intact for at least many minutes . Similar to hepatic merosomes , which appeared to be infectious and did not stain with annexin V , YO-PRO-1 , or PI , pulmonary merozoites were ultrastructurally well preserved , TUNEL-negative , and did not incorporate PI . Together , these data suggest that merosomal merozoites remain viable until their release into the pulmonary microvasculature . Based on the above assumptions , we propose that merozoite liberation in the lungs represents an integral part of the Plasmodium life cycle . Further support for our premise was found in the following observations and suggestions derived from them . The notion that merozoites shuttled out of the liver within merosomes that are protected from phagocytosis by Kupffer cells [8] was confirmed by demonstrating that murine Kupffer cells do not phagocytoze PyGFP merosomes in vitro ( unpublished data ) , in agreement with the finding that P . berghei merosomes are not ingested by a murine macrophage cell line in vitro [15] . Trager and Jensen's finding that P . falciparum merozoite invasion is enhanced by lack of flow and dense erythrocyte packing [47 , 48] supports our hypothesis that merozoites released within capillary beds have a better chance to invade erythrocytes than those released into larger vessels . We can imagine that capillary occlusion by arrested merosomes could be helpful by causing local stagnation of the pulmonary blood flow . We can also speculate that merosome arrest in lung septal capillaries allows Plasmodium to exploit the unique microenvironment of the blood-air barrier . Virtually nothing is known about the biology of the first-generation ( hepatic ) merozoites , but perhaps transient residence in the lungs provides these parasites with time and a suitable microenvironment to gain infectivity for erythrocytes . The well-oxygenated milieu of the terminal airways and the anastomozed nature of the pulmonary microvasculature [49] likely allow local occlusion of septal capillaries by merosomes without causing the necrotic tissue damage associated with infarction of microvessels in other organs . Many aspects of the process of merosome formation and release we describe are in agreement with earlier work , but others are not . For example , we found that similar to P . berghei–infected HepG2 cells , which detach in toto from the culture vessel after merozoite differentiation is complete [15] , merosomes exiting P . yoelii–infected mouse livers contain viable merozoites and initially do not expose PS on their surface . This confirms earlier predictions [14 , 15] that merozoites are safely shuttled out of the liver disguised as merosomes . The presence of intact mitochondria in mature EEFs indicates that Plasmodium liver stages are able to manipulate hepatocytes in a way that useful organelles ( such as mitochondria as a source of energy ) are preserved , even after merosome budding . Our interpretation , namely that Plasmodium controls certain host cell functions to the last minute , differs from the P . berghei HepG2 cell model , in which the parasites induce death and detachment of their host cells followed by merosome budding [15] . Further , the cell membrane of P . yoelii–infected hepatocytes remains in close apposition to that of neighboring parenchymal and endothelial cells until the very end of EEF differentiation , i . e . , up to the onset of merosome budding , as reported [5 , 12 , 13 , 50 , 51] . As merosomes are produced , the host cell gradually decreases in size and loses contact with neighboring cells as reported [15] . We observed that after releasing merosomes over several hours , the exhausted host cell eventually disintegrates . Some free merozoites still escaped and entered the sinusoidal lumen , thus being exposed to attack by Kupffer cells . In contrast , others proposed that the remaining host cell remnant is rapidly expelled in toto from the tissue with the resulting void immediately filled by neighboring cells [15 , 18] . We find that the necrotic remnant attracts neutrophils and mononuclear phagocytes , which eventually produce a small granuloma . Such granulomata are a frequent observation in P . yoelii– and particularly in P . berghei–infected mouse livers [5 , 8 , 27–31] . Rather than the void created by expulsion of an EEF being filled quickly , our in vivo observations suggest that hours , if not days , are required for phagocytic removal of parasite and host cell debris with subsequent repair of the structural damage before normal tissue architecture is restored . Although we found merosome formation to be the predominant mode of merozoite release from the liver , we observed a less frequent but still common alternative: EEFs undergoing what we interpret as decay . This alternative process of EEF ghost formation was rapid and typically complete within minutes to an hour . In contrast to merozoite release by merosome formation , ghost-forming EEFs did not detach from the surrounding tissue . EEF decay was accompanied by leakage of GFP into the surrounding tissue suggesting damage to the host cell membrane . It occurred in immature EEFs ( recognizable by a homogeneous green fluorescent cytoplasm ) and also in mature EEFs ( containing fully formed merozoites ) without merosome formation regardless of maturity . Sometimes it was found as early as 42 h after sporozoite infection , hours before merozoite differentiation begins . The end result of this alternative process was the formation of large faintly fluorescent EEF ghosts containing some cellular debris and a few dead merozoites . We interpret this rapid conversion of EEFs to ghosts as abortive liver stage development . Merozoite content of EEFs has historically been difficult to estimate due to the large number of parasites and their high packing density . Based on measurements of the size and merozoite content of small merosomes combined with size measurements of EEFs and an appropriate mathematical algorithm [42] , we were able to calculate the number of merozoites in an EEF ( Table S2 ) . Under intravital imaging conditions , mature P . yoelii EEFs measured 40–75 μm and the calculated space effectively occupied by a merozoite is a sphere of 2 . 2 μm diameter . Using this effective size , we calculated that individual P . yoelii sporozoites produce roughly 4 , 200–29 , 000 merozoites per EEF . This number is in general agreement with older estimates of EEF merozoite content [5 , 12 , 22 , 23 , 52–57] ( Table S1 ) . An exception is P . falciparum , which produces considerably larger numbers of hepatic merozoites , most likely because of the small size of the parasites . As far as we know , our analysis of the number of merozoites produced in hepatocytes is the first such analysis based on actual merozoite counts and host cell measurements . Precision is limited by variations in measurements , but basing calculations on direct in vivo measurements enhances accuracy . Earlier studies conducted by us and others had suggested that merosome budding may precede completion of merozoite differentiation [14 , 15] . One factor that helped lead to this interpretation is that GFP expressed in the parasite stroma can obscure the parasites in mature EEFs . We now show that prior to merosome formation , the signal of the stromal GFP fluorescence equaled that of the merozoite cytoplasm , thus preventing clear definition of parasites enmeshed in the stroma . At the onset of merosome budding , the stromal GFP emission signal decreased abruptly thus revealing the presence of the already formed fluorescent parasites ( Figure 3A–3E and Video 5 ) . Two factors contribute to this reduction in fluorescence of material surrounding the parasites: dilution and loss of cytosolic GFP . Dilution of GFP results from PV disassembly and mixing of fluorescent parasite stroma with non-fluorescent host cytoplasm . Loss of GFP is caused by leakage of the fluorochrome into the environment . In agreement with reports that the hepatocyte membrane becomes permeable at late stages of infection with P . berghei [5] , we found that merosome-forming EEFs are typically surrounded by a halo of green fluorescence . Optimization of the imaging conditions allowed us to visualize the parasites inside mature EEFs and revealed that merosomes always contain mature merozoites . Thus , merozoites maturation precedes merosome formation . Depending on the approach used for measurement , the reported diameters of hepatic and pulmonary capillaries vary greatly . For example , when measured in perfusion-fixed liver tissue , the sinusoidal diameter ranged from 4–6 μm to 9–12 μm [58–60] . A crucially important factor is the pressure applied during perfusion fixation , because the sinusoidal diameter is known to vary with changes in blood pressure [61 , 62] . To determine the sinusoidal diameter under normal blood pressure conditions , we used live Tie2-GFP mice [32] , whose fluorescent endothelia clearly delineate the boundaries of the sinusoidal lumen [31] . In agreement with earlier in vivo microscopic studies , which reported a diameter of 6 μm for portal sinusoids and 7 μm for central sinusoids [58] , we found by intravital imaging that liver sinusoids measure 6 . 7 ± 1 . 9 μm in diameter . Similar differences between fixed and live specimens were reported for the size of alveolar capillaries . While vascular casts of the lung suggested that alveolar capillaries measure 6 . 69 ± 1 . 39 μm in diameter [63] , intravital measurements determined a functional diameter of only 1–4 μm [64 , 65] . Regardless which liver sinusoid and lung capillary measurements are relied upon and regardless of the drastic reduction in merosome size after leaving the liver , merosomes still exceed the size of the lumen of the microvasculature of both liver and lung . Since even the largest merosomes were eventually transported out of the liver , the much smaller extrahepatic merosomes would be expected to be malleable enough to be able to pass though the pulmonary capillary bed . Therefore it is somewhat surprising that the lungs effectively clear the blood of all merosomes , so virtually none were detectable in arterial blood harvested from the left ventricle , in the capillary beds of spleen , brain and kidney , or in tail vein blood . The fact that the velocity in pulmonary capillaries is somewhat higher than hepatic sinusoids [66–69] makes this more unexpected . Consequently , the possibility of a receptor-mediated mechanism for pulmonary merosome arrest cannot be excluded .
Anopheles stephensi mosquitoes were used to propagate wild-type P . yoelii ( strain 17 XNL ) or PyGFP [14 , 70] . Sporozoites were purified from the salivary glands of female A . stephensi mosquitoes [71] . Mice were ( 1 ) Balb/c ( Taconic Farms , Incorporated ) , ( 2 ) Swiss Webster ( Taconic Farms , Incorporated ) , or ( 3 ) Tie2-GFP mice , a transgenic strain that expresses GFP in vascular endothelial cells under control of the Tie2 promoter ( STOCK Tg ( TIE2GFP ) 287Sato/J; Jackson Laboratory ) [31 , 32] . Animals were maintained and used in accordance with recommendations in the guide for the Care and Use of Laboratory Animals . Mice were inoculated into the tail vein with 0 . 3–1 . 5 × 106 PyGFP sporozoites . At 30–66 h p . i . , the animals were surgically prepared for intravital imaging of liver and spleen as described [31] and anesthetized by intraperitoneal injection of a cocktail of 50 mg/kg ketamine ( Ketaset , Fort Dodge Animal Health ) , 10 mg/kg xylazine ( Rompun , Bayer ) , and 1 . 7 mg/kg acepromazine ( Boehringer Ingelheim Vetmedica ) . Reinjection of the anesthetics at 30-min intervals allowed intravital microscopic examination of the animals for at least 3 h [31] . After surgical preparation for intravital imaging , mice were placed onto the stage of an inverted Zeiss DMIRE2 microscope , equipped with a temperature-controlled Ludin chamber , and analyzed with a Leica TCS SP2 AOBS confocal microscope . Appropriate laser lines were used to excite GFP , various other fluorochromes , and the natural autofluorescence of the mouse tissues . Laser power was reduced to a minimum to avoid phototoxicity and bleaching . These optimized conditions allowed continuous scanning of live PyGFP for a period of up to 6 h without any apparent effect on viability . To assess parasite and host cell viability , some mice were i . v . injected with 1–2 μg/ml of the membrane-permeable nuclear dye Hoechst 33342 prior to confocal microscopy . Other mice received 1 μg/ml PI in addition to detect dead host cells and/or parasites . Mice were intravenously inoculated with 3 ×106 purified wt P . yoelii or 1 × 106 PyGFP salivary gland sporozoites and various organs were removed at 52 h after infection . Tissue slices were snap-frozen in liquid nitrogen or fixed with PBS containing 4% paraformaldehyde for immunofluorescence labeling of cryosections and with PBS containing 4% paraformaldehyde and 1% glutaraldehyde for electron microscopic examination [72 , 73] . At 30–66 h after infection with PyGFP , major organs such as spleen , brain , kidney , or lung were removed , placed into glass-bottom dishes , and kept moist with medium for confocal microscopy analysis . Blood was harvested from ( 1 ) the terminal hepatic vein , ( 2 ) the aorta , or ( 3 ) a tail vein . To increase the probability of detection , ten aliquots of 5 μl blood from each of these sites were spread over an area of 1 cm2 , allowed to dry , and stained with Giemsa without prior fixation . Merosomes were counted and expressed as average number ± STD . In parallel , the number of merozoites per merosome was determined accordingly . Two days after infection with 1 . 5 × 106 wt P . yoelii sporozoites , hepatic effluent and tail vein blood was harvested from the same animal and parasitemia and merosome content were determined using thin and thick blood smears , respectively . 20-μl hepatic effluent , containing 1 × 105 infected erythrocytes plus 167 merosomes , or tail vein blood containing the same number of infected erythrocytes but no merosomes , was intravenously inoculated into Swiss Webster mice ( three mice per group ) and the parasitemia was monitored daily by Giemsa staining . To improve the recovery of parasite material from the liver , merosomes were dislodged from hepatic sinusoids by perfusing mouse livers via the portal vein with oxygenated medium at 5 ml/min for 10–30 min . The effluent was collected in two fractions: fraction 1 was collected from the Vena cava inferior and contained mainly red blood cells; fraction 2 was collected from the Vena cava superior after ligation of the Vena cava inferior . The cells were washed and allowed to settle onto cover slips or glass-bottom dishes ( WillCo Wells ) treated with Alcian blue [74] for live cell imaging . Nuclei of merosomes were visualized with the membrane permeable nucleic acid stains Hoechst 33342 ( 1–2 μg/ml ) or SYTO-64 . Nuclei of dead parasites were determined with membrane impermeable PI ( 1 μg/ml ) . Merosome membranes were stained with 5 μg/ml FM 4–64 FX ( Molecular Probes ) . Annexin V Alexa Fluor 488 conjugate or YO-PRO-1 ( 0 . 1 μM ) were used to detect evidence of programmed cell death in live merosomes . Tissue sections were stained with a BrdU TUNEL assay kit ( Molecular Probes ) according to manufacturer's guidelines . Alcian blue–immobilized PyGFP merosomes were fixed and labeled with the red nuclear dye SYTO-64 . 3-D stacks were scanned by confocal microscopy and the number of merozoite nuclei was counted using a 3-D object count plug-in of ImageJ ( NIH freeware ) . Merozoite number and merosome diameter were then entered into a formula for efficient packing of equal small spheres in a large sphere ( n = 0 . 7405 [1–2D] / D3 + 1 / [2D2]; D = dmerozoite / dliver stage ) [42] to determine the effective diameter/volume merozoites occupy inside merosomes . Based on these calculations and the diameter of PyGFP liver stages measured by intravital microscopy , the merozoite content of P . yoelii liver stages was estimated in relation to size . Frozen sections of 10-μm thickness were prepared with a Reichert-Jung Frigocut cryostat . Parasites were labeled with a mAb directed against the P . yoelii merozoite surface protein MSP-1 , a kind gift from W . Bergman [75] . A rabbit antiserum , which was originally generated against the PVM-associated protein from P . berghei , but exhibits cross-reactivity with P . yoelii UIS4 [38 , 76] , was used to label the PV in P . yoelii–infected hepatocytes . Affinity-purified goat IgG against the murine asialoglycoprotein receptor ASGR1 was from R&D Systems . Incubation with the primary antibodies was followed with protein A conjugated to fluorescein isothiocyanate ( PA-FITC; Molecular Probes ) , anti-goat IgG conjugated to Texas Red ( GAR-TR; Molecular Probes ) , or goat anti-rabbit IgG conjugated to Texas Red ( GAM-TX; Molecular Probes ) in color-matching fluorochrome combinations . In case of a single FITC label , the specimens were counterstained with 0 . 1% Evans blue in PBS . Immunofluorescence-labeled frozen tissue sections were examined by confocal microscopy . Mouse liver or lung tissue was fixed with 1% glutaraldehyde and 4% paraformaldehyde in PBS , post-fixed with 1% osmium tetroxide and 1 . 5% potassium hexacyanoferrate , stained en bloc with 1% uranyl acetate , dehydrated in ethanol , and embedded in Epon as described [72 , 73] . Semithin sections were cut with an RMC MT-7 ultramicrotome and photographs were taken with Kodak Ektachrome 160T slide film using a Nikon FX-35DX/UFX-DX camera/exposure system . Thin sections were post-stained with uranyl acetate and lead citrate and viewed with a Zeiss EM 910 electron microscope [73] . Electron microscopy negatives and Ektachrome slides were scanned with a Hewlett Packard Scanjet 5370C . All digital , electron , or confocal microscopy images were processed using Image-Pro Plus ( Media Cybernetics ) , Adobe Photoshop ( Adobe ) , and AutoDeBlur ( AutoQuant Imaging , Incorporated ) software .
|
The malaria parasite Plasmodium undergoes one large round of multiplication in the liver before beginning the blood phase of the life cycle , the phase that causes the typical episodes of fever and chills . Using intravital microscopy and fluorescent parasites , we studied the mode and dynamics of parasite release from the liver , a critical stage in the malaria life cycle . Earlier work had indicated that infected liver cells could release packets of dozens to hundreds of parasites enveloped by host cell membrane , structures now known as merosomes . We report here that this is the predominant mechanism of parasite release from the liver . The host-derived merosome membrane lacks a marker for phagocytic engulfment , thus allowing safe passage through the gauntlet of Kupffer cells , highly active liver macrophages . Merosomes remain intact during passage through the heart and become sequestered within lung capillaries where the membrane eventually disintegrates liberating the parasites into the lung circulation . We propose that this previously unknown part of the life cycle of Plasmodium facilitates red blood cell invasion , thus jump-starting the blood phase of the life cycle and the onset of clinical malaria .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"infectious",
"diseases",
"cell",
"biology",
"plasmodium",
"gastroenterology",
"and",
"hepatology",
"in",
"vitro",
"immunology",
"microbiology",
"mus",
"(mouse)"
] |
2007
|
Release of Hepatic Plasmodium yoelii Merozoites into the Pulmonary Microvasculature
|
Nosocomial infections are increasingly being recognised as a major patient safety issue . The modern hospital environment and associated health care practices have provided a niche for the rapid evolution of microbial pathogens that are well adapted to surviving and proliferating in this setting , after which they can infect susceptible patients . This is clearly the case for bacterial pathogens such as Methicillin Resistant Staphylococcus aureus ( MRSA ) and Vancomycin Resistant Enterococcus ( VRE ) species , both of which have acquired resistance to antimicrobial agents as well as enhanced survival and virulence properties that present serious therapeutic dilemmas for treating physicians . It has recently become apparent that the spore-forming bacterium Clostridium difficile also falls within this category . Since 2000 , there has been a striking increase in C . difficile nosocomial infections worldwide , predominantly due to the emergence of epidemic or hypervirulent isolates that appear to possess extended antibiotic resistance and virulence properties . Various hypotheses have been proposed for the emergence of these strains , and for their persistence and increased virulence , but supportive experimental data are lacking . Here we describe a genetic approach using isogenic strains to identify a factor linked to the development of hypervirulence in C . difficile . This study provides evidence that a naturally occurring mutation in a negative regulator of toxin production , the anti-sigma factor TcdC , is an important factor in the development of hypervirulence in epidemic C . difficile isolates , presumably because the mutation leads to significantly increased toxin production , a contentious hypothesis until now . These results have important implications for C . difficile pathogenesis and virulence since they suggest that strains carrying a similar mutation have the inherent potential to develop a hypervirulent phenotype .
C . difficile is the causative agent of a spectrum of gastrointestinal diseases , collectively known as C . difficile infections , or CDI , that are induced by treatment with antibiotics that disrupt the normal gastrointestinal microbiota . CDI can range from mild diarrhoea , through moderately serious disease , to severe life-threatening pseudomembranous colitis , a chronic , often fatal , gastrointestinal disease [1] . During the past decade , there has been an astonishing increase in the rate and prevalence of C . difficile infections in many parts of the world , including the UK , USA , Canada and Europe , largely due to the emergence of a “hypervirulent” or epidemic group of isolates belonging to the BI/NAP1/027 category [2] , [3] . These strains are highly resistant to fluoroquinolones [3] and are associated with more severe disease and higher mortality rates [4]–[7] . C . difficile now also causes disease in those previously not at risk , such as children and pregnant women , with community-associated C . difficile disease being increasingly common [8]–[10] . The reasons for the emergence of these strains , and for their increased virulence , remain largely speculative . The use of fluoroquinolones , and the emergence of fluoroquinolone resistant strains , are undoubtedly driving factors in these new epidemics [11] , however , the reasons for the heightened virulence and persistence of these strains are unknown . Genotypic and phenotypic comparison of the hypervirulent BI/NAP1/027 isolates to historical strains has identified numerous differences that may contribute to hypervirulence . Phenotypically , these differences may include the production of a toxin known as binary toxin , or CDT [3] , and a higher sporulation rate [12] . Whole genome comparisons have identified numerous genetic differences with BI/NAP1/027 strains having an additional 234 genes compared to the well characterised strain 630 [13] , including five unique genetic regions that are absent from both strain 630 and non-epidemic 027 strains [4] . Fundamentally , however , the factors directly resulting in the development of hypervirulence by these strains remain unknown . The major virulence factors of C . difficile are two members of the large clostridial cytotoxin family , toxin A and toxin B , encoded by the tcdA and tcdB genes , respectively , which are potent monoglucosyltransferases that irreversibly modify members of the Rho family of host regulatory proteins [14] . Two recent studies definitively showed that toxin B plays a major role in the virulence of C . difficile [15] , [16] . The role of toxin A in disease was less clear however , with conflicting data concerning toxin A reported [15] , [16] . Epidemic strains are reported to produce significantly more toxin A and toxin B than other strains [2] . The tcdA and tcdB genes are located on the chromosome within a region known as the pathogenicity locus or PaLoc [17] . In addition to tcdA and tcdB , the PaLoc encodes three additional genes designated tcdR , tcdE and tcdC , which encode an alternative sigma factor , TcdR [18] , a putative holin , TcdE [19] , and an anti-sigma factor , TcdC [20] , respectively . The expression of toxins A and B is controlled in a complex manner by several factors , including TcdR and TcdC . TcdC is thought to negatively regulate toxin production by interacting with TcdR or with TcdR-containing RNA polymerase holoenzyme or both [20] , TcdR is essential for toxin production [18] . BI/NAP1/027 C . difficile strains have a nonsense mutation in tcdC , which results in the production of a truncated protein that no longer negatively regulates TcdR . This mutation is postulated to be responsible for the increased toxin production observed in vitro in these strains [2] . Accordingly , this observation has prompted debate over the importance of the tcdC mutation in the hypervirulent phenotype . However , there is currently a lack of experimental evidence to support this hypothesis , with inconsistent reports in the published literature [20]–[22] . Despite their important impact worldwide on public health little is known about the virulence factors of BI/NAP1/027 strains and many important questions about the pathogenesis of disease caused by these strains remain to be answered , especially the role played by TcdC . BI/NAP1/027 isolates have proven difficult to genetically manipulate , which has hampered our ability to study these strains at the molecular level . To address these questions , here we use a novel Tn916-based plasmid conjugation system to facilitate the efficient transfer of plasmids into BI/NAP1/027 strains of C . difficile . Using this system , we have demonstrated conclusively the role of TcdC as a negative regulator of toxin production in C . difficile . Furthermore , using the hamster model of infection , we provide evidence to show that the tcdC mutation found in BI/NAP1/027 strains is an important factor in the development of hypervirulence by these strains . This study is the first to use isogenic strains to identify a factor involved in the development of a hypervirulent phenotype in C . difficile , and also represents the first in vivo demonstration of the role of TcdC in the pathogenesis of C . difficile disease .
To determine if mutation of the tcdC gene in C . difficile BI/NAP1/027 isolates leads to the development of a hypervirulent phenotype it was necessary to construct isogenic BI/NAP1/027 strains that only differed in their ability to produce a functional TcdC protein . To construct the isogenic strains required for this analysis , genetic manipulation of BI/NAP1/027 isolates was required . The genetic manipulation of these strains has proved difficult and attempts to transfer plasmids into BI/NAP1/027 strains using published methods , which rely on RP4-mediated conjugation from Escherichia coli [23]–[27] , were not successful , even though transfer of plasmids into the genetically amenable strains JIR8094 , an erythromycin sensitive derivative of strain 630 [24] , and CD37 was readily achieved ( Table S1 ) . To overcome this barrier and to facilitate DNA transfer into the strains of interest , we developed a novel plasmid transfer system that exploits the conjugation apparatus encoded by the broad-host range transposon Tn916 . The oriT region of Tn916 ( oriTTn916 ) [28] was cloned into the catP-containing C . difficile shuttle plasmid pMTL9361Cm [29] , generating pDLL4 . This plasmid was introduced into C . perfringens strain JIR4225 , which contains five copies of Tn916 [30] and plate matings were performed between this donor strain and several C . difficile strains , including a BI/NAP1/027 strain , M7404 , which is a Canadian epidemic isolate [29] . Transconjugants from these matings were isolated on medium supplemented with thiamphenicol and cefoxitin . The efficiency of plasmid transfer into strain M7404 was 1 . 2×102–4×104 transconjugants/ml of plated culture . Analysis of transconjugants using PCR specific for the catP gene together with restriction analysis confirmed that all putative colonies carried pDLL4 ( data not shown ) , verifying successful plasmid transfer into the BI/NAP1/027 strain M7404 . Similar plasmid transfer efficiencies were obtained for numerous other C . difficile strains ( Table S1 ) , highlighting the utility of this methodology for the genetic manipulation of clinically relevant strains . To complement the tcdC mutation in a BI/NAP1/027 strain the intact tcdC gene from strain VPI10463 , together with 300 bp of its upstream region , was cloned into the shuttle plasmid pDLL4 , generating pDLL17 . This plasmid was transferred by Tn916-mediated conjugation from C . perfringens strain JIR4225 to C . difficile strain M7404 as before . PCR was subsequently used to confirm the presence of plasmid pDLL17 in representative transconjugants ( data not shown ) . To determine whether the presence of pDLL17 complemented the TcdC deficiency of M7404 , Western immunoblots using TcdC-specific antibodies were performed . Lysates were collected from the wild-type M7404 , the pDLL4-carrying vector control strain M7404 ( VC ) and the pDLL17-tcdC+ strain M7404 ( tcdC+ ) , as well as strains VPI10463 and the PaLoc-deficient strain VPI11186 , which served as positive and negative controls , respectively . An additional control strain , M7404 ( cured ) , was generated by serially passaging strain M7404 ( tcdC+ ) on non-selective growth medium and curing the plasmid from this strain . Loss of the plasmid was confirmed by sensitivity of the strain to thiamphenicol followed by PCR analysis to verify the absence of several plasmid encoded genes ( data not shown ) . As Figure 1A shows , whilst no TcdC could be detected in the lysates of the negative control strain , the wild-type M7404 , M7404 ( VC ) and the plasmid-cured strain M7404 ( cured ) , a 34-kDa protein that reacted with TcdC-specific antibodies was detected in lysates from the tcdC+-complemented strain M7404 ( tcdC+ ) . This band was the same size as the immunoreactive TcdC protein produced by the positive control strain VPI10463 , confirming that the tcdC mutation in the BI/NAP1/027 epidemic isolate M7404 was efficiently complemented in trans . Since complementation was performed using a multicopy plasmid , we also quantified TcdC production levels from strain M7404 ( tcdC+ ) in comparison to strain VPI10463 using a time-course assay . Previous studies involving transcriptional analysis of PaLoc genes during different growth phases showed that tcdC is expressed in early exponential phase but not in stationary phase , whereas the other PaLoc genes show the opposite expression pattern [31] . VPI10463 ( Figure 1B ) and M7404 ( tcdC+ ) ( Figure 1C ) exhibited similar TcdC expression patterns , with higher levels of TcdC observed in early exponential phase and negligible amounts detected beyond 16 hours , suggesting that the regulatory regions governing tcdC expression have been retained on the tcdC-carrying fragment used to construct pDLL17 . In addition to the kinetics of TcdC expression in strain M7404 ( tcdC+ ) mirroring that of VPI10463 , a similar amount of protein was also detected at each time point with VPI10463 producing 1 . 3- to 1 . 6-fold more protein ( Figure 1B ) than M7404 ( tcdC+ ) ( Figure 1C ) . Therefore , although tcdC complementation was achieved using a multicopy plasmid vector , a physiologically relevant amount of TcdC protein was expressed during the appropriate growth phases in strain M7404 ( tcdC+ ) . To determine the effect of TcdC on toxin production in strain M7404 , a combination of Western immunoblots and cytotoxicity assays were performed using supernatants collected from strain M7404 and isogenic M7404 derivatives carrying the vector pDLL4 , pDLL17 , the cured strain M7404 ( cured ) and the PaLoc-negative control strain CD37 . To assess toxin A production , Western immunoblotting was performed using TcdA-specific antibodies ( Figure 2A ) . The results showed that the presence of the tcdC+ plasmid pDLL17 resulted in a dramatic decrease in the amount of toxin A produced by M7404 ( tcdC+ ) when compared to the wild-type strain . By contrast , M7404 ( cured ) produced qualitatively similar levels of toxin A to wild-type , as did M7404 carrying the vector plasmid , whereas the PaLoc-negative strain CD37 produced no detectable toxin A , as expected . Vero cell cytotoxicity assays , which predominantly measure toxin B activity [15] , were then performed to quantitatively determine the effect of TcdC on toxin production from these strains . As previously observed with toxin A , the amount of toxin produced from strain M7404 was significantly reduced when functional TcdC was restored ( Figure 2B ) . The amount of toxin produced in the TcdC-complemented strain was approximately 16–32-fold less , and therefore significantly lower ( p = 0 . 0001; unpaired t-test , 95% confidence interval ) , than in the vector control-carrying M7404 derivative . There was , however , no significant difference in toxin activity levels between strains M7404 , the vector control strain M7404 ( VC ) or M7404 ( cured ) ( Figure 2B ) . A kinetic analysis of toxin production also clearly showed that the presence of TcdC delayed the onset of toxin production in M7404 ( tcdC+ ) in comparison to M7404 carrying the vector plasmid , mirroring the delayed toxin production observed from the tcdC+ 630-strain derivative JIR8094 ( Figure 2C ) . We also determined if TcdC-mediated repression of toxin production was at the transcriptional level and evaluated the effect of tcdC complementation on the expression of the other PaLoc-encoded genes , tcdR and tcdE . Quantitative real-time PCR ( qRT-PCR ) analysis using RNA extracted from the wild-type strain and its isogenic derivatives was performed to ascertain the relative transcription levels of the tcdA , tcdB , tcdR and tcdE genes . As shown in Figure 2D , an approximate 13- and 23-fold reduction in tcdA- and tcdB-specific mRNA levels , respectively , in strain M7404 ( tcdC+ ) was observed compared to M7404 . Similar observations were made for tcdR and tcdE expression levels , with 33-fold and 21-fold less tcdR and tcdE mRNA , respectively , in the tcdC-complemented strain compared to the wild-type . No significant differences in the expression levels of these four genes were detected when M7404 , the vector-carrying derivative or the cured strain were compared . These data conclusively demonstrate that TcdC negatively regulates toxin production in C . difficile and show that repression occurs at the transcriptional level . To define the role of TcdC in the virulence of a BI/NAP1/027 C . difficile isolate , female Golden Syrian hamsters were infected with spores of strain M7404 carrying either the vector control or the tcdC+ plasmid ( n = 10 and n = 12 , respectively ) . For comparative purposes , a group of hamsters ( n = 14 ) was also infected with strain 630 , a strain previously characterised as being less virulent than other clinical isolates [32] . Following infection , all C . difficile strains were found to be equally efficient at colonising the hamsters ( data not shown ) . Infection of colonised hamsters was allowed to proceed and animals were monitored by telemetry . The end point of infection was achieved when the core body temperature of the hamsters dropped to 35°C . This parameter has previously been shown to be a reliable indicator of non-recoverable disease [33] . At this point , the animals were immediately culled for animal ethics reasons . Bacteria were then isolated from the culled hamsters , the bacterial load quantified and isolates subjected to MVLA analysis [33] to confirm that these isolates were the same strain as originally used for infection . Hamsters infected with the M7404 ( tcdC+ ) derivative showed a significant delay ( p = 0 . 0003; Logrank ( Mantel-Cox ) test; 95% confidence interval ) in the mean time taken to reach non-recoverable disease ( 2370 minutes or 39 . 5 hours ) in comparison to the vector-carrying M7404 group ( M7404 ( VC ) ) , with a mean time of 1869 minutes or 31 . 15 hours ( Figure 3 ) . In one of the hamsters colonised with the M7404 ( VC ) strain , the time taken to reach the end point of infection was substantially longer than the other hamsters in this group ( 2814 minutes or 46 . 9 hours ) . This hamster was shown by statistical analysis ( p = 0 . 0405; Grubbs test; 95% confidence interval ) to be an outlier and was therefore excluded from the experimental analysis . Note that statistical significance would be retained upon inclusion of this outlier . Interestingly , whilst the mean time to the end point of infection in the strain 630 group of hamsters ( 2701 minutes or 45 . 02 hours ) was significantly longer than that of hamsters infected with M7404 ( VC ) ( p = 0 . 0001; Logrank ( Mantel-Cox ) test; 95% confidence interval ) , there was no significant difference in the mean time taken to achieve non-recoverable disease in the 630 group compared to the M7404 ( tcdC+ ) derivative , indicating that the virulence of the TcdC-complemented strain was equivalent to that of strain 630 . It is apparent from these virulence experiments that the expression of TcdC in a BI/NAP1/027 isolate has an important effect on virulence , resulting in a significant delay in the time needed to reach non-recoverable disease . These data therefore provide compelling evidence that the naturally occurring mutation of tcdC in BI/NAP1/027 isolates is an important factor in the development of a hypervirulent phenotype by these strains . The results presented here show that a BI/NAP1/027 strain complemented with tcdC is not as virulent as its isogenic vector-carrying control , suggesting that any C . difficile strain that acquires a null TcdC phenotype has the potential to develop a hypervirulent phenotype . Furthermore , phylogenetic studies have shown C . difficile to be a genetically diverse species , with disease-causing isolates seemingly arising from multiple lineages , suggesting that virulence in these strains may have evolved independently [4] , [34] . The tcdC status of a diverse group of clinical isolates was therefore determined in parallel to the genetic studies described above . One hundred Australian clinical isolates were initially analysed by toxinotyping , a typing method which categorises strains according to variation in the PaLoc region; BI/NAP1/027 strains belong to toxinotype III [35] . Approximately 5% of these strains were found to belong to a toxinotype associated with a tcdC mutation . PCR and sequence analysis was then used to confirm the presence of tcdC mutations in each of these isolates . A single BI/NAP1/027 was identified in this survey , designated strain KI [36] . One other strain , DLL3053 , was particularly interesting since it belonged to toxinotype group III but was not a BI/NAP1/027 isolate . This strain harboured the well documented single base pair deletion at nucleotide position 117 of tcdC , and the 18 base pair in-frame deletion from nucleotide 330 to 347 [37] . Two strains , DLL3054 and DLL3055 , were toxinotype V strains with a nonsense mutation at nucleotide position 184 ( C184T ) and a 39 base pair deletion from nucleotides 341 to 379 [37] whereas strain DLL3056 was toxinotype XIV and possessed a nonsense mutation at nucleotide position 191 ( C191A ) and an in-frame 36 base pair deletion from nucleotide 300 to 336 . To examine the impact of tcdC mutations on toxin production in these clinical isolates , Vero cell cytotoxicity assays were performed . These assays , which predominantly detect toxin B , determine the relative amounts of toxin produced by each strain since the expression of toxins A and B is coordinately regulated [23] , [31] . Strains JIR8094 , a derivative of strain 630 [24] , and VPI10463 [38] , which both possess intact tcdC genes , were used as positive reference controls , and the PaLoc-negative strain CD37 was used as a negative control . Strains were grown in glucose-free medium since glucose has previously been shown to repress toxin production [23] . As shown in Figure 4A , the relative amount of toxin produced by the tcdC clinical isolates varied over approximately a 10-fold range . In agreement with the work of others [12] , [22] , these results show that the presence of mutations within tcdC was not directly correlated with high level toxin production . Strain DLL3053 for example , did not produce levels of toxin significantly different from that of strain JIR8094 , which is considered to be a low toxin producer [39] . Of all the isolates with tcdC mutations , strain KI [36] produced the most toxin , and at significantly ( p = 0 . 0406; unpaired t-test , 95% confidence interval ) higher levels that were approximately 10-fold more than strain DLL3053 even though both strains have identical tcdC alleles . In comparison to the TcdC-positive reference strains , all of the tcdC-deficient clinical isolates produced more toxin than strain JIR8094 apart from strain DLL3053 ( Figure 4A ) . Conversely , however , all strains produced significantly less toxin than VPI10463 ( p = 0 . 0099–0 . 0202; unpaired t-test , 95% confidence interval ) , including the BI/NAP1/027 strain KI . VPI10463 produced over 100-fold more toxin than strains JIR8094 and DLL3053 and over 30-fold more than KI ( Figure 4B ) . This observation is in agreement with recently published findings , which showed that VPI10463 produced significantly more toxin than other strains [12]; however , in that study , strains were grown using glucose-rich BHI medium so differential effects of glucose on toxin production by each strain could not be ruled out . In agreement with other studies [12] , [22] , [37] , our data therefore suggest that the tcdC-status alone of C . difficile isolates is not an accurate predictor of high-level toxin production .
The hypothesis that the naturally-occurring tcdC mutation in epidemic BI/NAP1/027 isolates contributes to hypervirulence is widely accepted , despite a lack of supportive experimental evidence . Indeed , the exact role of TcdC in the pathogenesis of C . difficile disease has remained controversial with conflicting findings reported in the literature [20]–[22] . As a result , several published studies have suggested that there is a need to assess isogenic tcdC strains in order to conclusively determine the role of this gene in the virulence of C . difficile [12] , [22] , [40] . We have now constructed such isogenic strains and compared them in an animal model . The results conclusively show that TcdC negatively regulates toxin production in C . difficile . Most importantly , complementation of the tcdC mutation in the BI/NAP1/027 epidemic isolate M7404 clearly showed that this mutation is an important factor in the development of hypervirulence by this strain since the genetic complementation of tcdC reduced virulence in comparison to the wild-type strain . To elucidate the role of TcdC in hypervirulence , it was necessary to construct an isogenic panel of BI/NAP1/027 strains that were identical except for the presence or absence of the wild-type tcdC gene . Despite the publication of studies describing the successful transfer of plasmids into the BI/NAP1/027 isolate R20291 [25] , [26] , [41] , this group of strains has remained difficult to work with at the molecular genetic level . As such , a new system that utilised the conjugation apparatus of Tn916 was developed in this study and used successfully to genetically manipulate a number of clinically relevant isolates , including a BI/NAP1/027 strain of C . difficile . Tn916 is a broad host-range conjugative transposon that was recently used to transfer plasmids into genetically intractable strains of Enterococcus faecium [28] and has been shown to transfer into C . difficile [42] , [43] . C . perfringens was chosen for use as a donor strain in anticipation that it may be more proficient for the transfer of plasmids into C . difficile in comparison to the more distantly related E . coli . The addition of oriTTn916 onto the shuttle vector pMTL9361Cm facilitated the efficient transfer of this plasmid into strain M7404 from a Tn916-carrying C . perfringens strain . Furthermore , this system has been successfully used to transfer shuttle plasmids into every C . difficile isolate tested so far ( Table S1 ) . Most importantly , this new technology facilitated the complementation of the tcdC mutation in strain M7404 enabling the role of TcdC in the virulence of BI/NAP1/027 strains of C . difficile to be investigated . Previous in vitro studies have shown that TcdC is able to sequester the TcdR sigma factor , preventing its association with core RNA polymerase and blocking toxin gene expression [20] . These experiments suggested that TcdC was important in the regulation of toxin production by C . difficile , but the in vivo role of this protein was not determined . Conversely , several studies on C . difficile clinical isolates [22] , [37] showed that the absence of a functional tcdC gene was not an accurate predictor of high level toxin production or increased disease severity , indicating that TcdC may not play an important role in virulence in these strains [22] , [37] . The analysis of Australian clinical isolates in the present study is in accordance with these latter studies in that isolates with naturally occurring tcdC mutations were found to produce toxin at a range of different levels that were not necessarily high . However , since these strains , and those in the other studies [22] , [37] , are not isogenic it is not possible to draw conclusions about the importance of tcdC in the context of toxin yield or virulence . By contrast , the isogenic tcdC strains studied here clearly show that TcdC is a negative regulator of toxin production since the tcdC complemented BI/NAP1/027 C . difficile strain produced significantly less toxin A and B than the non-complemented control strains . The finding that TcdC-status is not correlated with toxin production in clinical isolates highlights the limitation of accurately assigning gene function by studying non-isogenic strains , particularly in a highly heterogeneous species such as C . difficile . In this context , it might be of interest to study the function of tcdC in isogenic strains generated in a different genetic background such as a ribotype 078 isolate . Analysis of PaLoc gene expression by qRT-PCR demonstrated that TcdC exerts regulatory control of toxin production at the transcriptional level , and this is in keeping with its proposed role as an anti-sigma factor [20] . The observation that the expression of tcdR and tcdE is reduced in the tcdC-complemented strain , together with tcdA and tcdB , is probably because of autoregulation of tcdR since TcdR upregulates its own expression and that of the other PaLoc genes [23] . The virulence of strain M7404 was reduced upon complementation of tcdC , clearly demonstrating that the tcdC mutation in BI/NAP1/027 strains has a significant impact on virulence and is likely to be an important factor in the development of hypervirulence by these strains . Surprisingly , the virulence of M7404 ( tcdC+ ) was found to be equivalent to that of strain 630 , which has been shown in other studies to be reduced in virulence in comparison to other isolates , including three other BI-type strains [32] . These findings have important implications for C . difficile virulence since they suggest that strains carrying tcdC mutations have the inherent potential to develop hypervirulence . The recent emergence of a new class of hypervirulent strains , ribotype 078 [44] , may be one such example . These isolates encode a non-functional TcdC protein [37] , produce significantly more toxin than non-epidemic strains , are associated with more severe disease as well as higher rates of mortality and are increasingly being identified as the causative agent of CDI [44] , [45] . Although these experiments show that TcdC-status alone can modulate virulence it is probable that multiple factors working synergistically are necessary for the development of hypervirulence in the BI/NAP1/027 strains . It is likely that the accumulation of multiple genetic changes in addition to the tcdC mutation has enabled BI/NAP1/027 strains to become the predominant disease-causing isolates in numerous countries . Of particular importance might be variations in the functional activity of the encoded toxins since these isolates were recently shown to produce a toxin B that shows variation across the C-terminal receptor binding domain of the protein [46] , resulting in more potent activity across a wider range of cell lines in comparison to toxin B from the historical , non-epidemic strain 630 [4] . Furthermore , using the zebrafish embryo model of intoxication , the BI/NAP1/027 toxin B was recently shown to have pronounced in vivo cytotoxic activity in comparison to toxin B from VPI10463 , another historical non-epidemic isolate , with greater tissue tropism and more extensive tissue destruction observed [47] . Since toxin B is thought to be one of the major virulence factors of C . difficile [15] , [16] these observations suggest that TcdB variations might play an important role in the hypervirulent phenotype . There are other factors that may influence C . difficile hypervirulence . The BI/NAP1/027 strains encode an additional toxin known as binary toxin or CDT [2] . The role of this toxin in CDI remains to be elucidated but a recent study showed that CDT induces the formation of microtubule-based protrusions on the host cell surface thereby increasing C . difficile adherence to epithelial cells . Moreover , intestinal colonisation of gnotobiotic mice with a BI/NAP1/027 C . difficile strain was significantly reduced in mice treated with CDT-neutralising antibodies in comparison to control mice [48] . These findings suggest that CDT may be an important colonisation factor , enhancing the ability of BI/NAP1/027 strains to initiate infection as well as causing adjunctive tissue damage during later stages of infection , potentially leading to more severe disease . Many BI/NAP1/027 strains are also more proficient at sporulation than non-epidemic C . difficile strains [12] , [49] . C . difficile spores are highly infectious [50] and play a critical role in the transmission of CDI and perhaps in disease relapse , which is a serious problem in patients with CDI [51] . In this context , enhanced sporulation is ostensibly an important adaptation by BI/NAP1/027 isolates , which would result in larger numbers of spores being shed from infected patients and an increased environmental spore load , ultimately leading to higher transmission rates . Finally , the development of fluoroquinolone resistance , in particular to moxifloxacin and gatifloxacin , is unquestionably a major factor in epidemics caused by BI/NAP1/027 strains [3] , [11] . In this regard , the hypothetical co-evolution of enhanced virulence traits and antibiotic resistance in C . difficile mirrors trends seen with other significant nosocomial pathogens such as Methicillin Resistant Staphylococcus aureus ( MRSA ) [52] and Vancomycin Resistant Enterococcus ( VRE ) species [53] . In summary , it is clear that our findings represent an important breakthrough in our understanding of the development of hypervirulence in prevailing C . difficile isolates and will provide a significant reference point for future studies on epidemic strains and their control .
This study was carried out in strict accordance with the recommendations in the United Kingdoms Home Office Animals ( Scientific Procedures ) Act of 1986 which outlines the regulation of the use of laboratory animals for the use of animals in scientific procedures . The experiments described were subject to approval by the University of Glasgow Ethics Committee and by a designated Home Office Inspector ( Project Number 60/4218 ) . All experiments were subject to the 3 R consideration ( refine , reduce and replace ) and all efforts were made to minimize suffering . The characteristics and origins of all recombinant strains and plasmids are shown in Table 1 and Table S2 , respectively . All bacteriological culture media were obtained from Oxoid . C . difficile strains were cultured in BHIS [54] or TY medium [15] , unless otherwise stated , in an atmosphere of 10% H2 , 10% CO2 , and 80% N2 at 37°C in a Coy anaerobic chamber . Escherichia coli was cultured in 2×YT medium aerobically at 37°C , with shaking for broth cultures . All antibiotics were purchased from Sigma-Aldrich and were used at the following concentrations: cycloserine ( Cs , 250 µg/ml ) , cefoxitin ( Cf , 8 µg/ml ) , thiamphenicol ( Tm , 10 µg/ml ) or tetracycline ( Tc , 10 µg/ml ) , chloramphenicol ( Cm , 25 µg/ml ) . Plasmid DNA was isolated using a QIAprep spin miniprep kit ( Qiagen ) . Genomic DNA was prepared using a DNeasy tissue kit ( Qiagen ) . Standard methods for the digestion , modification , ligation , and analysis of plasmid and genomic DNA were used [55] . Nucleotide sequence analysis was carried out using a PRISM BigDye Terminator cycle sequencing kit ( Applied Biosystems ) and detection was performed by Micromon at Monash University . Oligonucleotide primer sequences are listed below . Unless otherwise stated , all PCR experiments were carried out with Phusion DNA polymerase ( New England Biolabs ) and the 2× Failsafe PCR buffer E ( Epicentre ) according to the manufacturer's instructions . For construction of the Tn916 transferrable clostridial shuttle vector , PCR was performed using primers DLP33 ( 5′-GAATTCGCCCTTTTTTATACTCCCCTTG-3′ ) and DLP34 ( 5′-GAATTCGCCCTCAAAGGACGAATATGTCGC-3′ ) and chromosomal DNA extracted from Clostridium perfringens strain JIR4225 [30] . The resulting 700 bp DNA fragment , which contained the oriT region of Tn916 , was TOPO-cloned into pCR-Blunt II-TOPO according to the manufacturer's instructions ( Invitrogen ) . The fragment was then excised from pCR-Blunt II-TOPO using EcoRI and cloned into the equivalent sites of plasmid pMTL9361Cm [29] , resulting in plasmid pDLL4 . For construction of the tcdC-carrying plasmid , PCR was performed using primers DLP35 ( 5′-CTGCAGCCACCTCTAAATCACTGAGTCACTTAATTAC-3′ ) and DLP36 ( 5′-CTGCAGAGCCTTGTAACTGTTTATTTGC-3′ ) and C . difficile strain VPI10463 genomic DNA in order to amplify a 1085 bp fragment encompassing the tcdC gene and upstream region . This fragment was then TOPO-cloned into pCR-Blunt II-TOPO , before being excised with PstI and subcloned into the equivalent site of plasmid pDLL4 , resulting in the final construct pDLL17 . The conjugation procedure utilising E . coli HB101 ( pVS520 ) as the conjugative donor was carried out as previously described [29] . Recombinant plasmids were introduced into C . perfringens strain JIR4225 as before [56] . Conjugations utilising C . perfringens JIR4225 were then performed as follows: separate 90 ml BHIS broth cultures were inoculated with 1 ml aliquots from an overnight C . difficile recipient strain or C . perfringens donor strain starter culture and grown to mid-exponential phase . Approximately 1 ml was removed from each culture , mixed and centrifuged . The cell pellet was then resuspended in phosphate-buffered saline ( PBS ) , spread onto a BHIS agar plate and incubated overnight at 37°C . Bacterial growth was harvested in sterile PBS before being spread onto BHIS agar supplemented with thiamphenicol and incubated overnight as before . Bacterial growth was again harvested with PBS and dilutions spread onto BHIS agar supplemented with cefoxitin and thiamphenicol or tetracycline , and the plates incubated under anaerobic conditions for 24 to 72 h . The toxins were partially purified by ammonium sulphate precipitation from culture supernatants harvested after growth for 72 hours and toxin A was then detected by Western blotting as described previously [15] . For non-quantitative TcdC-specific Western Blots , crude extracts of C . difficile were prepared by sonication of samples taken from cultures that had been grown for 12 hours under anaerobic conditions . The crude extracts from each strain were then subjected to electrophoresis in a 15% SDS-PAGE gel and transferred to a nitrocellulose membrane using standard methods [55] . Membranes were treated with anti-TcdC antibody [57] and detected following treatment with goat anti-mouse IgG-alkaline phosphatase conjugated secondary antibody using standard procedures . For quantitative TcdC-specific Western blots , cultures of C . difficile VPI10463 and C . difficile M7404 ( tcdC+ ) were grown under anaerobic conditions and samples were removed every 4 hours for 24 or 28 hours respectively . Each sample was normalized to an optical density ( 600 nm ) of 0 . 9 prior to lysis , to ensure that the same number of cells was present . Lysates were then prepared by sonication prior to SDS-PAGE gel electrophoresis , transfer and detection , as described above . Following detection , the amount of TcdC in each lysate was quantified by densitometric analysis , using purified recombinant TcdC protein ( rTcdC ) as the standard and the ImageJ software package , according to published methods [58] . Toxin B was detected in C . difficile culture supernatants harvested after growth for 72 hours by Vero cell cytotoxicity assays as described previously [15] , except that each well was seeded with 1×105 cells . Total RNA was extracted from C . difficile cultures grown for 12 h in TY media . Reverse transcription was performed using AMV Reverse Transcriptase ( Promega ) using random hexamer oligonucleotides primers and 2 µg template RNA . The cDNA samples were then purified using Qiaquick Columns ( Qiagen ) . PaLoc gene specific primers were designed using Primer 3 software ( Geneious Software ) . qRT-PCR was performed using an AB7300 real-time PCR instrument ( Applied Biosystems ) . Reactions were carried out using the FastStart Universal SYBR Green Master Mix ( Roche ) with 40 ng of cDNA as template . Standard curves were generated for each primer pair using C . difficile genomic DNA , and melt curve analysis was performed following each qRT-PCR reaction to verify amplification specificity . Samples were normalised using the C . difficile rrnA gene . Spores were prepared from C . difficile cultures grown in 500 ml of BHI broth . Cultures were pelleted by centrifugation for 10 mins and re-suspended in 50% ethanol . The material was then vortexed every 10 min for 1 h before centrifugation for 10 mins . The pellet was then treated with 1% Sarkosyl in PBS for 1 h at room temperature and again pelleted by centrifugation , followed by incubation overnight at 37°C with lysozyme ( 10 mg/ml ) in 125 mM Tris-HCl buffer ( pH 8 . 0 ) . The sample was treated in a sonicating water bath ( 3 pulses of 3 min each; 1510 Branson ) before centrifugation through a 50% sucrose gradient for 20 mins . The pellet was incubated in 2 ml of PBS containing 200 mM EDTA , 300 ng/ml proteinase K and 1% Sarkosyl for 30 mins at 37°C before centrifugation through a 50% sucrose gradient for 20 mins . The final pellet was then washed twice in sterile distilled water before finally being resuspended in 1 ml of sterile water . Spore preparations were stored at −80°C prior to use . Female Golden Syrian hamsters purchased from Harlan Olac UK were used for all animal experiments . Telemetry chips ( Vitalview Emitter ) were inserted by laparotomy into the body cavity of the animals at least 3 weeks before infection with C . difficile . Animal experiments were then carried out as described previously [33] , except that animals received 1×104 spores of C . difficile . Animals were culled when core body temperature dropped below 35°C . This study was carried out in strict accordance with the recommendations in the United Kingdoms Home Office Animals ( Scientific Procedures ) Act of 1986 which outlines the regulation of the use of laboratory animals for the use of animals in scientific procedures . The experiments described were subject to approval by the University of Glasgow Ethics Committee and by a designated Home Office Inspector ( Project Number 60/4218 ) . All experiments were subject to the 3 R consideration ( refine , reduce and replace ) and all efforts were made to minimize suffering . To estimate colonisation , hamsters were sacrificed and the gut region from the caecum to the anus removed . The tissues were homogenised in PBS using a Stomacher and viable counts were performed on the homogenate as described previously [33] . To confirm that the bacteria isolated from the hamster were the same strain as originally used for infection , genomic DNA was isolated and subjected to MVLA as described previously [59] . Plasmid rescue was performed as previously described [29] followed by restriction digest analysis to confirm plasmid integrity .
|
Hospital infections are increasingly being recognised as a major patient safety issue with the hospital environment providing a niche for the rapid evolution of microbial pathogens that are well adapted to infecting susceptible patients . The spore-forming Clostridium difficile is one such bacterium , which causes disease in patients undergoing antibiotic therapy . Since 2000 , there has been a striking increase in C . difficile infections due to the emergence of hypervirulent isolates that appear to possess extended antibiotic resistance and virulence properties . Here we use a genetic approach to identify a factor linked to the development of hypervirulence in C . difficile . This study shows that a naturally occurring mutation in a negative regulator of toxin production , the anti-sigma factor TcdC , is an important factor contributing to the development of hypervirulence in epidemic isolates , presumably because it leads to significantly increased toxin production . These results have important implications for C . difficile pathogenesis since they suggest that strains carrying a similar mutation have the inherent potential to develop a hypervirulent phenotype . This study has increased our understanding of how these new variant strains cause disease and why they are more harmful , which is critical for the development of improved strategies for preventing and treating these infections .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"biology"
] |
2011
|
The Anti-Sigma Factor TcdC Modulates Hypervirulence in an Epidemic BI/NAP1/027 Clinical Isolate of Clostridium difficile
|
The mechanisms by which adaptive phenotypes spread within an evolving population after their emergence are understood fairly well . Much less is known about the factors that influence the evolutionary accessibility of such phenotypes , a pre-requisite for their emergence in a population . Here , we investigate the influence of environmental quality on the accessibility of adaptive phenotypes of Escherichia coli's central metabolic network . We used an established flux-balance model of metabolism as the basis for a genotype-phenotype map ( GPM ) . We quantified the effects of seven qualitatively different environments ( corresponding to both carbohydrate and gluconeogenic metabolic substrates ) on the structure of this GPM . We found that the GPM has a more rugged structure in qualitatively poorer environments , suggesting that adaptive phenotypes could be intrinsically less accessible in such environments . Nevertheless , on average ∼74% of the genotype can be altered by neutral drift , in the environment where the GPM is most rugged; this could allow evolving populations to circumvent such ruggedness . Furthermore , we found that the normalized mutual information ( NMI ) of genotype differences relative to phenotype differences , which measures the GPM's capacity to transmit information about phenotype differences , is positively correlated with ( simulation-based ) estimates of the accessibility of adaptive phenotypes in different environments . These results are consistent with the predictions of a simple analytic theory that makes explicit the relationship between the NMI and the speed of adaptation . The results suggest an intuitive information-theoretic principle for evolutionary adaptation; adaptation could be faster in environments where the GPM has a greater capacity to transmit information about phenotype differences . More generally , our results provide insight into fundamental environment-specific differences in the accessibility of adaptive phenotypes , and they suggest opportunities for research at the interface between information theory and evolutionary biology .
During adaptation , a population “moves” in genotype space in search of genotypes associated with high-fitness phenotypes . The success of adaptation depends crucially on the accessibility of such adaptive phenotypes . While adaptive phenotypes rely on natural selection for their fixation , their accessibility depends , primarily , on the structure of the genotype to phenotype mapping ( GPM ) and , secondarily , on the forces that move a population in genotype space – i . e . selection and genetic drift ( see Materials and Methods for relevant definitions ) . In particular , accessible phenotypes must be linked by a path of viable phenotypes to the initial phenotype of a population . In addition , the structure of the GPM determines the dominant mechanism by which a population moves in genotype space; for a smooth GPM that contains extensive neutral networks of genotypes associated with individual adaptive phenotypes , the motion may occur predominantly by genetic drift , with selection acting only occasionally to move a population from one neutral network onto another [1]–[3] . On the other hand , for very rugged GPMs having a low degree of neutrality , movement in genotype space may be mostly mediated by selection . By studying the factors that influence biologically relevant GPMs , we may gain insight into the accessibility of adaptive phenotypes . To that end , we have taken advantage of recent advances in the understanding of bacterial metabolic networks [4]–[8] to investigate the influence of environmental quality on the structure of E . coli's central metabolic network GPM [9] , [10] ( see Table S1 ) . We used the latest gene-protein reaction-associations data on the metabolic network [10] to identify all the genes involved in the network's central metabolic pathways ( i . e . , respiration , the tricarboxylic acid cycle , glycogen/gluconeogenesis , pyruvate metabolism , the pentose shunt ) . We found 166 such genes ( see Table S1 ) , and defined the network's genome to be an ordered list of these genes . Mutations to a given gene are allowed to change the gene's state from “on” to “off” ( deleterious mutations ) and vice-versa ( compensatory mutations ) . A genotype is defined as a particular configuration of on-off states of the 166 genes that make up the genome . The Hamming distance between any two genotypes is the number of differences in the states of corresponding genes . We define a genotype's phenotype ( equivalent , for our purposes , to fitness ) using a model of metabolic flux . Specifically , a growing body of experimental and theoretical work [7] , [11]–[15] suggests that under conditions of carbon limitation , E . coli ( and other bacteria ) organize their metabolic fluxes so as to optimize the production of biomass , and that experimentally realized optimal biomass yields can be predicted with reasonable accuracy by the mathematical method of flux-balance analysis [4]–[7] . Therefore , we used the optimal biomass yield predicted by means of flux-balance analysis as a biologically grounded proxy for the phenotype/fitness of a particular genotype of the metabolic network ( Further details on the definition of the metabolic network GPM are given in Materials and Methods ) . We then used statistical and information-theoretic methods to investigate the structure of the GPM under conditions in which one of seven compounds ( henceforth called “environments” ) served as the primary metabolic substrate . Note that an advantage to studying E . coli's central metabolic network is that its GPM is systemic ( as are organismal GPMs ) , and it has a very rich structure; the network phenotype is an emergent property of interactions among gene products and between these products and the intracellular and extracellular environments of the bacterium . The interaction rules are numerous ( i . e . , of the order of the number of genes ) and , in some cases , complex ( see Figure 1 ) .
We found that in all environments ( except acetate ) large genotype changes ( >∼30 ) induce phenotype differences that follow an interesting bi-modal distribution . This bi-modal distribution is characteristic of the expected distribution of phenotype differences between randomly sampled genotypes , suggesting that in the considered environments the E . coli metabolic network maps onto two dominant clusters of similar metabolic phenotypes . In acetate , the poorest environment , the distribution of phenotype differences induced by large genotype changes was essentially uni-modal , suggesting the existence of only one dominant cluster of similar metabolic phenotypes . The CL was shorter in poorer environments , suggesting that the GPM could have a more rugged structure in such environments and , hence , it may be intrinsically more difficult to find adaptive phenotypes . Note that in poorer environments there may be fewer possibilities of re-routing fluxes through the metabolic network in order to maintain biomass yields following gene deletion; this could account for the faster decay of the correlation between biomass yields attained before and after gene deletions and , hence , the lower CLs computed in these environments . In spite of the predicted ruggedness of the GPM in acetate , the poorest of the considered environments , very long ( ∼74% of the genotype length ) neutral walks could still be performed on the GPM , suggesting that neutral drift can alter a substantial fraction of the phenotype during evolution . In other words , a population evolving in acetate could explore large portions of genotype space by drifting on neutral networks , increasing its likelihood of discovering adaptive phenotypes . Furthermore , the NMI was largest in acetate and smallest in lactose , suggesting that the information-transmission capacity of the GPM does not necessarily increase in better environments . In order to gain further intuition about how qualitative changes to the environment could influence the dynamics of adaptation , we simulated the adaption of E . coli populations to qualitatively different environments . The speed of adaptation to a given environment was positively correlated with the NMI associated with that environment; adaptation appeared to increase with the GPM's capacity to transmit information about phenotype differences under given environmental conditions . In contrast , the relative speed of adaptation to different environments was inconsistent with expectations based on the environment-specific CLs . This suggests that the CL , and the degree of ruggedness of the GPM that it measures , may not capture enough information about features of the GPM that influence the speed of adaptation . The above results were found to be consistent with the predictions of a mathematical theory that , under the assumptions of Fisher's fundamental theorem of natural selection ( e . g . , see [22] ) , demonstrated the existence of a positive correlation between the NMI and the rate of fitness increase ( i . e . , the speed of adaptation ) . Together , the above results suggest that environmental quality could have a fundamental influence on the outcome of adaptation . Note that previous work [24] , [25] showed that in a changing environment , the speed of adaptation may increase with the mutation rate and also with the propensity of point mutations to have phenotypic effects . The mathematical theory presented here provides a complementary perspective: if both the environment and the mutation rate are fixed , then the speed of adaptation may increase with the amount of information that genetic variation provides about phenotypic variation and , due to the symmetry of the mutual information , with the amount of information that phenotypic variation provides about underlying genetic variation . This suggests an intriguing connection between the “predictability” of the genetic basis of fitness increases of a particular magnitude and the rate at which such increases occur . Also , note that the NMI is , in essence , a measure of adaptation potential ( or evolvability ) . It is applicable to a wider range of data types ( both numeric and symbolic data types ) than related measures of evolvability used in quantitative genetics , such as the ( “narrow-sense” ) heritability of phenotype [26] , defined as the ratio of the additive genetic variance in phenotype to the total variance in phenotype . We conclude by pointing out some limitations of our empirical GPM model , and we discuss possible directions for future work . Firstly , our approach to analyzing E . coli's metabolic network GPM did not take into account transcriptional regulation , which has been shown [27] , [28] to mediate dynamic microbial responses to environmental perturbations . By accounting for transcriptional regulation we would endow the metabolic network with a much richer structure that may provide additional information about the evolutionary accessibility of the network's adaptive phenotypes . Nevertheless , there is ample experimental evidence [7] , [10]–[15] that the model underlying our approach to analyzing the network is sufficient for predicting bacterial growth phenotypes in various environments . Secondly , we only considered genotype changes that turn a gene either off ( i . e . , deleterious changes to the gene or to its associated transcription factor ) or on ( as could happen when compensatory changes occur ) . This was motivated by practical considerations: computational prediction of graduated phenotypic consequences of genotype changes is currently not feasible on a genomic scale . Future improvements in our ability to make such predictions will allow for better modeling of metabolic network GPMs . The GPM model we studied will add to the suite of available models ( e . g . , see [25] , [29] , [30] ) that have enabled the investigation of important questions in evolutionary biology . In addition , the insights we presented could contribute to the understanding of evolutionary processes at both the molecular and population levels . At the molecular level , the NMI could be useful for understanding the evolvability of proteins . For example , one expects the nucleotide sequences of proteins that are particularly important for the adaptation of a pathogen to the immune response of its host ( e . g . , the hemagglutinin protein of influenza viruses ) to occupy regions of genotype space associated with NMI values that are significantly greater than random expectations . To test this hypothesis , the NMI of the nucleotide sequence variation in a population sample of nucleotide sequences of a pathogen's protein relative to the corresponding amino acid sequence/protein structure variation can be computed . The computed NMI can be subsequently compared to the distribution of NMIs obtained from appropriately randomized ( e . g . , see [21] , [31] ) versions of the original sample of nucleotide sequences to determine its statistical significance . In addition , since the NMI affords an analytically tractable measure of evolvability , it could be useful to the mathematical investigation of the evolutionarily important relationship between evolvability and robustness ( e . g . , see [32] ) . Of particular interest is the derivation of a broadly applicable mathematical description of this relationship . Previous simulation studies of the RNA GPM ( e . g . , see [1] , [33] ) showed that evolvability can increase with the robustness of RNA structures to nucleotide changes . In contrast , a recent simulation study of GPMs generated by a model gene network showed that the fraction of phenotypically consequential point mutations to a genotype of the network , which is inversely correlated with the network's robustness , increased with evolvability , during adaptation to a changing environment [25] . It is not clear whether these conflicting results can be obtained from different instantiations of the same mathematical model or whether they are fundamentally irreconcilable . Additional insight could come from mathematical investigations of simple model GPMs ( e . g . , see [30] ) using analytically tractable measures of evolvability ( e . g . , the NMI ) and robustness ( e . g . , the CL; see [18] ) . These investigations could yield important insights into the possible existence of general mathematical rules underlying the relationship between evolvability and robustness .
|
Adaptation involves the discovery by mutation and spread through populations of traits ( or “phenotypes” ) that have high fitness under prevailing environmental conditions . While the spread of adaptive phenotypes through populations is mediated by natural selection , the likelihood of their discovery by mutation depends primarily on the relationship between genetic information and phenotypes ( the genotype-phenotype mapping , or GPM ) . Elucidating the factors that influence the structure of the GPM is therefore critical to understanding the adaptation process . We investigated the influence of environmental quality on GPM structure for a well-studied model of Escherichia coli's metabolism . Our results suggest that the GPM is more rugged in qualitatively poorer environments and , therefore , the discovery of adaptive phenotypes may be intrinsically less likely in such environments . Nevertheless , we found that the GPM contains large neutral networks in all studied environments , suggesting that populations adapting to these environments could circumvent the frequent “hill descents” that would otherwise be required by a rugged GPM . Moreover , we demonstrated that adaptation proceeds faster in environments for which the GPM transmits information about phenotype differences more efficiently , providing a connection between information theory and evolutionary theory . These results have implications for understanding constraints on adaptation in nature .
|
[
"Abstract",
"Introduction",
"Discussion"
] |
[
"computational",
"biology/evolutionary",
"modeling",
"computational",
"biology/metabolic",
"networks",
"computational",
"biology/systems",
"biology"
] |
2009
|
On the Accessibility of Adaptive Phenotypes of a Bacterial Metabolic Network
|
Endoglin is an auxiliary receptor for members of the TGF-β superfamily and plays an important role in the homeostasis of the vessel wall . Mutations in endoglin gene ( ENG ) or in the closely related TGF-β receptor type I ACVRL1/ALK1 are responsible for a rare dominant vascular dysplasia , the Hereditary Hemorrhagic Telangiectasia ( HHT ) , or Rendu-Osler-Weber syndrome . Endoglin is also expressed in human macrophages , but its role in macrophage function remains unknown . In this work , we show that endoglin expression is triggered during the monocyte-macrophage differentiation process , both in vitro and during the in vivo differentiation of blood monocytes recruited to foci of inflammation in wild-type C57BL/6 mice . To analyze the role of endoglin in macrophages in vivo , an endoglin myeloid lineage specific knock-out mouse line ( Engfl/flLysMCre ) was generated . These mice show a predisposition to develop spontaneous infections by opportunistic bacteria . Engfl/flLysMCre mice also display increased survival following LPS-induced peritonitis , suggesting a delayed immune response . Phagocytic activity is impaired in peritoneal macrophages , altering one of the main functions of macrophages which contributes to the initiation of the immune response . We also observed altered expression of TGF-β1 target genes in endoglin deficient peritoneal macrophages . Overall , the altered immune activity of endoglin deficient macrophages could help to explain the higher rate of infectious diseases seen in HHT1 patients .
Endoglin was originally described as a type I integral membrane protein with an extracellular domain of 561 amino acids , a hydrophobic transmembrane domain , and a 47-residue cytosolic domain [1] . It is mainly expressed in endothelial cells and plays a pivotal role modulating cellular responses to TGF-β [1 , 2 , 3] . Mice lacking endoglin die at E10 . 5-E11 . 5 from angiogenic and cardiovascular defects [4 , 5 , 6] . Mutations in ENG are responsible for the Hereditary Hemorrhagic Telangiectasia type 1 ( HHT1 ) [7] . HHT , or Rendu-Osler-Weber syndrome , is a rare disease with a prevalence of 1/5 , 000 to 1/8 , 000 and is an autosomal dominant disorder characterized by multisystemic vascular dysplasia and recurrent hemorrhages [8] . In endothelial cells , endoglin promotes a stimulatory effect mediated by TβRII/ALK1 signaling , and inhibitory signals transduced by TβRII/ALK5 signaling complexes [3 , 9 , 10 , 11] . Decreased endoglin expression in endothelial cells from HHT1 donors leads to impaired TGF-β signaling , a disorganized cytoskeleton , and failure to form vascular cord-like structures in vitro [12] . In addition , several reports have shown a role for endoglin during hematopoiesis and myeloid lineage development . Endoglin regulates hematopoiesis by modulating the TGF-β signaling pathway in early development [13] . Moreover , in the absence of endoglin , myelopoiesis and definitive erythropoiesis are severely impaired . In contrast , lymphopoiesis appears to be only mildly affected [14] . Furthermore , in endoglin knock-out embryos , hematopoietic colony activity and numbers of erythroid progenitors are severely reduced [15] . TGF-β is a regulatory cytokine with a pivotal role in regulating immune responses [16] . On monocyte/macrophage ( Mo/MΦ ) cell populations , the action of TGF-β appears to depend on the differentiation stage of the cells . Generally , TGF-β stimulates cells in the resting state ( Mo ) , whereas activated cells ( MΦ ) are inhibited by it [17] . The role of endoglin on MΦ and immune cell functions remains unknown , although endoglin was identified in differentiated MΦ from human peripheral blood Mo more than 20 years ago [18 , 19] . Furthermore , as an extension of this , Sanz-Rodríguez and colleagues [20] described that up-regulation of endoglin during the differentiation of peripheral blood Mo in culture , is age-dependent and impaired in Mo from HHT patients , a fact that could be related to a high frequency of infectious diseases observed in HHT patients [21 , 22] . Moreover , endoglin expression in human macrophages is important for blood cell-mediated vascular repair [23] . In addition , colitic Eng+/- mice show an impaired resolution of inflammation characterized by increased macrophage and neutrophil infiltration , but with a reduction in expression of NAPH oxidase 2 ( Nox-2 ) and myeloperoxidase , two key phagocytic respiratory enzymes [24] . These findings suggest that endoglin is required for fully functional myeloid cells and prompted us to develop a mouse model with endoglin depletion in the myeloid lineage to analyze the role of endoglin in macrophages . Active TGF-β1 binds to TβRII/ALK5 receptor complex and exerts inhibitory signals for immune cells . The presence of endoglin in MΦ and its capacity to modulate TGF-β signal via ALK1/Smad1/5/8 opens a new door for TGF-β signaling in MΦ . Endoglin expression in human macrophages is well characterized [18 , 19 , 25] , but its expression on murine macrophages is not well established . In the present report we describe endoglin up-regulation during in vitro transition of murine peripheral blood Mo towards MΦ , and evaluate endoglin expression on tissue-resident MΦ isolated from liver and peritoneal cavity ( PerC ) of wild-type C57BL/6 mice . Following the identification of endoglin expression in MΦ , we examined its role in the innate immune system by disrupting endoglin expression in the myeloid lineage in vivo . Several HHT mouse models have been developed to investigate the underlying mechanisms leading to vascular malformations in HHT , mainly focusing on the role of endoglin in endothelial cells [26] . Here we present an HHT1 mouse model to dissect the function of endoglin in MΦ , and consequently to elucidate its role in the innate immune response . Mice lacking endoglin expression in MΦ exhibit increased susceptibility to spontaneous infections , an impairment of phagocytic activity , and an aberrant leukocyte recruitment to the site of infection . Endoglin deletion also results in decreased levels of the pro-inflammatory cytokines TNF-α , IL-1β and IL-6 , which correlated with a weaker septic response following lipopolysaccharide ( LPS ) injection . Altogether , these results suggest that endoglin is involved in the regulation of the innate immune response and provide , for the first time , evidence for its role in TGF-β signaling in MΦ in vivo . Furthermore , this impairment of the innate immune response seen when endoglin is absent from MΦ may help to explain the high frequency of infectious diseases observed in HHT patients [21 , 22] .
To follow endoglin expression during in vitro differentiation of cultured mouse Mo , flow cytometry analyses of peripheral blood Mo from wild-type mice were carried out . FS vs SS analysis of Peripheral Blood Leukocytes ( PBLs ) cultured for 3h allowed the identification of Mo as Ly6Gneg CD11bhigh CD3neg population . Three hours after culture , Mo are negative for endoglin expression . Twenty-four hours after culture , 22% of Mo are positive for endoglin expression and there is a gradual increase in the percentage of endoglin positive cells reaching a plateau between 4 and 7 days of 85% ( Fig 1A and 1B ) . In parallel experiments , MΦ were in vitro differentiated for 7 days from bone marrow precursors , with GM-CSF or M-CSF to polarize MΦ towards M1 and M2 phenotypes , respectively [27 , 28] . Quantitative PCR and flow cytometry analysis showed expression of endoglin in bone marrow derived macrophages ( BMDM ) , both M1 and M2 subtypes from 3 different wild-type mice ( Fig 1C and 1D ) . These data also suggest that endoglin is associated with in vitro differentiation of murine MΦ . To examine endoglin expression in vivo , tissue-resident MΦ were isolated as single cell suspensions from liver and PerC . Cells were analyzed by flow cytometry and endoglin expression was measured on the Ly6Gneg F4/80pos cell subpopulation ( Fig 2 ) . F4/80 , widely used as a mouse MΦ marker [29 , 30] , is highly expressed on Kupffer cells and resident PerC MΦ , but is weakly expressed in other resident MΦ , as alveolar MΦ or even absent , as in marginal and white pulp splenic MΦ [29 , 31] . In the comparative study shown here , F4/80 expression levels are higher in PerC MΦ than in F4/80pos cells present in the liver cellular suspension , likely Kupffer cells . Flow cytometric analysis showed that , PerC MΦ as well as putative Kupffer cells , are positive for endoglin expression . Remarkably , F4/80pos cells from liver express more endoglin than resting PerC macrophages ( Fig 2 ) . Next , we analyzed in vivo regulation of endoglin expression in the Mo-MΦ differentiation process . Peritonitis was induced by injecting Zymosan A in the peritoneal cavity , the so called Zymosan induced peritonitis or ZIP model in order to investigate endoglin expression in MΦ derived from peripheral blood Mo and recruited to the PerC . Cells from unstimulated mice ( steady-state ) were used as controls . Endoglin expression was measured on PerC MΦ defined as Ly6Gneg F4/80pos , easily distinguishable from the Ly6Gpos F4/80neg granulocyte population ( Fig 3A ) . In the PerC , two subpopulations of MΦ have been identified and are defined as Small Peritoneal Macrophages ( SPMs ) and Large Peritoneal Macrophages ( LPMs ) , due to size differences [31] . They are also characterized by their different F4/80 expression levels: LPMs are Ly6Gneg F4/80high while SPMs are Ly6Gneg F4/80low . In steady-state condition , we observed a main peak of F4/80high expression levels , suggesting that PerC MΦ is highly enriched for LPM , as previously reported for other mouse strains [31] . Until forty-eight hours after 1 mg of Zymosan injection , the predominant PerC MΦ subset is that of SPM , as previously described [32] . As is shown in Fig 3A , twelve hours after ZIP , the main peak of F4/80pos cells corresponds to SPM , and no LPM is detected . Granulocytes do not express endoglin , while endoglin expression in MΦ changes over time ( Fig 3B ) . The first MΦ differentiated from peripheral blood Mo and identified as F4/80pos cells ( SPM ) , are negative for endoglin expression . Indeed , endoglin expression remains undetectable until day 3 . At this time point endoglin expression begins to increase after ZIP . Remarkably , there is more endoglin on the brighter F4/80pos cells , likely the LPM macrophage subset , while granulocytes do not express endoglin at all . Thus , endoglin expression is induced during the in vivo differentiation process of Mo towards MΦs . These data indicate that the endoglin up-regulation during the Mo to MΦ differentiation process occurs in vivo , albeit over a slower time course than observed in in vitro assays . The new LPM population is easily distinguishable 2 weeks after ZIP , when the inflammatory response is resolved and the transmigrated Mo are completely differentiated to MΦ and replenish the PerC . In order to investigate the role of endoglin in MΦ function either in resident tissue MΦ or in activated Mo/MΦ in vivo , endoglin specific myeloid lineage knock-out mice were obtained by crossing a strain expressing Cre recombinase from the lysozyme 2 gene ( Lyz2 ) , the LysMCre strain , with a floxed endoglin strain containing loxP sites flanking exons 5 & 6 , the Engfl/fl strain . The scheme of crosses is shown in S1 Fig . The mice were all of the same genetic background ( C57BL/6 ) and were genotyped by genomic PCR using DNA isolated from tail tissue ( S1B and S1C Fig ) . Activity of Cre recombinase was confirmed in MΦ from liver , lung , spleen , heart and peritoneal cavity by the detection of the EngΔ5–6 allele by genomic PCR in Engwt/flLysMCre and Engfl/flLysMCre mice . The specificity of Cre recombinase under the control of the Lyz2 promoter was checked by immunohistochemistry in liver . The Lyz2 promoter is specific to the myeloid lineage [30] , so it is expected that endoglin expression will be maintained in other cellular types e . g . endothelial cells . In fact , endoglin expression appears unaltered in hepatic sinusoids and in endothelium of hepatic veins of Engfl/flLysMCre mice ( Fig 4A ) . Moreover , the serum levels of soluble endoglin ( sEng ) , directly related to the levels of endothelial endoglin expression [33] , remain unaltered ( Fig 4B ) . The efficiency of Eng deletion was complete as confirmed by real-time qPCR analysis of peritoneal MΦ . The Eng mRNA levels were almost undetectable in MΦ from Engfl/flLysMCre mice and were intermediate in Engwt/flLysMCre mice compared to control mice ( Fig 4C ) . These results were also confirmed at protein level by flow cytometry ( Fig 4D ) . All the strains were kept in the same room , and under the same breeding conditions . Interestingly , we observed local spontaneous infections in a notable percentage of reproductive individuals ( 32 . 5% of animals ) of the Engfl/flLysMCre genotype , and one of 30 Engwt/flLysMCre males ( Fig 5 ) . The percentage of spontaneous infections reflects the incidence only in adult mice that were interbred to maintain the experimental strains . We only detected one single spontaneous infection in an Engfl/flLysMCre male housed with its littermates . Infections affected both sexes equally , and the most frequent localization was in the abdominal region surrounding the urogenital area . Usually , when one individual developed an infection , as they were housed in the same cage , they transmitted the infection to their breeding partner . Females underwent a normal gestation , but if infected , they usually ate their litter . Also , if the infection in females appeared immediately after set up , they did not become pregnant . Unfortunately , this led to a decrease of breeding efficiency . Necropsis of infected individuals revealed a splenomegaly secondary to infectious processes in all animals with visible infection symptoms ( Fig 5D ) . To characterize the bacterial strains responsible for the infections , samples from infected animals were assessed and analyzed at the microbiological department of Complutense University ( Madrid ) . Several opportunistic bacteria were identified , but infections were mainly due to Staphylococcus aureus ( Fig 5E ) . Therefore , mice with MΦ lacking endoglin show increased susceptibility to infection by opportunistic bacteria . As the previous results suggested an immune-compromised phenotype following endoglin deletion in MΦ , we next assessed the primary immune responses in the three genotypes: endoglin KO , heterozygous and controls , to test the role of endoglin in MΦ during the innate immune response . For this purpose a LPS septic shock was induced and survival of animals was monitored for 5 days . Mice with normal MΦ ( Engwt/wtLysMCre ) were significantly more susceptible to LPS treatment than heterozygous ( Engwt/flLysMCre ) and KO ( Engfl/flLysMCre ) mice . During the first 36h following the LPS injection , animals lacking endoglin in MΦ showed a delayed endotoxin-induced mortality and a higher survival at the 120 hour endpoint ( Fig 6A ) . Animals alive at the endpoint had completely recovered from the septic shock , with healthy and normal appearance and appetite . At this endpoint , no pain signals were observed . Since production of pro-inflammatory cytokines is rapidly activated following LPS injection , levels of TNF-α , IL-1β and IL-6 in PerC and in blood serum , were analyzed at early times 1 & 3 hours post-injection , to compare responses between the three genotypes . The absence of endoglin did not affect TNF-α serum concentrations ( Fig 6B ) , but these were significantly lower in PerC 1h post-LPS injection in Engwt/flLysMCre and Engfl/flLysMCre mice than in control animals ( Fig 6B ) . On the other hand , the increase of IL-1β and IL-6 serum levels , 3 hours after LPS injection , was significantly lower in Engwt/flLysMCre and Engfl/flLysMCre mice compared to controls ( Fig 6C ) . To functionally support the differences in pro-inflammatory cytokines found in PerC among the different genotypes , a lymphocytic cell line , SR . D10-CD4-F1 was used in in vitro migration assays to measure the migratory response to peritoneal exudates from the different genotypes . As can be seen in Fig 6D , peritoneal exudates from mice with endoglin deficiency were significantly less effective in recruiting lymphocytes than exudates from control mice . Because cytokines are chemoattractants that direct leukocytes to sites of inflammation , we next investigated if endoglin expression in MΦ played a role in leukocyte transmigration , in an in vivo model of acute inflammation . To this end , peritonitis was induced by injecting Zymosan A in PerC ( ZIP ) . The number of resident PerC cells in quiescence is unaffected by the presence or absence of endoglin expression in MΦ ( Fig 7A ) , and the total number of leukocytes and subpopulations in peripheral blood are similar between the strains ( Table 1 ) . However , twenty-four hours after Zymosan challenge , Engfl/flLysMCre mice show a significantly lower influx of leukocytes to PerC compared to control mice Engwt/wtLysMCre ( Fig 7B ) . The cell influx is mainly represented by blood granulocytes ( CD11bpos Ly6Gpos ) and blood Mo differentiated to MΦ ( CD11bpos F4/80low ) ( Fig 7C ) . Phagocytosis can be measured by the incorporation of fluorescently labeled Zymosan A particles by resident LPM . When endoglin was reduced or absent , MΦ exhibit deficient phagocytosis of Zymosan particles ( Fig 8 ) . Phagocytic activity ( represented by the percentage of MΦ that have incorporated fluorescent particles; positive for CFSE signal ) and phagocytic efficiency ( represented by the CFSE Mean Fluorescence Intensity ( MFI ) ) in F4/80pos cells were both decreased in MΦ from Engfl/flLysMCre compared to control mice ( Engwt/wtLysMCre ) . In heterozygous mice , the phagocytic efficiency is also clearly decreased compared to control mice . As endoglin is a TGF-β1 co-receptor , we evaluated the expression of selected downstream genes: Acvrl1 , Serpine1 , Id1 , Nos2 , Mmp12 and Inhba in in vitro cultures of control , heterozygous and KO PerC MΦ ( Fig 9 ) . Macrophages were selected by adherence to plastic flasks and cultured in DMEM supplemented with 10% FCS for 24 hours . Gene expression was also checked in untreated cultured macrophages . Endoglin deficiency in MΦ led to a reduced expression of Acvrl1 and Mmp12 genes . We also observed reduced expression of Nos2 and Inhba in endoglin-deficient MΦ . Of note , there was a trend suggesting intermediate loss of target gene expression in MΦs that were heterozygous for endoglin expression with levels of Nos2 and Inhba significantly reduced compared with controls . Serpine1 expression in MΦs from Engfl/flLysMCre mice was significantly increased compared to those of Engwt/flLysMCre or Engwt/wtLysMCre . No statistically significant changes in Id1 expression were detected .
Endoglin is expressed during the in vitro differentiation of human Mo [18 , 19 , 25] , but its expression in murine MΦ has remained elusive and controversial . Some authors could not detect endoglin by immunohistochemistry in murine atherosclerotic plaques [34] and had considered that murine MΦ do not express endoglin . However , more recently , endoglin transcripts in murine MΦ were reported by semi-quantitative RT-PCR [35] . The results reported in the present manuscript , show the expression of endoglin in murine MΦ in three different contexts: ( i ) during in vitro differentiation of Mo , ( ii ) during in vivo transition of peripheral blood Mo to Mɸ following an inflammatory process , and ( iii ) on tissue-resident Mɸ isolated from liver and peritoneal cavity of untreated C57BL/6 mice . Taken together this shows that endoglin is expressed in murine MΦ and that endoglin is also a marker of Mo differentiation towards MΦ , since endoglin is not detected on circulating Mo ( Fig 1A and 1B ) . However , the endoglin levels that we found on differentiated blood Mo to MΦ in mouse are lower ( 22% of murine Mo were endogline positive ) compared to those reported for cultured human Mo , where 94% of differentiated MΦ are positive for endoglin expression after 21h in culture [25] . Endoglin is expressed in terminally differentiated M1/M2 macrophages . In this context , Aristorena and colleagues [35] recently demonstrated that overexpression of S-endoglin in U937 cells ( a human promonocytic cell line ) impairs differentiation to the pro-inflammatory M1 phenotype . S-endoglin counteracts the signaling prompted by L-endoglin via ALK1/Smad1/5 [36] . These data suggest that L-endoglin would be necessary for the complete differentiation , at least to M1 phenotype . On the other hand , we have seen that both M2 and M1 macrophages express endoglin , suggesting that endoglin expression is associated with a complete M2 or M1 phenotype . The expression of endoglin during Mo in vivo differentiation is compatible with the involvement of endoglin in cellular trafficking to the target tissues . Previous studies described that endoglin on endothelial cells is involved in adhesion to the extracellular matrix [37] , and in promoting leukocyte adhesion to vascular endothelium [38] . Both processes are involved in inflammation and leukocyte extravasation , however , the role of MΦ endoglin in innate immunity and inflammation is not yet well established . Constitutive overexpression of endoglin in U937 cells showed the deregulation of hundreds of genes compared to the parental line . These genes are involved in cellular movement , cell adhesion and transmigration [35 , 39] . In the present work , we have followed in vivo endoglin expression in SPM and LPM ( the two macrophage subsets present in PerC ) during leukocyte recruitment induced by a ZIP process . Subsequently to Zymosan challenge , granulocytes are the predominant myeloid population in PerC , and they did not express endoglin . Following Zymosan injection , MΦ seem to disappear from PerC . The “macrophage disappearance reaction” is attributed to their migration to the omentum [40] . Following the granulocytes influx after Zymosan challenge , recruited peripheral blood Mo differentiate to SPM being the predominant MΦ population in PerC . A discrete LPM population is detected 24 hours after ZIP and seems to be derived from SPM . This new LPM population starts to become positive for endoglin expression 3 days after ZIP , as do SPM cells . Thus , we suggest that SPM contribute to the replenishment of LPM after ZIP , similar to other situations where Mo-derived MΦ make substantial contribution to the population of resident MΦ [41] . Moreover , the results shown in this work indicate that endoglin is a marker of differentiated MΦ , suggesting that endoglin could affect early inflammatory events orchestrated by resident MΦ , not from Mo during the process of transmigration , although a transient endoglin up-regulation during this event cannot be ruled out . Nonetheless , caution is required when comparing in vitro and in vivo data , and our results indicate that up-regulation of endoglin during Mo to MΦ differentiation in vivo is much slower than Mo differentiation in vitro . To elucidate the role of endoglin in differentiated MΦ a mouse strain lacking endoglin in MΦ was generated using the LysMCre model . The value of myeloid-specific promoters in transgenic mice has been discussed since knock-out of MΦ markers such as CD11b , CD11c and F4/80 has no impact on MΦ numbers and remarkably , little impact on MΦ function [30] . The LysMCre models do not really allow a distinction among myeloid cell types since Cre is expressed in Mo , MΦ and granulocytes . Cre-mediated excision is effective in the majority of MΦ and granulocytes but considerably lower in CD11c+ DCs [30] . In our LysMCre model , we have not seen an alteration on resident PerC MΦ numbers but we have observed effective action of Cre recombinase on endoglin floxed gene leading to reduced phagocytosis , one of the main functions of resident MΦ . Endoglin deletion in MΦ predisposes animals to develop infections by opportunistic bacteria , where S . aureus is the most predominant pathogen identified . In immunocompetent animals , S . aureus colonization of the skin , intestinal tract , or nasopharynx is generally asymptomatic while in immunocompromised or immunodeficient animals , may cause pyogenic ( abscess ) infections . In our model , mice lacking endoglin on MΦ ( Engfl/flLysMCre ) were susceptible to develop infections by opportunistic bacteria after being set up in breeding pairs . We postulate that minor wounds due to the physical interactions between mice prior to and during mating were responsible for the development of infections in Engfl/flLysMCre mice . The low incidence of spontaneous infections in Engwt/flLysMCre mice may be due to a threshold effect where endoglin has to fall below a critical level in MΦ in order to show the infectious phenotype . In this regard , mouse models of HHT suggest that AVMs in HHT patients occur following loss of heterozygosity [26] . All mice were maintained in the same room and the same conditions , but environmental factors could be relevant in determining the type and frequency of infections . Mice lacking endoglin in MΦ show other characteristics compatible with an immunocompromised phenotype . The lack of endoglin in MΦ impairs phagocytosis , and this may be affecting the initiation of the innate immune response . Upon bacteria recognition and phagocytosis , MΦ orchestrate coordinated inflammatory responses involving recruitment of neutrophils and other inflammatory cells [42] . Engfl/flLysMCre mice , and to a lesser extend Engwt/flLysMCre mice , showed a deficient recruitment of inflammatory cells to sites of infection . These differences may be influenced by the absence or deficiency of endoglin expression in resident PerC MΦ . Furthermore , animals with reduced or absent endoglin expression in MΦ showed an extended survival in the first hours following induction of septic shock . and a reduced production of inflammatory cytokines . Altogether , these data suggest that endoglin expression in resident MΦ is relevant for initiation of the innate immune response . While endoglin plays a pivotal role modulating TGF-β signaling pathways on endothelial cells , its role on TGF-β signaling in MΦ is not well known . Stable transfectants of two different alternatively spliced isoforms of endoglin in the human promonocytic cell line U937 showed that endoglin isoforms counteracts TGF-β1 inhibition of proliferation and migration [3] and displayed a differential gene expression pattern , mainly affecting biological function related to cell movement and the expression of INHBA , a TGF-β family member [39] . In the immune system , TGF-β initially plays a pro-inflammatory role , acting as a chemoattractant for Mo , and triggering the production of inflammatory mediators . However , TGF-β later functions mainly as an inhibitory molecule , when Mo differentiate into MΦ . This TGF-β inhibitory function contributes to resolve inflammation and prevents the development of immunopathologies [17] . Due to their involvement in the TGF-β signaling pathways , the expression of two relevant target genes of TGF-β; Serpine1 ( Pai-1 gene ) and Id1 , controlled by TGF-β/ALK5/Smad2/3 and by TGF-β/ALK1/Smad1/5/8 pathways in endothelium , respectively , were analyzed . While , Id1 expression seems to be independent of endoglin , Serpine1 is highly expressed on MΦ lacking endoglin compared to MΦ from control mice . This overexpression is in agreement with previous reports where endoglin and PAI-1 levels are inversely correlated [43] . Since Smad3 is a critical mediator of TGF-β inhibition of MΦ activation [44] , these results suggest that the absence of endoglin could act by increasing TGF-β signaling via ALK5/Smad2/3 on MΦ . Moreover , haploinsufficiency or absence of endoglin in MΦ leads to the repression of Acvrl1 expression . These data agree with previous reports where endothelial cells from HHT1 patients also show a downregulation of ACVRL1 expression [45] . Our observations of decreased levels of Nos2 , Mmp12 and Inhba transcripts in MΦ with reduced endoglin expression , would further strengthen the idea that endoglin counteracts in MΦ the known inhibitory effects of TGF-β1 signaling through the ALK5/Smad2/3 pathway . Furthermore , Nos2 , MMP12 and Activin A are all markers of MΦ activation [44 , 46] . Nos2 expression helps to control bacterial infections such as S . aureus [47] and MMP12 plays a role in early killing of S . aureus in the phagolysosome of MΦ [48] . Thus , the decreased expression of Nos2 and Mmp12 in Engfl/flLysMCre mice would impair bacterial clearance by MΦ . The decreased capacity for MΦ activation seen in Engfl/flLysMCre mice is compatible with a weaker primary immune response . The affected immune functions in Engfl/flLysMCre mice suggest a possible explanation for certain infectious events seen in HHT patients [21 , 22] such as rare infections , osteomyelitis , sepsis and extracerebral abscesses , among others . In fact , it has been reported that HHT patients display abnormalities of phagocytosis and oxidative burst exerted by neutrophils and Mo [49] , although another study did not find this impairment [22] . An explanation for this discrepancy is that patients with HHT exhibit a great diversity of clinical manifestations due to an incomplete penetrance of the disease and the influence of environmental factors . In addition , the cellular types analyzed are not the most appropriate since neutrophils do not express endoglin and in blood Mo , it is almost undetectable [18 , 20] . Future studies on a larger patient cohort and focusing on differentiated MΦ , would be more suitable to determine the effects of endoglin mutations on the innate immune response in HHT1 patients . The inclusion in the international guidelines of immunological assessment during management of HHT patients would be useful to prevent serious infectious outcomes . In this context , preventive protocols for vaccination [50] and review of antibiotic prophylaxis for hospitalized HHT patients should improve their clinical management and outcomes .
Specified pathogen-free C57BL/6 male 10-12-week-old mice were used in the experiments . Mice were housed under specific pathogen-free conditions at the department of Animal Resources facilities in the Centro de Investigaciones Biológicas ( CSIC ) . LysMCre mice were provided by Dr . Mercedes Ricote ( Fundación Centro Nacional de Investigaciones Cardiovasculares , Madrid , Spain ) . Endoglin floxed mice ( Engfl/fl ) were generated as described [51] and were crossed with LysMCre individuals to generate mice with specific Eng gene deletion in the myeloid lineage . The first heterozygous offspring containing the loxP-targeted Eng gene ( Engwt/fl ) and the Cre transgene ( LysMCre ) were backcrossed for 10 generations selecting heterozygous individuals ( Engwt/flLysMCre ) to achieve homogeneity . The offspring Engwt/wtLysMCre resulting from the 10th generation of backcrosses between Engwt/flLysMCre mice were then interbred to increase the number of control mice ( Engwt/wtLysMCre ) . Engwt/wtLysMCre were crossed with Engfl/flLysMCre mice to obtain the heterozygous experimental animals ( Engwt/flLysMCre ) , and Engfl/flLysMCre individuals were interbred to maintain the strain Engfl/flLysMCre . Mice were genotyped by PCR using the primers X 5’-CCACGCCTTTGACCTTGC 3’ , Y 5’-GGTCAGCCAGTCTAGCCAAG 3’ , Z 5’-GTGGTTGCCATTCAAGTGTG 3’ as described [51] and primers Cre Fw 5’-AGGTGTAGAGAAGGCACTTAGC 3’ and Cre Rv 5’-CTAATCGCCATCTTCCAGCAGG 3’ . DNA was extracted from tails using the REDExtract-N-Amp Tissue PCR Kit ( Sigma #XNAT ) . From PerC MΦ and different tissues , DNA was obtained using QIAamp DNA Mini Kit ( QIAGEN #51304 ) . For isolation of peripheral blood cells , blood samples were obtained by cardiac puncture using heparin as anticoagulant . Blood samples were treated twice with red blood cell lysis buffer ( 1g/L KHCO3 , 8 . 3g/L NH4Cl , 0 . 019% EDTA ) for 2 min at RT . Samples from a total of 10 mice were pooled . Mo were isolated by incubating the total blood leukocyte fraction at 37°C and , 5% CO2 , in autologous plasma-coated plastic flasks . Non-adherent cells were removed by extensive washing with a pre-warmed Hanks’ solution . Adherent cells were trypsinized at selected times . For isolation of Kupffer cells , liver was removed from the PerC and rinsed in Krebs-Ringer-Buffer ( KRB-1000; Zen-Bio Inc . , NC , USA ) . To obtain a single-cell suspension from mouse liver , the gentle MACS Dissociator ( #130-095-937; Miltenyi Biotec GmbH , Bergisch Gladbach , Germany ) was used following the manufacturer’s guide , using collagenase IV treatment ( C5138; Sigma-Aldrich , Saint Louis , MO , USA ) . For collection of PerC cells , 10 mL of PBS were injected in the PerC . After an abdominal soft massage , between 8 . 5 to 9 . 5 mL were recovered . The suspension obtained from peritoneal lavage was centrifuged at 1 , 200 rpm for 5 min to recover cells . Blood contaminated samples were discarded . Bone Marrow-Derived Macrophages ( BMDM ) were obtained by flushing mouse femurs with ice-cold PBS . From 5 to 8x106 cells , were cultured in DMEM supplemented with 10% heat-inactivated FCS and 50μM β-mercaptoethanol , containing human macrophage CSF ( M-CSF ) ( 25 ng/mL ) or murine GM-CSF ( 1000U/mL; ImmunoTools GmbH ) , respectively , for 7 days to obtain 95%-pure CD11bpos BMDMs . Cytokines were added every two days . After this , media were discarded and cells rinsed twice with saline solution . Peritoneal MΦ were harvested from Engwt/wtLysMCre , Engwt/flLysMCre and Engfl/flLysMCre mice as described above . After 1 hour of incubation at 37°C , non-adherent cells were discarded by extensive washes with warm Hanks’ solution and then incubation was continued for 24h at 37°C . SR . D10-CD4negF1 was a CD4neg mutant cell line cloned from the mouse CD4pos TH2 cell line D10 . G4 . 1 [52] . It was grown in Click medium with 10% FCS , 9% ( v/v ) β-mercaptoetanol , 5 U/mL IL-2; 10 U/mL IL-4 , and 25 pg/mL IL-1α . Peripheral blood was obtained by puncturing the cava vein . Blood was drawn into EDTA-coated tubes ( Monlab SL , Spain ) . Complete blood profiles and hemoglobin levels were obtained using an Abacus Junior Vet ( Diatron ) hematology analyzer . Values are shown as absolute counts and referenced against the normal range established for mice . Ten week-old mice were i . p . injected with 1 mg of Zymosan A from Saccharomyces cerevisiae ( Z4250; Sigma-Aldrich ) in 0 . 5 mL of sterile PBS . Isolation of PerC cells was performed as described above . For in vivo evaluation of endoglin expression on MΦ surface , samples were collected at 12h , 24h , 36h , 3 , 7 and 14 days after injection of Zymosan A , and stained for flow cytometry analysis . Samples of unstimulated mice were used as time 0 . For other experiments , animals were similarly i . p . injected with 1 mg of Zymosan A in 0 . 5 mL of sterile PBS . Twenty-four hours later , PerC exudates were recovered to evaluate the leukocyte recruitment to the PerC after ZIP . The number of total leukocytes in PerC was evaluated in a CASY Cell Counter . Events were considered leukocytes above a threshold of 5 . 7 μm diameter . Percentage of different leukocyte subpopulations in PerC was evaluated by flow cytometry . BMDMs and cell suspensions were blocked with PBS containing 5% of rabbit serum for 20min at 4°C , followed by incubation with an Ab against endoglin ( eBioscience , 14–1051 ) or rat anti-mouse isotype control ( eBioscience , 14–4321 ) for 1h at 4°C . Thereafter , cells were washed twice with 1% BSA in PBS and incubated with a FITC-conjugated F ( ab’ ) 2 rabbit anti-rat IgG ( Invitrogen A11006 ) for 20 min at 4°C . After endoglin staining , cell suspensions were washed twice and incubated at 4°C during 20 min with the following monoclonal antibodies: PE anti-mouse F4/80 ( Biolegend; 122616 ) , PE anti-CD11b ( Immunostep; M11BPE ) , APC anti-mouse Ly6G ( Immunostep; MLY6GA ) , anti-mouse CD19 FITC ( Immunostep; M19F ) , anti-mouse CD3e FITC ( Immunostep; 220911 ) . For isotype controls , antibodies were: PE rat IgG2b , κ ( Biolegend; 400608 ) , rat IgG2a APC ( Immunostep; 220812/RIGG2A ) and Alexa 488 rat IgG2a , κ ( Biolegend; 400525 ) . Unbound antibodies were removed by washing twice with PBS containing 1% of BSA . Flow cytometry analyses were performed with a Beckman Coulter FC500 cytometer . Cells were first selected on the basis of their FS vs SS properties . Dead cells were localized by propidium iodide ( Sigma #81845 ) exclusion to set the gating area of interest . A minimum of 5 , 000 stained cells per sample was analyzed . Upon gating , levels of endoglin were analyzed on the CD11bpos Ly6Gneg CD3neg population of adherent cells for in vitro assay and on the F4/80pos Ly6Gneg CD19neg of cell suspensions from liver and PerC . Flow cytometry experiments were carried out by fitting isotype controls to the first decade on log histograms , setting upper limit at 10° . A residual percentage of positive cells lower than 5% above this limit was considered as a negative control . Endoglin expression is represented by the percentage of positive cells , and cells are considered positive for endoglin expression when the population is over 5% . Immediately after sacrifice , mice were perfused with freshly prepared 1% paraformaldehyde ( PFA ) . Liver was excised and fixed in 1% PFA for 12h at 4°C , and 15% and 30% sucrose solution , until the specimens were decanted , and frozen in OCT . Cryosections were incubated with an Ab against endoglin ( eBioscience , 14–1051 ) or rat anti-mouse isotype control ( eBioscience , 14–4321 , overnight at 4°C . Endoglin was detected following 1 hour incubation with Alexa 488 anti-rat ( Molecular Probes #A-11006 ) . All the incubations were done in the presence of 5% goat serum in PBS . Staining was visualized by laser confocal scanning microscopy ( TCS-SP2-AOBS; Leica ) . For sEng measurements , serum from 10–12 week old mice isolated from cava vein was used . Blood samples were collected and centrifuged at 2 , 000xg for 20 minutes to collect serum from whole blood . Serum was collected and kept at -20°C until analysis . The levels of sEng were determined using the Mouse Endoglin/CD105 Quantikine ELISA sandwich kit ( R&D Systems #MNDG00 ) , following the manufacturer’s guide . Individuals with spontaneous infections were sacrificed . Infected areas were cleaned under sterile conditions with sterile PBS , excised and transferred to sterile eppendorf tubes and kept at 4°C . Samples were sent to the Microbiology department of Clinical Veterinary Hospital ( Complutense University , Madrid ) . Isolation of microorganisms was carried out by selective media and identification was achieved by API strips . Spleens were excised from the PerC and rinsed in PBS . Spleen length was measured in all animals suspected of infection ( n = 22 ) and in 10 healthy Engfl/flLysMCre mice . Twelve-week-old mice of each genotype were i . p . injected with 40 mg/kg LPS ( E . coli 0111;B4; Sigma ) . The survival rate was followed for 5 days in Engwt/wtLysMCre ( n = 33 ) , Engwt/flLysMCre ( n = 24 ) and Engfl/flLysMCre ( n = 34 ) mice . TNF-α , IL-1β and IL-6 levels were analyzed 1 hour after ZIP in PerC samples , and 3 hours post-ZIP in serum samples . Blood samples were obtained by puncture of posterior cava vein , and centrifuged at 2 , 000g for 20 min at 4°C to obtain serum samples . PerC exudates were obtained by i . p . injection of 10 mL of sterile PBS . Between 8 . 5–9 . 5 mL were recovered . TNF-α , IL-1β and IL-6 levels were quantified with ELISA kits ( Quantikine R&D Systems ) . 1 . 5 x 105 SR . D10-CD4neg lymphocytes in a final volume of 100μL of serum free DMEM were placed on the upper compartment of individual Transwell ( Costar ) chamber wells with pores of 5 μm diameter . Cells were allowed to migrate for 3 hours towards PerC exudates from Engwt/wtLysMCre , Engwt/flLysMCre and Engfl/flLysMCre mice . Cells migrating to the lower compartment were counted by flow cytometry . The percentage of migration was normalized to exudates from control mice ( Engwt/wtLysMCre ) . Zymosan particles at a final concentration of 1mg/mL in PBS were CFSE labelled ( 45 μM ) for 15 min at RT , washed three times with PBS and sonicated in RPMI DMEM during 15 min before assay . Mice were i . p . injected with 50 μg of CFSE-stained zymosan particles in 500μL of sterile PBS . After 90 min , mice were anesthetized to isolate PerC exudates . Cellular suspension was processed for F4/80 flow cytometry detection . Phagocytic activity was calculated as the percentage of PerC MΦ that incorporated CFSE-Zymosan particles ( F4/80posCFSEpos ) . Phagocytic efficiency represents the Mean Fluorescence Intensity ( MFI ) of CFSE in F4/80pos cells . Total cellular RNA was extracted using the NucleoSpin RNA II ( Macherey-Nagel , Düren , Germany ) . Six hundred nanogram of total RNA was reverse transcribed in a final volume of 20 μL with the Kit First Strand cDNA Synthesis ( Roche , Mannheim , Germany ) , using random primers . One μL of fresh cDNA was subjected to real time PCR ( in triplicate ) using Sybr-green master mix from Bio-Rad . Transcripts of Eng , Acvrl1 , Serpine1 , Id1 , Nos2 , Mmp12 , Inhba and 18s ( as endogenous control ) , were amplified using specific primers with the following sequences: Eng Fw: 5’ CGATAGCAGCACTGGATGAC 3’ Rv: 5’ AGAATGGTGCCTTTGGGTCT 3’ Acvrl1 Fw: 5’ TGACCTCAAGAGTCGCAATG 3’ Rv: 5’ CTCGGGTGCCATGTATCTTT 3’ Serpine1 Fw: 5’ GTCTTTCCGACCAAGAGCAG 3’ Rv: 5’ GACAAAGGCTGTGGAGGAAG 3’ Id1 Fw: 5’ GCGAGATCAGTGCCTTGG 3’ Rv; 5’ CTCCTGAAGGGCTGGAGTC 3’ Inhba Fw: 5’ ATCATCACCTTTGCCGAGTC 3’ Rv: 5’ TCACTGCCTTCCTTGGAAAT 3’ Nos2 Fw: 5’ TGGCCACCAAGCTGAACT 3’ Rv: 5’ TTCATGATAACGTTTCTGGCTCT 3’ Mmp12 Fw: 5’ CCACTTCGCCAAAAGGTTTA 3’ Rv: 5’ GGGGTAAGCAGGGTCCAT 3’ 18s Fw; 5’-CTCAACACGGGAAACCTCAC 3’ Rv: 5’- CGCTCCACCAACTAAGAACG 3’ . A pool of RNA from 3 individuals was used for each condition . The experiment was repeated five times . Data were analyzed using SPSS version 11 . 0 . 0 ( SPSS Inc . , Chicago , IL ) . Reported values are expressed as median and SEM unless otherwise described . Comparisons between control ( Engwt/wtLysMCre ) and experimental mice ( Engwt/flLysMCre and Engfl/flLysMCre ) were made using ANOVA . Bonferroni’s post hoc multiple comparisons testing was performed to detect significant differences . All comparisons were two-tailed . Survival rates were represented as a Kaplan–Meier curve , and the results were analyzed with a log-rank ( Mantel–Cox ) t test . Statistical significance is displayed as * ( P<0 . 05 ) , ** ( P<0 . 01 ) or *** ( P<0 . 001 ) . Mice were maintained under specific pathogen-free conditions at department of Animal Resources facilities in the Centro de Investigaciones Biológicas ( CSIC ) . All animals were handled in strict accordance with good animal practice as defined by the national animal welfare bodies ( RD 1201/2005 BOE #252 ) . The experimental design and all animal work were approved by our institutional Ethical Committee following the guidelines of EU Directive 2010/63/UE for animal experiments . For the isolation of organs and body fluids , all animals were deeply anesthetized by isoflurane and sacrificed by cervical dislocation .
|
Endoglin is a transmembrane protein and an auxiliary receptor for TGF-β with an important role in the homeostasis of the vessel wall . However , endoglin was originally identified as a human cell surface antigen expressed in a pre-B leukemic cell line . Mutations in ENG are responsible for the Hereditary Hemorrhagic Telangiectasia type 1 ( HHT1 ) or Rendu-Osler-Weber syndrome . HHT is a rare disease , with a prevalence of 1/5 , 000 to 1/8 , 000 . It is an autosomal dominant disorder characterized by a multisystemic vascular dysplasia , recurrent hemorrhages and arteriovenous malformations in internal organs . Interestingly , endoglin expression is also triggered during the monocyte-macrophage differentiation process . In our laboratory , we described that up-regulation of endoglin during in vitro differentiation of blood monocytes is age-dependent and impaired in monocytes from HHT patients , suggesting a role of endoglin in macrophages . In the present work , we first analyzed endoglin expression during differentiation of peripheral blood monocytes to macrophages under in vitro and in vivo conditions . Next , to investigate endoglin’s role in macrophage function in vivo , a myeloid-lineage specific endoglin knock-out mouse line was generated ( Engfl/flLysMCre ) . Endoglin deficiency in macrophages predisposed animals to spontaneous infections and led to delayed endotoxin-induced mortality . Phagocytic activity by peritoneal macrophages was reduced in the absence of endoglin and altered expression of TGF-β target genes was consistent with an altered balance of TGF-β signaling . The results show a novel role for endoglin in mouse macrophages , which if analogous in human macrophages , may explain , at least in part , the increased infection rates seen in HHT patients .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"blood",
"cells",
"flow",
"cytometry",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"body",
"fluids",
"pathology",
"and",
"laboratory",
"medicine",
"immunology",
"opportunistic",
"infections",
"animal",
"models",
"model",
"organisms",
"signs",
"and",
"symptoms",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"white",
"blood",
"cells",
"inflammation",
"animal",
"cells",
"mouse",
"models",
"hematology",
"immune",
"response",
"spectrophotometry",
"cytophotometry",
"signal",
"transduction",
"blood",
"cell",
"biology",
"anatomy",
"physiology",
"tgf-beta",
"signaling",
"cascade",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"macrophages",
"cell",
"signaling",
"spectrum",
"analysis",
"techniques",
"signaling",
"cascades"
] |
2016
|
Mice Lacking Endoglin in Macrophages Show an Impaired Immune Response
|
Herpesviruses are large DNA viruses which depend on many nuclear functions , and therefore on host transport factors to ensure specific nuclear import of viral and host components . While some import cargoes bind directly to certain transport factors , most recruit importin β1 via importin α . We identified importin α1 in a small targeted siRNA screen to be important for herpes simplex virus ( HSV-1 ) gene expression . Production of infectious virions was delayed in the absence of importin α1 , but not in cells lacking importin α3 or importin α4 . While nuclear targeting of the incoming capsids , of the HSV-1 transcription activator VP16 , and of the viral genomes were not affected , the nuclear import of the HSV-1 proteins ICP4 and ICP0 , required for efficient viral transcription , and of ICP8 and pUL42 , necessary for DNA replication , were reduced . Furthermore , quantitative electron microscopy showed that fibroblasts lacking importin α1 contained overall fewer nuclear capsids , but an increased proportion of mature nuclear capsids indicating that capsid formation and capsid egress into the cytoplasm were impaired . In neurons , importin α1 was also not required for nuclear targeting of incoming capsids , but for nuclear import of ICP4 and for the formation of nuclear capsid assembly compartments . Our data suggest that importin α1 is specifically required for the nuclear localization of several important HSV1 proteins , capsid assembly , and capsid egress into the cytoplasm , and may become rate limiting in situ upon infection at low multiplicity or in terminally differentiated cells such as neurons .
Herpesviruses such as herpes simplex virus ( HSV ) , human cytomegalovirus or Epstein-Barr virus cause human diseases ranging from minor ailments to life threatening acute infections , blindness or cancers , particularly in immunocompromised patients . They are complex DNA viruses that depend on many nuclear functions; e . g . triggering the release of the viral genomes from incoming capsids , nuclear import of viral genomes , viral gene expression , genome replication , assembly of progeny capsids , genome packaging into capsids and nuclear capsid egress . Despite these multiple interactions , little is known about the host transport factors that herpesviruses rely on for import through the nuclear pore complexes ( NPCs ) during infection . NPCs are the gateways for bidirectional trafficking between cytoplasm and nucleoplasm . The GTPase Ran controls the activity of transport factors to achieve active nuclear import and export of host and viral cargoes . While some import cargoes bind directly to a member of the importin β superfamily , the majority requires one of the importin α isoforms as an adaptor to interact with importin β1 . All importin α isoforms share an N-terminal auto-inhibitory importin β1 binding domain followed by a helical core domain of 10 stacked armadillo repeats ( ARM ) , and a small C-terminal acidic cluster; the 7 human importin α isoforms have an amino acid sequence conservation of 42% ( [1–4]; reviewed by [5 , 6] ) . Classical mono-partite nuclear localization signals ( NLSs ) utilize a major binding site on ARM 2 to 4 , and bipartite NLSs in addition to ARM 2 to 4 a minor binding site on ARM 6 to 8 [7 , 8] . Furthermore , the C-terminal acidic domain and ARM 9 and 10 contain a third binding site for non-canonical binding motifs [3 , 9–11] . Different importin α isoforms bind to similar , if not identical NLSs in vitro , and their recognition mechanisms are structurally conserved from yeast to human; yet , the affinities to specific importin α isoforms can vary considerably , and they display striking differences in cargo recognition in vivo ( [2 , 8 , 12–14]; reviewed in [5 , 6] ) . Importin α links its cargo to importin β1 , which in turn binds to NPC proteins to import such ternary complexes into the nucleoplasm , where they disassemble upon interaction with RanGTP ( reviewed in [7 , 15–17] . The nuclear import of several herpesvirus proteins has been shown in transient expression experiments to occur via binding of their NLS to importin α and thus indirectly to importin β1 . However , few studies have investigated the specificity of importin α usage in vitro , let alone in vivo in the context of a viral infection . Among the herpesviruses , interactions of host nuclear transport factors with viral proteins have been investigated at most for herpes simplex virus type 1 ( HSV-1 ) , an alphaherpesvirus that productively replicates in epithelial cells , fibroblasts and neurons . After viral fusion with a host membrane , the incoming capsids utilize dynein for microtubule-mediated transport to the nucleus [18–22] . Capsids covered by inner tegument proteins can bind to the NPCs on nuclei isolated from rat liver or reconstituted from Xenopus laevis egg extracts [23 , 24] . Incoming capsids lacking the large inner tegument protein pUL36 are not targeted to nuclei , and antibodies directed against pUL36 reduce nuclear targeting [25–27] . O’Hare and collaborators have characterized a conserved N-terminal NLS in pUL36 that is essential for targeting incoming capsids to the nucleus and for genome release [28 , 29] . A likely scenario is that this NLS interacts with host nuclear transport factors to mediate capsid docking to the NPCs . Furthermore , importin β , the RanGTP/GDP cycle and capsid-NPC interactions are required to trigger genome uncoating from capsids; however , a function for importin α could not be uncovered in these in vitro assays [23] . HSV-1 promotors in general contain regulatory sequences common with host genes , and are sequentially regulated with immediate-early , early and late gene expression kinetics unless the incoming genomes are repressed and silenced by facultative heterochromatin ( reviewed in [30–33] ) . The tegument viral protein VP16 dissociates from incoming capsids and complexes with the host cell factor HCF-1 and the POU homeodomain protein Oct-1 to keep immediate-early HSV1 promotors de-repressed for transcription ( reviewed in [34] ) . VP16 does not seem to contain an own NLS but piggy-backs onto HCF-1 in the cytosol for co-import into the nucleus; VP16 is not imported into the nucleoplasm , when the NLS in HCF-1 has been mutated [35] . In the nucleoplasm , VP16/HCF binds to Oct-1 that is already associated with HSV-1 promotors [36] . The NLS of Oct-4 interacts with importin α1 , Oct-6 with importin α5 , while the one of Oct-1 has not been characterized [11 , 37 , 38] . In addition to binding sites for VP16 , immediate-early HSV-1 promotors also include response elements for the host transcription factors SP1 and GABP [39] . HSV-1 early and late promotors also contain SP1 transcription factor binding sites , and the transcription of viral genes increases after DNA replication due to the increased template number [32 , 40 , 41] . The major transactivator ICP4 ( infected cell protein 4 ) , the regulators ICP22 and ICP27 , and the E3-ubiquitin ligase ICP0 are immediate-early nuclear HSV-1 proteins important for early and late transcription . While their NLSs have been mapped , their nuclear transport factors are not known [42–45] . ICP4 is required for maximal expression from early and late promotors; it recruits the host RNA polymerase II and other host factors , ICP22 and ICP27 , and stabilizes the pre-initiation complex [46] . ICP27 is required for efficient viral transcription and translation of some early and early-late genes and perhaps all true late genes . It needs to shuttle between the cytosol and the nucleoplasm to enhance the nuclear export of intron-lacking viral mRNAs and thus their expression ( reviewed in [47] ) . ICP0 also increases the expression of early and late genes; particularly at a low MOI and in vivo ( reviewed in [48] ) . The formation of the nuclear HSV-1 DNA replication compartments results in host chromatin marginalization towards the nuclear rim , and requires seven HSV-1 proteins synthesized with early kinetics . These are the origin-binding protein pUL9 , the ssDNA binding protein ICP8 ( pUL29 ) , the heterotrimeric helicase-primase complex ( pUL5 , pUL8 , pUL52 ) , and the DNA polymerase with the catalytic subunit pUL30 and its processivity factor pUL42 ( reviewed in [49 , 50] ) . An NLS of pUL9 has been mapped to its amino acid residues 793 to 804 [51] , and the nuclear localization of ICP8 is mediated by its 28 C-terminal amino acid residues [52] . In contrast , the subunits of the primase/helicase complex remain cytosolic when translated in isolation; but their assembly is sufficient to generate an NLS for nuclear import in the case of HSV-1 , Epstein-Barr virus , and Kaposi sarcoma herpesvirus [53–55] . The NLSs of the DNA polymerase subunits have been well characterized for HSV-1 , the human cytomegalovirus , Epstein-Barr virus and Kaposi sarcoma herpesvirus ( reviewed in [56] . Capsid assembly and packaging of the viral genomes also occur in the nucleoplasm , but the major capsid protein VP5 , the capsid protein VP23 , and the small capsid protein VP26 are not capable of nuclear import on their own [57] . VP5 requires the capsid scaffolding protein VP22a for localization to the cell nucleus [58] , and a non-classical NLS of the triplex capsid protein VP19c is responsible for the nuclear import of the other triplex protein VP23 [57 , 59 , 60] . Furthermore , the NLSs of pUL15 and pUL33 , of the terminase that catalyzes genome packaging into preassembled capsids , have been characterized in detail [61] . Thus although some few direct interactions between host transport factors and viral nuclear proteins have been elucidated , host transport factors required for specific steps in the herpesvirus life cycle have not been identified yet . Considering that herpesviruses rely on so many nuclear functions , we conducted an RNAi screen to identify nuclear transport factors that are relevant for efficient HSV-1 gene expression . Of the 17 host factors that we had targeted , importin β1 , importin α1 , importin α6 , and transportin 1 were required for efficient HSV-1 gene expression while importin 11 , importin 8 , transportin 3 and importin 9 seemed to repress HSV-1 . Our experiments with fibroblasts from knock-out mice or transduced with lentiviral vectors encoding for shRNAs to perturb the expression of specific importin α isoforms showed that efficient nuclear import of the HSV-1 immediate-early proteins ICP4 and ICP0 , and the early proteins ICP8 and DNA polymerase required importin α1 and importin α3 but was restricted by importin α4 . Furthermore , the assembly of nuclear capsids , capsid egress into the cytoplasm and formation of infectious virions were reduced in the absence of importin α1 , while nuclear targeting of incoming capsids , nuclear import of VP16 and of incoming genomes were not impaired . Similarly , when the expression of importin α1 had been silenced in neurons , nuclear targeting of incoming capsids from the somal plasma membrane or the axonal compartment were also not impaired , but the nuclear import of ICP4 , HSV-1 gene expression , and the formation of nuclear capsid compartments was prevented . Our data indicate that the nuclear import of several important HSV-1 proteins and thus efficient HSV-1 infection depend specifically on importin α1 in fibroblasts , and even more so in neurons .
To identify nuclear transport factors required for HSV-1 replication , we transfected HeLa cells with specific siRNAs and infected them at 72 hpt ( hour post transfection ) with the reporter strain HSV1 ( 17+ ) Lox-pMCMVGFP which expresses GFP under the control of a murine cytomegalovirus promoter . At 12 hours post infection ( hpi ) , the HSV1-mediated GFP expression ( Fig 1A ) , and the cell density based on DNA staining were measured in a plate reader ( Fig 1B ) . The GFP signals upon transfection of a scrambled siRNA were normalized to 100% and the background signals of a mock-infected control to 0% . An siRNA directed against the GFP transcripts or treatment with nocodazole served as controls , and reduced HSV1-mediated GFP expression by more than 75% , as expected [62 , 63] . Nocodazole depolymerizes microtubules that are required for efficient transport of incoming capsids to the nuclear pores , and thus for viral gene expression in epithelial cells [18 , 20 , 64 , 65] . Franceschini et al . ( 2014 ) have developed an algorithm to subtract some off-target effects of siRNAs with promiscuous seed regions [66] . We applied their criteria to our data which resulted in re-calculating the effect of 4 siRNAs on HSV-1 gene expression ( c . f . S1 Table , GFPcorr ) . Silencing the expression of some nuclear transport factors reduced cell density , particularly in the case of importin β1 ( KPNB1 ) , which is involved in many physiological processes [67 , 68] . We therefore determined the ratios of the GFPcorr signals over the DNA signals , and ranked the nuclear transport factors according to a decreasing inhibition of HSV1-mediated GFP expression per cell upon siRNA treatment ( Fig 1C; S1 Table ) . Individual siRNAs targeting importin β1 ( gene KPNB1 ) , importin α1 ( KPNA2 ) , importin α6 ( KPNA5 ) , or transportin 1 ( TNPO1 ) decreased HSV-1 mediated GFP/DNA expression on average by more than 30% , whereas most siRNAs directed against importin α7 ( KPNA6 ) , importin 4 ( IPO4 ) , importin α3 ( KPNA4 ) , importin 7 ( IPO7 ) , importin α4 ( KPNA3 ) , importin α5 ( KPNA1 ) , transportin 2 ( TPNO2 ) , or Ran binding protein 5 ( RANBP5 ) on average had little effect . In contrast , HSV1-mediated GFP expression was markedly increased by some siRNAs aiming at importin 13 ( IPO13 ) , importin 11 ( IPO11 ) , importin 8 ( IPO8 ) , transportin 3 ( TPNO3 ) , or importin 9 ( IPO9 ) . These data suggested that HSV1-mediated GFP expression in human HeLa cells particularly depended on importin ß1 , importin α1 , importin α6 , and transportin 1 , but might have been restricted by the activities of importin 13 , importin 11 , importin 9 , transportin 3 , and importin 9 . The nuclear transport factors that were required for efficient HSV-1 mediated GFP expression might contribute to ( i ) the release of the incoming HSV-1 genomes into the nucleoplasm , ( ii ) the nuclear import of host transcription factors operating on the MCMV promoter , such as NF-ΚB , AP-1 , and SP-1 , or ( iii ) the nuclear import of host or viral factors required for HSV-1 DNA replication , since the amount of the GFP reporter protein depends on the copy number of HSV-1 genomes in the nucleus . Since we had already shown that importin β1 promotes targeting of incoming HSV-1 capsids to NPCs and viral genome uncoating [23] , we focused on the next potential hit , importin α1 ( KPNA2 ) . Promiscuous siRNA seed regions might result in off-target effects [66] , and importin α isoforms are highly homologous; we therefore decided to use murine embryonic fibroblasts ( MEFs ) derived from specific importin α knock-out mice for functional experiments . Like others , we use the numbering of the human proteins also for their closest murine homologs: importin α1 ( hImp α1 , gene KPNA2; mImp α2 , kpna2 ) and importin α8 ( KPNA7; kpna7 ) for members of the RCH-family , importin α3 ( hImp α3 , KPNA4; mImp α4 , Kpna4 ) and importin α4 ( hImp α4 , KPNA3; mImp α3 , Kpna3 ) for the QIP family , and importin α5 ( KPNA1; Kpna1 ) , importin α6 ( KPNA5; no murine homolog ) , and importin α7 ( KPNA6; Kpna6 ) for the SRP family [5 , 69–72] . Mouse embryonic fibroblasts ( MEFs ) derived from importin α1 ( MEF-Impα1-/- ) , importin α3 ( MEF-Impα3-/- ) , or importin α4 ( MEF-Impα4-/- ) [73] knock-out mice lacked the respective importin α proteins while the levels of other importins had not been reduced ( S1A Fig ) . These data validate the specificity of the polyclonal anti-peptide antibodies and the respective MEF lines used in this study . The first step of the HSV-1 life cycle suggested to recruit an importin α via an NLS is docking of incoming capsids at the NPCs [23–25 , 28] . We therefore infected MEFs with HSV1 ( 17+ ) Lox-CheVP26 in the presence of cycloheximide to prevent synthesis of progeny HSV-1 proteins , and analyzed the subcellular localization of incoming capsids by confocal fluorescence microscopy . In this HSV-1 reporter strain , the small capsid protein VP26 has been tagged with monomeric Cherry ( CheVP26; [74–76] ) . At 4 hpi , many HSV-1 capsids , detected by CheVP26 ( Fig 2Aii; red in Fig 2Aiv ) and/or by antibody labeling ( Fig 2Aiii , green in Fig 2Aiv ) , had accumulated at the nuclear rims ( Fig 2Ai and blue line Fig 2Aii , 2Aiii and 2Aiv ) of MEFwt ( Fig 2A ) . As in epithelial cells [18 , 20] , nocodazole treatment reduced nuclear targeting in MEFwt , and instead the capsids were dispersed throughout the entire cytoplasm ( Fig 2B ) . In contrast , incoming capsids accumulated at the nuclear rims of MEF-Impα1-/- ( Fig 2C ) , MEF-Impα3-/- ( Fig 2D ) or MEF-Impα4-/- ( Fig 2E ) . Thus , HSV-1 internalization into cells and nuclear targeting of incoming capsids were not impaired in MEF-Impα1-/- , MEF-Impα3-/- , or MEF-Impα4-/- . As efficient HSV-1 gene expression depends on genome uncoating from the capsids and release into the nucleoplasm , we examined the nuclear import of incoming HSV-1 genomes . MEFs were inoculated with HSV1 ( 17+ ) at a high MOI in the presence of cycloheximide , denatured at 3 hpi with an ethanol/acetic acid mixture , and hybridized with a Cy3-labeled DNA probe specific for HSV-1 . The cytoplasm and the nuclei of the MEFwt contained many spots of hybridized HSV-1 genomes and mRNAs ( S2Aiii Fig ) . In contrast , there were no signals for HSV-1 in mock-treated cells ( S2Biii Fig ) . The amount of nuclear HSV-1 nucleic acids appeared similar to MEFwt in MEF-Impα1-/- ( S2Ciii Fig ) , MEF-Impα3-/- ( S2Diii Fig ) , and MEF-Impα4-/- ( S2Eiii Fig ) . Efficient HSV-1 gene expression also depends on nuclear VP16 , and we therefore investigated its subcellular localization upon inoculation in the presence of cycloheximide . At 4 hpi , HSV1-VP16 had accumulated to a similar extent in the nuclei of MEFwt ( S2Fi Fig ) , MEFwt treated with nocodazole ( S2Gi Fig ) , MEF-Impα1-/- ( S2Hi Fig ) , MEF-Impα3-/- ( S2Ii Fig ) , and MEF-Impα4-/- ( S2Ji Fig ) . In MEFwt inoculated with the mutant HSV1 ( 17+ ) Lox-ΔgB [77] , VP16 had not reached the nucleoplasm as expected , but been retained in virions , located either at the plasma membrane or within endosomes ( S2Ki Fig ) . Glycoprotein B ( gB ) is essential for HSV-1 cell entry as it catalyzes the fusion of viral with host membranes [78 , 79] . Consistent with an unimpaired nuclear targeting of incoming capsids , of genomes , and of VP16 , we furthermore did not detect any major reorganization of the microtubule network or the distribution of NPC proteins among the different MEF lines ( S3 Fig ) . Taken together , HSV-1 internalization , nuclear targeting of incoming capsids , nuclear import of HSV-1 genomes , and nuclear import of VP16 occurred with similar efficiencies in MEFwt , MEF-Impα1-/- , MEF-Impα3-/- and MEF-Impα4-/- . To determine whether importin α1 is required for viral protein expression , MEFwt , MEF-Impα1-/- , MEF-Impα3-/- , or MEF-Impα4-/- were infected with HSV1 ( 17+ ) Lox and analyzed at 6 hpi by immunoblot . For calibration , we compared the lanes of the knock-out cell lines to lanes in which 25% , 50% or 100% of a comparably infected MEFwt lysate had been loaded ( S4 Fig; WT , 25 , 50 , 100 ) . By 6 hpi , MEFwt and the 3 knock-out lines expressed the immediate-early protein ICP4 , the early protein ICP8 , and the late tegument proteins VP16 and VP22 ( S4A Fig ) . In contrast , when MEFwt had been inoculated in the presence of nocodazole these proteins were barely detected . A quantitation showed that the expression of ICP4 , ICP8 and the late tegument protein VP22 were moderately reduced in the absence of importin α1 , but increased in cells lacking importin α4 ( S4B Fig ) . These data indicate that neither importin α1 , importin α3 , or importin α4 were obligatory but that importin α1 facilitated efficient HSV-1 protein expression while importin α4 restricted it to a certain extent . We next determined the impact of different importin α isoforms on the subcellular localization of several HSV-1 proteins required for early gene expression and for DNA replication . MEFwt , MEF-Impα1-/- , MEF-Impα3-/- , or MEF-Impα4-/- were infected with HSV1 ( 17+ ) Lox-CheVP26 , labeled for various HSV-1 proteins , stained for DNA , and analyzed by confocal fluorescence microscopy . By 4 hpi , ICP4 was detected in most nuclei of MEFwt although its amount varied considerably among individual cells ( Fig 3Ai ) . After infection of MEFwt in the presence of nocodazole , ICP4 was not detected ( Fig 3Bi ) , whereas in MEF-Impα1-/- ( Fig 3Ci ) and in MEF-Impα3-/- ( Fig 3Di ) there was some nuclear ICP4 , although considerably less than in MEFwt or MEF-Impα4-/- ( Fig 3Ei ) . A quantification of more than 150 cells for each condition showed that the control nocodazole treatment prevented nuclear localization of ICP4 , and that there was significantly less nuclear ICP4 in MEF-Impα1-/- and in MEF-Impα3-/- , but more in MEF-Impα4-/- when compared to MEFwt ( Fig 3F ) . Similar results were obtained for ICP0 ( S5 Fig ) . Infection in the presence of nocodazole had also prevented ICP0 expression ( S5B Fig ) , and there was less nuclear ICP0 in MEF-Impα1-/- ( S5Ci Fig ) and in MEF-Impα3-/- ( S5Di Fig ) , but not in MEF-Impα4-/- ( S5Ei Fig ) when compared to MEFwt ( S5Ai Fig ) . The quantification confirmed that the nuclear localization of ICP0 depended on both importin α1 and importin α3 , but not on importin α4 ( Fig 3G ) . Seven HSV-1 early proteins including ICP8 and pUL42 catalyze nuclear viral DNA replication . By 6 hpi , ICP8 was detected in most nuclei of MEFwt although its amount varied also among cells . ICP8 was diffusively distributed over the entire nucleoplasm , but clearly enriched in certain nuclear regions ( S5Fi Fig ) which are the sites of HSV-1 DNA replication ( reviewed in [49 , 50] ) . Infection of MEFwt in the presence of nocodazole did not reveal any ICP8 ( S5Gi Fig ) , whereas in MEF-Impα1-/- ( S5Hi Fig ) and MEF-Impα3-/- ( S5Ii Fig ) , there was some nuclear ICP8 , although considerably less than in MEFwt ( S5Fi Fig ) or MEF-Impα4-/- ( S5Ji Fig ) . Similarly , the amount of nuclear pUL42 was lowered in MEF-Impα1-/- ( S5Mi Fig ) and in MEF-Impα3-/- ( S5Ni Fig ) when compared to MEF-Impα4-/- ( S5Oi Fig ) or MEFwt ( S5Ki Fig ) , and there was very little nuclear pUL42 if the MEFwt had been inoculated in the presence of nocodazole ( S5Li Fig ) . The quantification of these images showed that the nuclear localization of ICP8 was reduced in the absence of importin α1 to a similar level as treatment with nocodazole , and also reduced in the absence of importin α3 , but increased without importin α4 when compared to MEFwt ( Fig 3H ) . Similarly , the nuclear localization of pUL42 was also dependent on importin α1 and on importin α3 but not on importin α4 ( Fig 3I ) . While the MEF cell lines derived from knock-out animals unequivocally did not express the targeted importin α isoform , they may have compensated its absence during passage in cell culture by increased or decreased expression of other isoforms or related transport factors . As an additional approach , we therefore validated lentiviral vectors expressing shRNAs to silence the expression of importin α1 , importin α3 , or importin α4 without impairing the expression of other importin α isoforms ( S1B Fig ) . We then infected MEFwt transduced with specific shRNAs or a scrambled shRNA with HSV-1 using the same conditions as for the MEF knock-out lines . The nuclear localization of ICP4 ( S6A–S6E Fig , Fig 3J ) , ICP0 ( Fig 3K ) , ICP8 ( S6F–S6J Fig , Fig 3L ) , and pUL42 ( Fig 3M ) was significantly reduced upon silencing the expression of importin α1 or importin α3 . In contrast , silencing importin α4 expression did not affect the nuclear targeting of ICP4 , ICP0 or ICP8 , but increased the nuclear amounts of pUL42 . In summary , targeting importin α4 with shRNA did not affect the nuclear amounts of three HSV-1 proteins but lead to an increase of nuclear pUL42 . Similarly , the nuclear amount of ICP0 and pU42 was similar in MEF-Impα4-/- as in MEFwt , but increased for ICP4 and ICP8 . In contrast , importin α1 and importin α3 were required for efficient nuclear localization of the immediate-early expressed proteins ICP4 and ICP0 and the early expressed proteins ICP8 and pUL42 . As infection progressed to later phases of the viral life cycle , MEFwt , MEF-Impα1-/- , MEF-Impα3-/- , or MEF-Impα4-/- infected with HSV1 ( 17+ ) Lox-CheVP26 were analyzed for nuclear capsid compartments . By 8 hpi , the nuclei of MEFwt ( Fig 4Ai ) , MEF-Impα1-/- ( Fig 4Ci ) , MEF-Impα3-/- ( Fig 4Di ) , and MEF-Impα4-/- ( Fig 4Ei ) contained prominent amounts of nuclear capsid proteins but no nuclear capsid proteins were detected upon infection in the presence of nocodazole ( Fig 4Bi ) . A quantitation showed that the amount of nuclear capsid protein was similar in MEFwt , MEF-Impα1-/- , and MEF-Impα3-/- , and even increased in MEF-Impα4-/- ( Fig 4F ) . A similar experiment with MEFwt transduced with specific or scrambled shRNAs indicated a moderate reduction in the amount of nuclear capsid protein upon silencing importin α1 expression but no changes in the absence of importin α3 or α4 ( Fig 4G ) . However nuclear import of capsid proteins does not necessarily indicate proper nuclear capsid assembly . Consistent with an impairment of nuclear events during infection , the production of cell-associated infectious HSV-1 particles was reduced by one log for MEF-Impα1-/- , and delayed for MEF-Impα3-/- ( Fig 4H ) . Accordingly , the release of extracellular infectious virions was also delayed and reduced from MEF-Impα1-/- , and delayed from MEF-Impα3-/- when compared to MEFwt ( Fig 4I ) . To obtain further insights into capsid and virion assembly , we infected MEFwt ( Fig 5A ) or MEF-Impα1-/- ( Fig 5B ) with HSV ( 17+ ) Lox for 12 h , fixed them , and processed them for analysis by conventional electron microscopy . In both cell types , all known assembly intermediates had been formed: nuclear A , B and C capsids ( Fig 5Ai and 5Bi ) , primary enveloped virions between the inner and the outer nuclear envelope ( white star in Fig 5Ai ) , cytosolic capsids ( white arrowhead in Fig 5Aii and 5Bii ) , capsids in the process of secondary envelopment ( black arrowhead in Fig 5Aiii ) , intracellular vesicles harboring apparently intact virions ( black star in Fig 5Aii , 5Aiii , 5Aiv and 5Biii ) , and extracellular virions attached to the plasma membrane ( arrow in Fig 5Aiv and 5Biv ) . To quantify the amounts of these different assembly intermediates , we systematically evaluated entire cross sections of 10 randomly imaged cells for each cell line ( Table 1 ) . The amount of intracellular capsids per sampled area was reduced in MEF-Impα1-/- when compared to MEFwt . However , although there were fewer nuclear capsids the proportion of nuclear C capsids was increased . In contrast , while there were also fewer cytoplasmic capsids , the relative proportions of the different cytoplasmic capsids , such as cytosolic capsids , capsids in the process of being wrapped by cytoplasmic membranes , and enveloped capsids within transport vesicles was rather similar . Taken together these observations indicate that importin α1 is required for efficient nuclear capsid assembly and efficient capsid egress . However , those capsids that are translocated into the cytosol seem to associate with cytoplasmic membranes and to become enveloped to a similar extent to form infectious virions that are released from the infected cells also in the absence of importin α1 . Since importin α isoforms exhibit unique expression profiles in neurons [80] , we also investigated the role of importin α in post-mitotic primary neurons derived from the dorsal root ganglia ( DRG ) of adult mice . We have shown previously that such neurons are susceptible to productive HSV-1 infection [81–83] . We cultured DRG cells for 1 day , transduced them for 7 days with lentiviral vectors expressing an shRNA targeting importin α1 , importin α3 , importin α4 , or expressing a scrambled shRNA , and infected them then with HSV1 ( 17+ ) Lox-GFP . Immunoblotting showed that the expression of the respective importin α isoforms as well of the late tegument protein VP22 was clearly reduced in the DRG cultures when compared to the loading control p150Glued , a subunit of the dynein cofactor dynactin ( Fig 6A ) . We then used confocal fluorescence microscopy to limit our analysis to neurons identified by their typical morphology , their DNA staining pattern ( Fig 6Bi–6Fi ) , and expression of the neuronal β-tubulin-III ( [83]; see also Fig 7 below ) . Neurons expressing scrambled shRNA were well infected as indicated by a prominent HSV-1 mediated expression of GFP ( Fig 6Bii ) . In contrast , there was no GFP detected upon infection in the presence of nocodazole ( Fig 6Cii ) , silencing importin α1 ( Fig 6Dii ) , or silencing importin α3 ( Fig 6Eii ) , while silencing importin α4 did not impair GFP expression ( Fig 6Fii ) . Quantification showed that the levels of nuclear GFP were very heterogeneous among individual neurons and as strongly inhibited in the absence of importin α1 or importin α3 as in the presence of nocodazole ( Fig 6G ) . We focused the subsequent experiments on the role of neuronal importin α1 , since silencing importin α3 often induced changes of the chromatin architecture ( arrow in Fig 6Ei ) . Neurons transduced for shRNA expression were inoculated with HSV-1 in the presence of cycloheximide , fixed at 2 . 5 hpi , labeled with antibodies against capsids , stained for DNA , and analyzed by confocal fluorescence microscopy . Incoming HSV-1 capsids were as efficiently targeted to the nuclei ( Fig 7Ai and 7Ci ) of neurons expressing a scrambled shRNA ( Fig 7Aii ) as after silencing importin α1 ( Fig 7Cii ) . In contrast , nocodazole treatment reduced the number of incoming capsids reaching the neuronal nuclei ( Fig 7Bii ) . Since importin α can contribute to retrograde axonal transport of some cargos [84–87] , we also cultured DRG neurons in microfluidic chambers to selectively inoculate the neurons via the axons and not via the plasma membrane of the cell bodies for 4 h . However , nuclear targeting of HSV-1 capsids that in this experimental set-up was strictly dependent on axonal transport was as efficient in neurons expressing a scrambled shRNA ( Fig 7Dii ) as in neurons silenced for importin α1 expression ( Fig 7Eii ) . To further assess later stages of the HSV-1 life cycle , we infected neurons with HSV1 ( 17+ ) Lox-GFP for 4 h , and labeled them for DNA , ICP4 and capsids . Neurons expressing the scrambled shRNA were well infected as indicated by nuclear targeting of ICP4 ( Fig 8Aii ) , expression of the reporter GFP ( Fig 8Aiii ) , and nuclear and cytoplasmic progeny capsids ( Fig 8Aiv ) . In contrast , there was little expression of ICP4 ( Fig 8Cii ) or of GFP ( Fig 8Ciii ) , and only incoming capsids were detected at the nuclear rims ( Fig 8Civ ) after importin α1 expression had been silenced . When the neurons had been infected in the presence of nocodazole , the incoming capsids were rather distributed over the cytoplasm than at the nuclear rims ( Fig 8B ) . A quantitation of these signals in more than 50 neurons revealed that silencing importin α1 had reduced ICP4 ( Fig 8D ) and GFP ( Fig 8E ) expression and also the formation of nuclear capsid assembly compartments ( Fig 8F ) almost as efficiently as the nocodazole treatment . These experiments show that in primary neurons nuclear ICP4 expression , HSV1-mediated GFP expression , VP22 expression , and the formation of nuclear capsid assembly compartments depended on importin α1 .
The production of infectious cell-associated and extracellular virions was delayed and nuclear targeting of ICP4 , ICP0 , ICP8 and the DNA polymerase subunit pUL42 impaired in the MEF cells lacking importin α1 or importin α3 . In contrast , nuclear targeting of incoming capsids as well as nuclear import of VP16 and the HSV-1 genomes seemed not to be affected . Although we could not test this directly since we lack sufficiently powerful antibodies , we suppose that HCF-1 had co-imported VP16 into the nucleus , and together with other nuclear host transcription factors such as Oct-1 , SP1 and GABP initiated immediate-early transcription . The nuclear functions of HCF-1 are essential for cell viability , as regulatory processes controlled by this critical transcription factor do not operate properly , when HCF-1 is sequestered experimentally to the cytosol [96] . Consistent with this assumption , we detected similar expression levels of the immediate-early protein ICP4 by immunoblot in the different MEF lines . However , the nuclear import of ICP4 and another immediate-early protein ICP0 was severely impaired without importin α1 or importin α3 . Based on the coordinated interdependent and temporally regulated HSV-1 expression program reported in other systems [30 , 31 , 33] , we expected that reducing the nuclear amounts of ICP4 and ICP0 would delay subsequent steps of the HSV-1 life cycle . Yet , expression of the early and late proteins ICP8 , VP16 , and VP22 was not or only moderately reduced in MEFs lacking importin α1 or importin α3 , and even increased in the absence of importin α4 . Although HSV-1 gene expression seemed rather unperturbed , the nuclear import of the ssDNA binding protein ICP8 and the DNA polymerase processivity factor pUL42 were reduced in the absence of importin α1 or importin α3 . The two DNA polymerase subunits pUL30 and pUL42 of HSV-1 rely on several mechanisms for nuclear import , and can be imported individually or as a holoenzyme ( reviewed in [56] ) . HSV1-pUL30 comprises a non-canonical and a classical bipartite NLS , and binds to importin α5 , but other importin α isoforms have not been tested [97–99] . A bipartite NLS in HSV1-pUL42 has been shown to bind to importin α7 and to some extent to importin α1 but actually not to importin α3 [100]; nevertheless its nuclear import was reduced in the absence of importin α1 or importin α3 . pUL30 and pUL42 with mutated NLSs are still efficiently imported and targeted to the DNA replication compartments when co-expressed with the wild-type version of the other , but the holoenzyme is retained in the cytosol when the NLSs on both subunits are mutated [100] . Thus , it is possible that the lowered amounts of nuclear ICP8 were sufficient to sustain some DNA replication by a nuclear pUL30 despite reduced amounts of its accessory factor pUL42 . Importin α1 was a hit in our targeted RNAi screen for HSV1-mediated GFP expression; possibly because the nuclear HSV1 DNA replication had been reduced . Furthermore , the nuclear import of one of the host factors NF-ΚB , CREB/ATF , AP-1 , or SP1 that bind to the major immediate-early promotor of murine cytomegalovirus controlling GFP expression in our reporter virus might have been impaired [101 , 102] . Although immediate-early , early and late HSV-1 proteins had been synthesized , the electron microscopy analysis shows that the assembly of nuclear capsids , and thus the overall amount of capsids was significantly reduced in the absence of importin α1 . Furthermore , the targeting of the HSV-1 pUL31/pUL34 nuclear export complex to the inner nuclear membrane ( reviewed in [103 , 104] ) might have been impaired , leading to the reduced nuclear egress of progeny capsids , and the reduced amount of cytoplasmic capsids . Consistent with an overall reduced nuclear targeting of important HSV-1 proteins , a reduced formation of nuclear capsids , and a reduction in nuclear egress , the production of infectious HSV-1 virions was delayed but not prevented in MEF-Impα1-/- , and to some extent also in MEF-Impα3-/- . The specific requirement for importin α3 over importin α4 is remarkable , considering that their amino acid sequences are to 86% identical and to 92% conserved , and considering that importin α4 might even restrict certain steps of the HSV-1 replication cycle . It may nevertheless be possible that when one importin α is missing , the HSV-1 proteins could utilize another importin α homolog . In the differentiated , post-mitotic neurons , HSV-1 infection depended even more on importin α1 and importin α3 . When importin α1 expression had been reduced by RNAi , the amounts of ICP4 , HSV-1 mediated GFP , VP22 , as well as the formation of nuclear capsid assembly compartments were reduced , while nuclear targeting of incoming capsids was not inhibited irrespective of an inoculation via the somal plasma membrane or the axons . The distribution of importin α isoforms is highly regulated in different cell types and during development ( reviewed in [3 , 5 , 6] ) . During neuronal differentiation , expression changes from being initially high in importin α1 and low in importin α3 and importin α5 to low in importin α1 and high in importin α3 and importin α5 [105] . The importin α repertoire of post-mitotic neurons might be more limited than that of MEFs , and therefore silencing the expression of importin α1 or importin α3 had a stronger impact on HSV-1 infection in neurons . Having available the novel knock-out mice [73 , 106] , MEF lines lacking specific importin α isoforms [2 , 12 , 73] , and shRNA lentiviral vectors targeting specific importin α isoforms without influencing the expression of other importin α isoforms , we could validate antibodies specific for particular importin α isoforms or subfamilies . While importin α1 has been considered the general nuclear transport factor for cargoes with a classical NLS [2] , we and others could generate knock-out mice for specific importin αs suggesting that their host functions could be compensated at least to some extent [73 , 106] . Our study contributes to elucidating the mode of importin α isoform specificity in vivo that is so far only understood for a limited number of cargoes ( reviewed in [5] ) . Furthermore , not all binding reactions of a substrate to an importin α result in nuclear import of this substrate; for example , Oct-6 can bind to multiple importin-α isoforms , but while binding to importin α1 causes retention in the cytoplasm , binding to importin α5 results in nuclear import [11] . It will be interesting to determine , whether other alphaherpesviruses , betaherpesviruses , and gammaherpesviruses depend on the same importin α isoforms for viral protein import into the nucleus , capsid assembly , and capsid egress to the cytoplasm . Since the early years of the nuclear transport field , the interaction of viral proteins with import factors has been studied , and in several proteins of the herpesviruses and also other viruses replicating in the nucleus , NLS motifs recruiting specific import factors have been identified ( for review see [5 , 56] ) . Interestingly , the polymerase subunit PB2 of avian influenza A virus strains , an RNA virus replicating in the nucleus , preferentially binds to importin α3 , while mammalian adapted strains prefer importin α7 , and this switch might be a virulence factor in avian-mammalian host adaptation [107] . Other viruses actually do not utilize but disarm specific importin α isoforms . The structural protein VP24 of Ebola virus and the polymerase of hepatitis B virus block the nuclear import of STAT1 , and thus interferon signaling by competitive binding to importin α5 [108–110] . Although the exact intracellular concentration of different nuclear transport factors is hard to measure in situ , it will be interesting to determine to which extent the specific importin isoforms are expressed in epithelial cells , fibroblasts , neurons , and immune cells that are targeted by HSV-1 and other herpesviruses . In future work , it may be possible to reduce expression of all isoforms of one importin α subfamily in cell lines or in primary cells derived from tissues of these knock-out mice in order to reveal potentially redundant virus-host interactions . Further binding studies using recombinant HSV-1 proteins and limiting and competing amounts of different importins will dissect whether herpesvirus proteins comprise additional binding determinants that provide preferential specificity for importin α1 and importin α3 in addition to the already known NLSs . Finally , herpesviruses may also utilize NLSs of tegument proteins , e . g . the one in the N-terminal part of pUL36 , or in capsid proteins exposed on the surface of the incoming capsids to recruit specific importin α isoforms and importin β for capsid targeting to the nuclear pores for genome release into the nucleoplasm .
All cell lines were maintained as adherent cultures in a humidified incubator at 37°C and 5% CO2 and passaged twice per week . BHK 21 cells ( ATCC CCL-10 ) and Vero-D6 . 1 expressing HSV1-gB ( Helena Browne , University of Cambridge , personal communication; [78] ) were maintained in minimum essential medium ( MEM; Cytogen , Wetzlar , Germany ) supplemented with 10% ( v/v ) FCS ( PAA Laboratories GmbH , Cölbe , Germany; Life Technologies Gibco ) and Vero cells ( ATCC CCL-81 ) in MEM supplemented with 7 . 5% FCS . HeLaCNX cells [62] , human embryonic kidney cells ( HEK293T , ATCC CRL-11268; [111] ) and mouse embryonic fibroblasts ( MEFs ) derived from wild type ( MEFwt ) , MEF-Impα1-/- from importin α1-/- , MEF-Impα3-/- from importin α3-/- , and MEF-Impα4-/- from importin α4-/- [73] C57Bl/6 mice were cultured in Dulbecco’s modified Eagle’s medium ( DMEM ) -GlutaMAX-I ( Life Technologies Gibco , Darmstadt , Germany ) supplemented with 10% ( v/v ) FCS . Cells from DRG of adult C57Bl/6JHanZtm mice were cultured using established protocols [83 , 112–114] . The mice strain C57Bl/6JHanZtm ( not genetically modified ) were bred and maintained without any perturbation . On the day of the experiment , they were taken up from the animal facility , within 3 hours sedated with CO2-inhalation prior to killing by cervical dislocation without any prior experimental perturbation , and the DRG from the cervical , thoracic and lumbar levels of 3 to 4 mice were dissected afterwards . Those DRG were pooled in 1x HBSS-complete buffer ( Hank’s balanced salt solution , pH 7 . 4 with 5 mM HEPES and 10 mM D-Glucose ) , incubated with 20 mg/mL papain ( Sigma-Aldrich; in 0 . 4 mg/mL L-Cysteine , 0 . 5 mM EDTA , 1 . 5 mM CaCl2xH2O , pH 7 . 4 ) for 20 min at 37°C , with 10 mg/mL collagenase IV ( Invitrogen ) and 12 mg/mL dispase II ( Sigma-Aldrich ) for another 20 min at 37°C , and then triturated using Pasteur pipettes with narrowed ends . The cells were sedimented through 20% ( v/v ) Percoll ( Sigma-Aldrich ) cushions in CO2-independent medium ( Life Technologies Gibco , Carlsbad , CA , USA ) containing 10 mM D-glucose , 5 mM HEPES , 10% FCS , 100 U/mL penicillin and 0 . 1 mg/mL streptomycin , suspended in Ham’s F-12 nutrient mix medium with 10% FCS , 50 ng/mL 2 . 5S nerve growth factor ( Promega Corporation , Fitchburg , WI , US ) , 100 U/mL penicillin and 0 . 1 mg/mL streptomycin , and seeded onto cover slips of 20 mm diameter in 24-well plates or into microfluidic devices ( SND 450 , Xona Microfluidics , LLC , Temecula , CA , USA ) attached to 24 x 32 mm cover slips . The cover slips had been pre-coated with 0 . 01% ( w/v ) poly-L-lysine ( Sigma-Aldrich ) and 7 ng/μl murine laminin ( Invitrogen ) . The cells were cultured at 37°C and 5% CO2 in a humidified incubator , and the media were replaced twice a week . The mitosis inhibitor 1-β-D-arabinofuranosylcytosine ( Sigma-Aldrich ) was added at 1 to 2 div to a final concentration of 2 μM to suppress proliferation of dividing , non-neuronal cells , but removed at 4 div prior to HSV-1 infection . We used HSV1 ( 17+ ) Lox , HSV1 ( 17+ ) Lox-pMCMVGFP , or HSV1-GFP for short , which expresses soluble GFP under the control of the major immediate-early promoter of murine cytomegalovirus [62] , HSV1 ( 17+ ) Lox-CheVP26 , in which monomeric Cherry has been fused to the N-terminus of VP26 [76] , HSV1 ( 17+ ) Lox-CheVP26-UL37GFP [76] , and HSV1 ( 17+ ) Lox-ΔgB lacking the UL27 gene that encodes the essential glycoprotein gB [77] . Virus titers were assessed by plaque assays [115] , or for HSV1 ( 17+ ) Lox-ΔgB estimated by comparing an immunoblot analysis of extracellular viral particles to HSV1 ( 17+ ) Lox-pMCMVGFP expressing gB and used in parallel . For infection experiments , extracellular virus sedimented from the medium of infected BHK 21 cells was used [18 , 115] . The stocks of the different HSV-1 strains used for infection as well as the MEF-associated virus and the virus released from infected MEFs were titrated on Vero cells . At 4 , 8 , 12 , 16 and 20 hpi , the supernatants of infected MEF were collected and cleared by low-speed sedimentation , and the cells were scraped into 1 mL/well MNT buffer ( 30 mM MES , 100 mM KCl , 20 mM Tris , pH 7 . 4 ) and subjected to 3 cycles of freeze-thawing . Vero cells were cultured to just confluency in 6-well dishes , and incubated for 1 h at room temperature on a rocking platform with 10-fold serial dilutions of the different virus suspensions in CO2-independent medium ( Life Technologies Gibco ) with 0 . 1% [w/v] cell culture grade bovine serum albumin ( PAA Laboratories GmbH ) . The inoculum was removed and 2 mL/well growth medium containing 20 μg/mL pooled human IgG ( Sigma-Aldrich ) was added . The cells were incubated for 3 d , fixed in absolute methanol , and stained with 0 . 1% [w/v] crystal violet and 2% [v/v] ethanol in H2O . To stain DNA , we used 4’ , 6-diamidino-2-phenylindole ( DAPI; Roth ) or TO-PRO-3-iodide ( Life Technologies ) at final concentrations of 50 μg/mL or 1 to 2 μM , respectively . We used rabbit polyclonal antibodies ( pAbs ) raised against human importin α1 ( #70160 , Abcam ) , human importin α3 ( Enno 31; Pineda Antikörper Service , Berlin , Germany ) , human importin α4 ( Enno 32; Pineda Antikörper Service ) , human importin α5/α6/α7 ( MDC 220; [2] ) , HSV1-VP16 ( #631209 , BD Biosciences ) , HSV-1 tegumented capsids ( Remus , bleed V; [23] ) , or nuclear HSV-1 capsids . To generate a polyclonal serum directed against HSV-1 capsids ( SY4563 , anti-capsid ) , rabbits were immunized with purified nuclear capsids ( Kaneka Eurogentec S . A . , Seraing , Belgium ) . Mouse monoclonal antibodies ( mAb ) were directed against α-tubulin ( DM1A , Sigma-Aldrich ) , nuclear pore complexes ( mAb 414 , Abcam ) , actin ( mAb 1501 , Millipore ) , β-III-tubulin ( mAb 5564 , Millipore ) , p150Glued ( #610474 , BD Biosciences ) , HSV1-ICP0 ( mAb 11060 , sc-53070 , Santa Cruz Biotechnology ) , HSV1-ICP4 ( mAb 10F1 , ab6514 , Abcam ) , HSV1-ICP8 ( mAb 11E2 , ab20194 , Abcam ) , or HSV1-pUL42 ( ab19311 , Abcam ) . Secondary antibodies for immunoblotting were conjugated to fluorescent infrared dyes ( anti-rabbit IgG-IRDye 800CW , anti-mouse IgG-IRDye 680RD , LI-COR Biosciences ) , and for immunofluorescence microscopy to Cy3 ( goat-anti-rabbit IgG; Dianova ) , Cy5 ( goat anti-mouse IgG; Dianova ) , Alexa Fluor488 ( A488; goat anti-rabbit IgG; goat-anti-mouse IgG , Invitrogen ) or fluorescein isothiocyanate ( FITC; goat anti-rabbit IgG; Dianova ) . All secondary antibodies were highly pre-adsorbed to eliminate cross-reactivity to other species than the intended one . To silence importin α1 , importin α3 , or importin α4 by short hairpin RNAs ( shRNAs; Sigma Mission library; S3 Table ) or to express a non-mammalian shRNA control ( SHC002 , Sigma Mission library ) , we used lentiviral transduction . HEK 293T cells were transfected with 5 μg pRSVRev , 2 μg pMD2 . g ( Addgene Inc . , Cambridge , MA , USA , Cat . No . 12259 ) , 10 μg pCDNA3 . GP . CCCC , and 10 μg transfer plasmid per 10 cm dish as described previously ( [116]; plasmids provided by Axel Schambach ) . The supernatants were harvested at 36 and 48 h , and sedimented in a SW32 . Ti rotor at 24 , 000 rpm for 90 min at 4°C ( Beckman Coulter ) . The re-suspended lentiviral particles were snap frozen in liquid N2 and stored in single-use aliquots at -80°C . Cell culture supernatants and concentrated lentiviral stocks were titrated using a p24 ELISA [117] . MEFwt were transduced with lentiviral particles at 4 to 12 μg/mL p24 and at 1 dpt , selection with puromycin at 2 . 5 μg/mL was started . DRG cells were transferred after 1 day in vitro to neuronal growth media containing lentiviral particles at 4 to 12 μg/mL p24 but no AraC . After 2 dpt , the media were replaced by F12-complete with 2 μM AraC and 5 μg/mL puromycin to select for transduced cells . Small interfering RNAs ( siRNAs ) against human transport factors as well as scr siRNAs were from QIAGEN ( c . f . S1 Table; Hilden , Germany ) and the GFP silencer siRNAs from Ambion ( AM4626; Darmstadt , Germany ) . 3 , 500 to 4 , 000 HeLaCNX cells per well of 96-well plates were reverse transfected with 50 nM of siRNA using Lipofectamine 2000 ( Invitrogen , Life Technologies ) . After 3 days , cells were left untreated or pre-treated with 50 μM nocodazole for 1 h and infected with 4 x 106 PFU/mL of HSV1 ( 17+ ) Lox-pMCMVGFP for 12 h in the absence or presence of nocodazole . Cells were fixed with 3 . 4% paraformaldehyde ( PFA ) , permeabilized with 0 . 1% Triton-X-100 and stained with DAPI . DAPI and GFP fluorescence were measured using a fluorescence plate reader ( BioTek Synergy 2 , Bad Friedrichshall , Germany ) and the GFP background signal of the mock infected cells was subtracted . To allow comparison of different experiments , the median values of cells transfected with scr siRNAs of each experiment were set as 100% and GFP/well and DAPI/well values were calculated . To reduce the impact of potential off-target effects introduced by miRNAs binding the siRNA seed region , the results were corrected using a dataset of seed region phenotypes [66] . The seed regions of siRNAs classified by Franceschini et al . ( 2014 ) to result in off-target effects were compiled , and the mean of significantly altered seed region phenotypes were determined using a threshold of p< = 0 . 05 after Bonferroni correction [118] . Franceschini et al . ( 2014 ) propose an additive model with the seed phenotype contributing with a factor of 0 . 6 to the overall gene expression results . This adjusted seed phenotype was subtracted from the gene expression results , and the medians of GFP/well , GFPcorr/well or DAPI/well respectively were determined ( c . f . S1 Table , GFP , GFPcorr , DAPI ) . To normalize for potential effects of RNAi on cell density , GFPcorr/DAPI ratios were determined for each well , and the median from the single values was calculated ( c . f . S1 Table , GFPcorr/DAPI ) . For immunofluorescence microscopy , immunoblot analysis and viral growth curves , MEFs were seeded onto coverslips in 24-well plates at densities of 1 x 105 cells/well or into 6-well dishes at 2 . 5 x 105 cells/well , and on the next day pre-cooled and inoculated with HSV-1 in CO2-independent medium with 0 . 1% ( w/v ) cell culture grade bovine serum albumin ( BSA; PAA Laboratories GmbH ) . MEFs were inoculated for 1 h on ice for nuclear targeting assays , for 0 . 5 h on ice for measuring nuclear import of viral genomes and VP16 , and for 2 h at RT for measuring viral gene expression by immunoblot and measuring nuclear import of viral proteins by confocal fluorescence microscopy . DRG cells were inoculated for 0 . 5 h at RT . After washing off the unbound virions , the cells were shifted to growth medium at 37°C for the indicated times . We used 5 x 107 pfu/mL ( MOI of 100 ) of HSV1 ( 17+ ) Lox-CheVP26 to analyze nuclear targeting of incoming HSV-1 capsids , 1 x 108 pfu/mL ( MOI 200 ) of HSV1 ( 17+ ) Lox-pMCMVGFP or of HSV1 ( 17+ ) Lox–ΔgB to study the subcellular localization of incoming VP16 , 1 x 108 pfu/mL ( MOI of 200 ) of HSV1 ( 17+ ) Lox-CheVP26-UL37GFP to examine the nuclear import of incoming viral genomes , 0 . 5 to 1 . 25 x 106 pfu/mL ( MOI 2 to 5 ) of HSV1 ( 17+ ) Lox-CheVP26 to examine the synthesis of structural HSV-1 proteins by immunoblot , 0 . 5 to 1 . 25 x 106 pfu/mL ( MOI 2 to 5 ) of HSV1 ( 17+ ) Lox-CheVP26 to determine the subcellular localization of the immediate-early proteins ICP4 and ICP0 , the early proteins ICP8 , and the late structural protein CheVP26 . For the virus growth curves , the different MEF lines were infected with 1 . 3 x 106 pfu/mL ( MOI 5 ) of HSV1 ( 17+ ) Lox at a reduced level of 1% [v/v] FCS . Primary cells derived from the DRGs were infected with 2 . 5 x 107 pfu/mL for nuclear capsid targeting , with 5 x 106 pfu/mL for gene expression upon infection from the somal plasma membrane , or with 1 . 3 x 108 pfu/mL for nuclear capsid targeting upon infection from the axonal compartment in microfluidic chambers . In those experiments analyzing the subcellular localization of incoming HSV-1 capsids , incoming VP16 or incoming viral genomes , 0 . 5 mM cycloheximide ( Sigma-Aldrich ) was added to prevent synthesis of new viral proteins [18] . When nocodazole ( 25 or 50 μM for MEFs , 10 μM for neurons; Sigma-Aldrich ) was used to depolymerize microtubules , cells were pretreated for 1 h at 37°C , and the drug was present during all further steps . Cells were lysed in hot sample buffer ( 1% [w/v] SDS , 50 mM Tris-HCl , pH 6 . 8 , 1% [v/v] β-mercaptoethanol , 5% [v/v] glycerol bromophenol blue ) containing a protease inhibitor cocktail ( cOmplete Roche , #11873580001 ) , and the DNA was sheared using 20-gauge needles . The lysates were loaded onto linear 5 to 12% gradient or 10% SDS gels , and proteins were transferred to nitrocellulose membranes . Membranes were incubated with a blocking solution of 5% [w/v] low-fat milk in PBS followed by incubation with primary antibodies in blocking solution , washed with PBS containing 0 . 1% [w/v] Tween-20 and 0 . 5% milk , incubated with secondary antibodies in blocking solution , washed and scanned ( Odyssey Infrared Imaging System , LI-COR Biosciences , NE , USA ) . The band areas and mean intensities were measured using a rectangular selection tool to calculate the integrated intensity ( ImageJ version 1 . 50e , NIH , USA ) . The background was subtracted , the integrated intensities were normalized to untreated MEFwt , and the ratios of the respective viral protein to actin used as loading control were calculated . Infected cells were either simultaneously fixed and permeabilized with PHEMO-fix ( 68 mM PIPES , 25 mM HEPES , 15 mM EGTA , 3 mM MgCl2 , 10% [v/v] DMSO , 3 . 7% [w/v] PFA , 0 . 05% [v/v] glutaraldehyde , 0 . 5% [v/v] Triton X-100 , pH 6 . 9 ) for 10 min at 37°C and washed twice with PHEMO buffer ( 68 mM PIPES , 25 mM HEPES , 15 mM EGTA , 3 mM MgCl2 , 10% [v/v] DMSO , pH 6 . 9 ) , or fixed with 3% [w/v] PFA in PBS for 20 min at room temperature as described before [18 , 19] . Fixed cells were treated with 50 mM NH4Cl/PBS for 10 min , and permeabilized with 0 . 1% Triton X-100/PBS for 5 min in the case of PFA fixation . The HSV1-Fc receptor [119] and other unspecific antibody binding were blocked with 0 . 5% ( w/v ) BSA and 10% ( v/v ) serum from HSV1-negative volunteers . After the immunolabelling , the samples were embedded in Mowiol containing 10% [w/v] 1 , 4-diazabicyclo[2 . 2 . 2]octane , and imaged with plan-apochromatic 63x oil-immersion objectives with a numerical aperture of 1 . 4 with a confocal fluorescence microscope ( LSM 510 Meta; Carl Zeiss Microscopy , Jena , Germany; TCS SP6 , LEICA Microsystems , Wetzlar , Germay ) . Contrast and brightness were adjusted identically across each set of images ( Adobe Photoshop version 6 . 0 or version CS4 ) . Figures were assembled using Adobe Illustrator CC ( version 20 . 1 . 0 ) . To quantify the nuclear accumulation of ICP4 , ICP0 , ICP8 , pUL42 , GFP or capsid proteins , we developed a pipeline using the CellProfiler software ( [120]; http://cellprofiler . org/; BI-2013-070 , version 2 . 1 . 1 , NIH , USA ) that segmented the nuclei based on DAPI fluorescence and size , and then determined the mean fluorescence intensity of the labeling for the above mentioned proteins . To measure the number of capsids at the nuclear rim of neurons , nuclear corridors around the outer rim of the segmented nuclei were defined by both expanding and shrinking the nuclear area by several pixels . Then the number of capsids localized within that area was counted . Thresholds for the recognition of the capsid signal were based on the typical signal intensity and size of capsids and considering the background intensity of the anti-capsid antibody in uninfected neurons . For each protein , the average grey values per nuclei were calculated to compile box and whisker plots . The p values were determined with a Kruskal-Wallis test followed by Dunn’s multiple comparison testing ( software Prism , version 6; Graphpad , San Diego , CA , USA ) . To analyze the subcellular distribution of incoming HSV-1 genomes , cells were infected as described above and fixed with a mixture of 95% ethanol and 5% acetic acid , and processed for fluorescent in situ hybridization ( FISH ) . HSV-1 probe synthesis and hybridization were performed as described previously [121 , 122] using a HSV1 ( 17+ ) Lox-ΔUL36 genome cloned into a bacterial artificial chromosome [123] to generate a Cy3-labelled DNA probe . For detection of incoming HSV-1 genomes , the DNA probe was used at 20 μg per coverslip , and the samples were analyzed by confocal fluorescence microscopy . MEFwt or MEF-Impα1-/- seeded on glass cover slips were infected with HSV1 ( 17+ ) Lox with an MOI of 10 pfu/cell at 2 . 5 x 106 pfu/mL . The cells were fixed at 12 hpi with 2% glutaraldehyde and 2 . 5% formaldehyde in cacodylate buffer [130 mM ( CH3 ) 2AsO2H , pH 7 . 4 , 2 mM CaCl2 , 10 mM MgCl2] for 1 h at room temperature . Cells were contrasted with 1% ( w/v ) OsO4 in cacodylate buffer ( 165 mM ( CH3 ) 2AsO2H , pH 7 . 4 , 1 . 5% ( w/v ) K3[Fe ( CH ) 6] ) followed by 0 . 5% ( w/v ) uranyl acetate in 50% ( v/v ) ethanol overnight . The cells were embedded in Epon plasticServa , Heidelberg , Germany ) and 50 nm ultrathin sections were cut parallel to the substrate . Images were acquired with a Morgani transmission electron microscope ( FEI , Eindhoven , The Netherlands ) at 80 kV . Viral structures were counted and sectioned nuclear and cytoplasmic areas were measured using Fiji software ( fiji . sc ) . Human sera of exclusively adult , healthy , HSV-1 seronegative volunteers were obtained after written informed consent by the blood donors . Permission was granted by the Institution Review Board ( Hannover Medical School; Approval Number 893 ) . According to the German Animal Welfare Law §4 , killing of animals needs no approval , if the removal of organs serves scientific purposes , and if the mice had not undergone experimental treatment before . The animal care and sacrifices were performed in strict accordance with the German regulations of the Society for Laboratory Animal Science ( GV-SOLAS ) , the European Health Law of the Federation of Laboratory Animal Science Association ( FELASA ) and the German Animal Welfare Law . This study here does not contain animal experiments that require pre-approval , and the total number of killed mice was reported at the end of each year to the animal welfare deputy of Hannover Medical School . This information was registered annually as the number of animals killed according to §4 of the German Animal Welfare Law and the number of killed mice was registered with the animal welfare application number 2012/20 at the local state authority ( LAVES; Niedersächsisches Landesamt fuer Verbraucherschutz und Lebensmittelsicherheit , Oldenburg , Germany ) .
|
Nuclear pore complexes are highly selective gateways that penetrate the nuclear envelope for bidirectional trafficking between the cytoplasm and the nucleoplasm . Viral and host cargoes have to engage specific transport factors to achieve active nuclear import and export . Like many human and animal DNA viruses , herpesviruses are critically dependent on many functions of the host cell nucleus . Alphaherpesviruses such as herpes simplex virus ( HSV ) cause many diseases upon productive infection in epithelial cells , fibroblasts and neurons . Here , we asked which nuclear transport factors of the host cells help HSV-1 to translocate viral components into the nucleus for viral gene expression , nuclear capsid assembly , capsid egress into the cytoplasm , and production of infectious virions . Our data show that HSV-1 requires the nuclear import factor importin α1 for efficient replication and virus assembly in fibroblasts and in mature neurons . To our knowledge this is the first time that a specific importin α isoform is shown to be required for herpesvirus infection . Our study fosters our understanding on how the different but highly homologous importin α isoforms could fulfill specific functions in vivo which are only understood for a very limited number of host and viral cargos .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"nuclear",
"import",
"gene",
"regulation",
"cell",
"processes",
"microbiology",
"light",
"microscopy",
"viral",
"structure",
"neuroscience",
"virus",
"effects",
"on",
"host",
"gene",
"expression",
"microscopy",
"research",
"and",
"analysis",
"methods",
"small",
"interfering",
"rnas",
"animal",
"cells",
"fluorescence",
"microscopy",
"gene",
"expression",
"viral",
"packaging",
"viral",
"replication",
"virions",
"biochemistry",
"rna",
"cellular",
"neuroscience",
"cell",
"biology",
"nucleic",
"acids",
"virology",
"neurons",
"genetics",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"non-coding",
"rna"
] |
2018
|
Importin α1 is required for nuclear import of herpes simplex virus proteins and capsid assembly in fibroblasts and neurons
|
Through high coverage whole-genome sequencing and imputation of the identified variants into a large fraction of the Icelandic population , we found four independent signals in the low density lipoprotein receptor gene ( LDLR ) that associate with levels of non-high density lipoprotein cholesterol ( non-HDL-C ) and coronary artery disease ( CAD ) . Two signals are novel with respect to association with non-HDL-C and are represented by non-coding low frequency variants ( between 2–4% frequency ) , the splice region variant rs72658867-A in intron 14 and rs17248748-T in intron one . These two novel associations were replicated in three additional populations . Both variants lower non-HDL-C levels ( rs72658867-A , non-HDL-C effect = -0 . 44 mmol/l , Padj = 1 . 1 × 10−80 and rs17248748-T , non-HDL-C effect = -0 . 13 mmol/l , Padj = 1 . 3 × 10−12 ) and confer protection against CAD ( rs72658867-A , OR = 0 . 76 and Padj = 2 . 7 × 10−8 and rs17248748-T , OR = 0 . 92 and Padj = 0 . 022 ) . The LDLR splice region variant , rs72658867-A , located at position +5 in intron 14 ( NM_000527:c . 2140+5G>A ) , causes retention of intron 14 during transcription and is expected to produce a truncated LDL receptor lacking domains essential for function of the receptor . About half of the transcripts generated from chromosomes carrying rs72658867-A are characterized by this retention of the intron . The same variant also increases LDLR mRNA expression , however , the wild type transcripts do not exceed levels in non-carriers . This demonstrates that sequence variants that disrupt the LDL receptor can lower non-HDL-C and protect against CAD .
The low-density lipoprotein receptor ( LDLR ) is a cell-surface receptor responsible for binding and uptake of circulating cholesterol-containing lipoprotein particles . This uptake is the primary pathway for removal of cholesterol from the circulation [1] . It is well established that high levels of low-density lipoprotein-cholesterol ( LDL-C ) are a key risk factor for coronary artery disease ( CAD ) and is a primary target for therapeutic intervention [2] . Recent studies show that non-high density lipoprotein cholesterol ( non-HDL-C ) is a better predictor for cardiovascular risk than LDL-C as it encompasses all cholesterol containing pro-atherogenic lipoproteins such as very low-density lipoprotein ( VLDL ) , intermediate-density lipoprotein ( IDL ) , chylomicron remnants ( CR ) as well as LDL-C [3] . LDL receptors primarily clear LDL-C from blood but they also bind VLDL-C and remnant particles [4] . The LDL receptor and its role in LDL-C regulation was discovered 40 years ago when Goldstein and Brown set out to unravel the cause of familial hypercholesterolemia ( FH ) [5] , a severe autosomal dominant disorder characterized by high levels of LDL-C in blood and premature cardiovascular disease [6] . The most common sequence variants causing FH are loss-of-function LDLR mutations that disrupt the receptor function leading to reduced hepatic LDL-C clearance and elevated plasma LDL-C . So far over 1 , 200 rare LDLR mutations have been reported in FH families [7 , 8] . Common variants at the LDLR locus with modest effects on LDL-C levels and risk of coronary artery disease ( CAD ) in the general population have been identified through genome-wide association studies ( GWAS ) [9–11] . More recently GWAS studies based on whole-exome sequencing have confirmed the association between very rare LDLR missense and loss-of-function variants ( MAF <1% ) with LDL-C levels and risk of myocardial infarction ( MI ) [12 , 13] . The design of these studies , however , had little capacity to detect rare and low frequency non-coding variants at the LDLR locus that affect cholesterol levels and the risk of CAD and MI . High-coverage whole-genome sequencing ( WGS ) based GWAS in contrast have the potential to identify such variants if present . Here we applied high-coverage WGS to a large fraction of the Icelandic population to search for LDLR sequence variants affecting serum levels of non-HDL-C in the general population . We found four highly significant variants each representing independent signals at the LDLR locus that associate with levels of non-HDL-C and risk of CAD . Two of these associations are novel and represented by non-coding variants of low frequency that lower non-HDL-C levels and protect against CAD . One of them affects splicing of the LDLR that is predicted to truncate the receptor .
In our study we elected to use the measurement non-HDL-C instead of LDL-C as it encompasses all potential atherogenic cholesterol containing lipoproteins including LDL-C . We examined the association of 7 , 351 sequence variants in a 1 Mb region centered on LDLR ( chr19:10 , 559 , 187–11 , 559 , 187 ( NCBI build36/hg18 ) ) with non-HDL-C levels in 119 , 146 Icelanders . These sequence variants ( SNPs and INDELs ) were identified by WGS of 2 , 636 Icelanders and imputed , assisted by long-range phased haplotypes , into 104 , 220 Icelanders genotyped with Illumina SNP arrays [14] . In addition , we used genealogical information to calculate genotype probabilities for 294 , 212 first and second degree relatives of array genotyped individuals[15] . After performing stepwise conditional analysis we identified four highly significant variants each representing an independent signal at the LDLR locus ( Fig 1 and Table 1 and S1 Table ) . Two of the variants are non-coding and low frequency and are novel with respect to association with non-HDL-C , rs72658867-A , a splice region variant at position +5 in intron 14 of LDLR ( NM_000527 . 4:c . 2140+5G>A , minor allelic frequency ( MAF ) = 2 . 2% ) , and rs17248748-T , a variant in the first intron of LDLR ( NM_000527 . 4:c . 68-4859C>T , MAF = 3 . 4% ) , that lower non-HDL-C by 0 . 44 mmol/l ( Padj = 2 . 0 × 10−70 ) and 0 . 13 mmol/l ( Padj = 5 . 0 × 10−11 ) , respectively . The splice region variant rs72658867-A has been described in FH families , however , it has been disputed whether it is pathogenic [16–19] . The third signal is captured by a common variant rs17248720-T ( NM_000527 . 4:c . -2038C>T , MAF = 8 . 8% ) located at the 5’ end of LDLR that lowers non-HDL-C by 0 . 24 mmol/l ( Padj = 1 . 8× 10−80 ) and has been reported to lower LDL-C levels with similar effect as shown here and confer protection against CAD [9 , 20] . The fourth signal at the LDLR locus , is represented by a rare variant rs200238879-C ( MAF = 0 . 06% ) , reported to be an Icelandic founder FH mutation [21] . This variant is located in the donor splice site of intron 4 ( NM_000527 . 4:c . 694+2T>C ) and increases non-HDL-C serum levels by 1 . 33 mmol/l ( Padj = 2 . 2 × 10−22 ) . The four variants associate with LDL-C with similar effect sizes as with non-HDL-C , the P-values are however slightly higher due to smaller sample size for LDL-C ( S2 Table ) . None of the variants associate with high-density cholesterol ( HDL-C ) or triglycerides except rs72658867-A , that associates weakly with increased HDL-C levels ( Padj = 0 . 0035 ) ( Table 1 ) . No combination of non-HDL-C lowering alleles ( minor alleles ) of rs72658867-A ( splice region variant ) , rs17248748-T ( intronic ) and rs17248720-T ( common ) occur on the same haplotype in our data ( S1 Fig ) . The splice donor variant rs200238879-C is very rare and is weakly correlated with the other three variants ( S3 Table ) . The non-HDL-C lowering effects of the two low frequency variants and the common variant are additive ( S2 Fig and S4 Table ) . Homozygous carriers of each of these variants have lower non-HDL-C levels than heterozygotes , with the lowest values observed for the homozygous carriers of the splice region variant ( rs72658867-A ) . We attempted to follow up the association of the two novel variants with non-HDL-C by direct genotyping in samples from Denmark , the Netherlands and Iran . In all three populations we replicate the association of both variants with lower non-HDL-C with similar effect sizes as in Iceland ( effect of rs72658867-A on non-HDL-C is -0 . 41 mmol/l , P = 1 . 2 × 10−11 and for rs17248748-T is -0 . 14 mmol/l , P = 0 . 0082 ) ( Table 2 and S5 Table ) . Joined with the Icelandic discovery data the combined effect on non-HDL-C for rs72658867-A is -0 . 44 mmol/l , P = 1 . 1 × 10−80 and for rs17248748-T is -0 . 13 mmol/l , P = 1 . 3 × 10−12 ( Table 2 and S5 Table ) . We tested the four LDLR variants for association with CAD in a sample of 33 , 090 cases and 236 , 254 controls from Iceland . All variants associate with CAD in a direction consistent with the known correlation between non-HDL-C and CAD ( Table 3 ) . The three non-HDL-C lowering variants , rs72658867-A ( splice region variant ) , rs17248748-T ( intronic ) and rs17248720-T ( common ) , all associate with a reduced risk of CAD and both rs72658867-A and rs17248720-T delay the age at diagnosis of CAD ( Table 3 ) . The rare splice donor variant that raises non-HDL-C , rs200238879-C , increases CAD risk and lowers the age at diagnosis by almost nine years ( Table 3 ) . Further , consistent with the effect on CAD , rs72658867-A and rs17248720-T both associate with increased lifespan ( 0 . 59 and 0 . 61 years per allele , respectively ) and the rare splice donor mutation rs200238879-C associates with decreased lifespan ( -6 . 46 years per allele ) ( S6 Table ) . In our WGS data the coverage at the LDLR locus was high ( ~20X ) apart from a small region ( 50bp ) in intron 1 of low coverage that was analysed separately in an independent set of individuals ( n = 738 ) whole-genome sequenced with the Illumina PCR-free sequencing method ( S3 Fig ) . Our dataset is thus likely to represent all sequence variants ( SNPs and INDELs ) at the LDLR locus that are present in the Icelandic population at a frequency over 0 . 1% , allowing for fine mapping and identification of the causative variants of the four LDLR signals [14] . For that purpose we explored all variants in the Icelandic dataset that correlated ( r2>0 . 8 ) with the four index variants in a 2 Mb window centered on each variant . The rare splice donor variant rs200238879-C is the most likely causative variant for that signal as it correlates with no other coding mutation and the mutation is known to cause abnormal splicing of the LDLR [21] . For the common upstream variant rs17248720 we found 56 correlates ( 46 variants with r2>0 . 99 ) none of which are in the LDLR coding region ( S7 Table ) . Three of the most strongly correlated SNPs are located in sequence motifs with strong promoter or enhancer activities in the liver cell line ( HepG2 ) ( HaploReg v3 , see URLs ) , suggesting that any one of them could be causative . The novel intronic variant ( rs17248748-T ) has no strong correlates . It is located in a sequence motif within intron 1 with strong enhancer activity in the HepG2 liver cell line and binds regulatory proteins , including c/EBPbeta known to regulate transcription of LDLR ( HaploReg v3 , see URLs ) , supporting a causative role of the variant [22 , 23] . For the splice region variant rs72658867-A we found only one correlate , rs180760728-C , an intronic variant in the LDLR gene ( MAF = 1 . 98% , r2 = 0 . 89 ) . Conditional analysis revealed that rs180760728-C does not account for the non-HDL-C association of rs72658867-A ( Padj = 0 . 32 for rs180760728 adjusting for rs72658867; Padj = 8 . 7×10−11 for rs72658867-A adjusting for rs180760728-C ) . Furthermore , in the 1000 Genomes European ancestry Phase 3 dataset ( see URLs ) rs180760728-C is more weakly correlated with rs72658867-A ( r2 = 0 . 49 ) and rs180760728 is the only marker with r2>0 . 3 with rs72658867-A , indicating that rs72658867-A is likely the causative variant for this signal . The replication of the non-HDL-C associations of rs72658867-A and rs17248748-T with similar effect sizes in the three distinct populations further supports a causative role for these variants . To further characterize the two novel non-HDL-C signals at the LDLR locus we analysed the effect of rs72658867 and rs17248748 on LDLR mRNA expression in a microarray mRNA expression dataset for white blood cells ( 1 , 001 individuals ) and adipose tissue ( 667 individuals ) . The non-HDL-C lowering allele of the splice region variant rs72658867 associates with increased LDLR mRNA expression in blood ( ~22% increase , P = 1 . 2 × 10−11 ) ( S4 Fig ) and no other variant in a ~1Mb region centered on LDLR correlated more strongly with LDLR expression than rs72658867 ( S8 Table ) . These findings were replicated in an independent RNA-sequencing dataset from blood ( 252 individuals ) where similar increase was detected in carriers ( P = 0 . 0075 ) ( Fig 2A ) . Using that dataset we also performed allele specific analysis of heterozygous carriers and non-carriers and show that the chromosomes carrying rs72658867-A have greater expression than the chromosomes carrying the reference allele ( S9 Table and S5 Fig ) . In heterozygous carriers of rs72658867-A about 60% of the transcripts are derived from the mutated chromosome compared to a baseline proportion of 0 . 52 in non-carriers . In the adipose tissue microarray dataset the correlation with LDLR expression is in the same direction but much weaker ( P = 0 . 020 , S4 Fig ) . No significant correlation with LDLR expression in blood or adipose tissue was found for the intronic rs17248748 variant ( P = 0 . 40 and P = 0 . 10 , respectively ) . The splice region variant rs72658867 is located at position 5 of intron 14 , a position that is conserved and could potentially affect splicing . To investigate this , we analysed the mRNA sequence data from blood and observed abnormal splicing in rs72658867-A carriers , characterized by retention of intron 14 ( i . e . transcription through intron 14 in the LDLR transcripts ) ( Fig 2B ) . When looking at the proportion of RNA sequencing reads that are spliced from exon 14 to exon 15 ( correctly spliced ) out of the total number of reads that cover the last base of exon 14 we observed a mean proportion of 0 . 95 in non-carriers ( n = 238 ) compared to 0 . 71 in heterozygous carriers ( n = 15 ) ( Mann-Whitney test for location shift P = 6 . 0×10−9 ) . This indicates that approximately 30% of the transcripts in heterozygous carriers are abnormally spliced ( Fig 2C ) . Analysis of blood RNA sequence data from homozygous carriers of rs72658867- A ( n = 3 ) demonstrated that about half of the LDLR transcripts are characterized by intron retention ( S10 Table ) . Together these data from the hetero- and homozygous carriers indicate that from chromosomes carrying rs72658867-A , about half of the transcripts are normally spliced and half of them are abnormal . This also indicates that even though the total amount of LDLR transcripts is increased by 22% in heterozygotes ( Figs 2A and S4 ) , the estimated amount of normally spliced LDLR transcripts will not exceed 90% of normal levels . This was confirmed by RT-PCR analysis of LDLR mRNA from blood that showed similar levels of wild type LDLR mRNA in heterozygous rs72658867-A/G carriers ( n = 20 ) and rs72658867-G/G non-carriers ( n = 343 ) , ( Mann-Whitney test: P = 0 . 87 ) ( S6 Fig ) . The non-HDL-C lowering effect of rs72658867-A is thus not mediated by a net increase in the wild type LDLR transcripts . The retention of intron 14 alters the LDLR reading frame after amino acid position 713 ( end of exon 14 , NP000518:p . Thr713fsTer33 ) such that 33 amino acids are added until a premature stop codon is reached . It is unlikely that the introduction of the premature stop codon renders the transcript susceptible to nonsense-mediated decay as a high fraction of the LDLR transcripts in heterozygotes and homozygotes are characterized by retention of intron 14 ( Fig 2C and S10 Table ) . The abnormally spliced mRNA is predicted to produce a truncated LDLR lacking the O-linked glycan region and the transmembrane and cytoplasmic domains ( S7 Fig ) . The transmembrane domain anchors the LDLR in the lipid bilayer and endocytosis and intracellular transport of the LDLR are regulated via its cytoplasmic domain [24] but the role of the O-linked glycan region is unclear[25] .
We have identified , by high coverage whole-genome sequencing and subsequent imputation into a large fraction of the Icelandic population , four independent sequence variants at the LDLR locus that associate with levels of non-HDL-C and risk of CAD in the general population . Two of them are of low frequency and novel with respect to non-HDL-C association: a splice region variant ( rs72658867-A , c . 2140+5G>A ) and an intronic variant ( rs17248748-T ) . Both variants associate with lowering of non-HDL-C and protection against CAD . We show that the splice region variant causes retention of intron 14 altering the LDLR reading frame after amino acid position 713 such that 33 amino acids are added until a premature stop codon is reached . This splicing defect affects about half of the transcripts generated from the chromosome carrying the variant . The same variant also increases mRNA expression of LDLR that includes both normally and abnormally spliced transcripts and this increase seems to be driven by the chromosome carrying rs72658867-A . This study highlights the importance of including non-coding variants , in all segments of the frequency spectrum ( common , low frequency and rare ) , in GWAS . Although the abnormal transcript is predicted to translate into a truncated LDLR lacking domains essential for receptor function , the splice region variant associates with a strong non-HDL-C lowering effect and protection against CAD in the general population . These data contrast LDLR truncating mutations that lead to an increase in non-HDL-C because of reduced function of the LDLR [24–26] . The evidence for the non-HDL-C lowering effect of rs72658867-A is further strengthened by replication of the effect in three additional populations . Since the splice region mutation also increases expression of the LDLR mRNA , the non-HDL-C lowering could be mediated by an increase above normal levels in wild type LDLR transcripts . We , however , demonstrate that the wild type mRNA levels are comparable in heterozygous rs72658867-A carriers and non-carriers . Furthermore , others have shown that in a lymphoblastoid cell line generated from a heterozygous carrier of rs72658867-A , the membrane bound LDLR levels and internalization of LDL are similar to that of cell lines that do not carry LDLR mutations [27] . The non-HDL-C lowering effect of rs72658867-A is thus not mediated by a net increase in the wild type LDLR transcripts . LDLR mutations have been described that cause truncation of the receptor at similar location as rs72658867-A is predicted to do [24–26] , however , these mutations are different from rs72658867-A in that they appear to lead to reduction in wild type transcripts . In contrast to rs72658867-A , these mutations have been linked to FH and an increase in non-HDL-C . Perhaps the combination of a truncated receptor and normal wild type levels of the LDLR mediate the non-HDL-C lowering effect of rs72658867-A . The observed effect of rs72658867-A on LDLR splicing can be attributed to its spatial relation to the site of splicing ( c . 2140+5G>A ) . Allele specific analysis of mRNA sequence data indicates that the increase in LDLR transcripts in rs72658867-A carriers is derived from the chromosome carrying rs72658867-A . The LDLR splicing defect and increased LDLR mRNA expression are thus likely both mediated by the splice region variant rs72658867-A itself since in Iceland no other variant than rs72658867-A can fully explain the association with non-HDL-C and in the 1000G European ancestry data ( Phase 3 dataset , see URL ) no variant is correlated with r2>0 . 5 with rs72658867-A . It is however , unlikely that the effect of rs72658867-A on LDLR expression is mediated by the splicing defect itself . Based on ENCODE data , rs72658867 overlaps a RNA polymerase II binding site in number of different cell lines which may possibly reflect an enhancer site that could mediate altered LDLR expression . In conclusion we have identified two non-coding low frequency variants in the LDLR gene that associate with lower non-HDL-C and protection against CAD . One of them , the splice region variant rs72658867-A , affects splicing and introduces a premature stop codon that is expected to produce a truncated LDLR lacking the O-linked glycan region and the transmembrane and cytoplasmic domains , domains that are both essential for function of the receptor . The same mutation increases transcription of the LDLR , albeit the normal wild type transcript levels do not exceed levels detected in non-carriers . These data contrast the effects of other reported LDLR truncating mutations that increase LDL-C levels and the risk of CAD . Further functional studies are warranted to gain better understanding of the biology of the splice region variant rs72658867-A .
The study was approved by The National Bioethics Committee in Iceland ( Approval no . 07–085 , with amendments ) , and the Data Protection Authority in Iceland ( Approval no . 2007060474ÞS/— , with amendments ) . All donors of biological samples gave informed written consent . Long-range phasing of all chip-genotyped individuals was performed with methods previously described[34 , 35] . For the HumanHap series of chips , 304 , 937 SNPs were used for long-range phasing , whereas for the Omni series of chips 564 , 196 SNPs were included . The final set of SNPs used for long-range phasing was composed of 707 , 525 SNPs . A detailed description of imputation methods used for the Icelandic population was recently published[15] . In brief , SNPs and INDELs identified through sequencing were imputed into 104 , 220 chip-genotyped and long-range phased Icelanders . Approximately 28 . 3 million SNPs and small INDEL variants were imputed based on this set of individuals . The imputation quality score for the four highly significant LDLR sequence variants , rs17248720-T , rs17248748-T , rs200238879-C and rs72658867-A was 0 . 99 , 0 . 99 , 0 . 95 and 0 . 98 , respectively ( Table 1 ) . A generalized form of linear regression that accounts for relatedness between individuals was used to test for the association of quantitative traits with sequence variants[36] . Conditional analysis was performed by including the sequence variant being conditioned on as a covariate in the model under the null and the alternative in the generalized linear regression . Stepwise forward selection was used starting with sequence variants as a covariate in the model then adjusting for the most significant sequence variant and repeating that process until no variant remained significant in the region . Logistic regression was used to test for association between sequence variants and disease ( CAD ) , treating disease status as the response and expected genotype counts from imputation or allele counts from direct genotyping as covariates . Other available individual characteristics that correlate with disease status were also included in the model as nuisance variables . These characteristics were: sex , county of birth , current age or age at death ( first- and second-order terms included ) , blood sample availability for the individual and an indicator function for the overlap of the lifetime of the individual with the timespan of phenotype collection . Testing was performed using the likelihood ratio statistic . Individuals in both the Icelandic case and control groups are related , causing the χ2 test statistic to have a mean >1 and median >0 . 675 . We estimated the inflation factor λg based on a subset of about 300 , 000 common variants and the P-values were adjusted by dividing the corresponding χ2 values by this factor to adjust for both relatedness and potential population stratification[37] . Genomic control correction factors: non-HDL-C: 1 . 36 , triglycerides: 1 . 40 , HDL-C: 1 . 575 , CAD: 1 . 71 , CAD age of onset: 1 . 41 , lifespan: 1 . 49 . The informativeness of genotype imputation was estimated by the ratio of the variance of imputed expected allele counts and the variance of the actual allele counts: Var[E ( θ|chipdata ) ]Var ( θ ) , where θ∈{0 , 1} is the allele count . Var[E ( θ|chipdata ) ] was estimated by the observed variance of the imputed expected counts and Var ( θ ) was estimated by p ( 1 − p ) , where p is the allele frequency . Sequence variants with imputation information below 0 . 8 were excluded from the analysis . Sequence variants were annotated with information from Ensembl release 70 using Variant Effect Predictor ( VEP ) version 2 . 8[38] . Variants annotated as having high impact include loss-of-function variants , i . e . stop-gained variants , frameshift indels and essential splice variants , and moderate impact variants include missense , inframe indels , and splice region variants . Samples of RNA from human peripheral blood were hybridized to Agilent Technologies Human 25K microarrays as described previously[39] . We quantified expression changes between two samples as the mean logarithm ( log10 ) expression ratio ( MLR ) compared to a reference pool RNA sample . In comparing expression levels between groups of individuals with different genotypes , we denoted the expression level for each genotype as 10 ( average MLR ) , where the MLR is averaged over individuals with the particular genotype . We determined s . e . m . and significance by regressing the MLR values against the number of risk alleles carried . We took into account the effects of age , gender and differential cell type count in blood as explanatory variables in the regression . P-values were adjusted for familial relatedness of the individuals by simulation . Picard: http://broadinstitute . github . io/picard/command-line-overview . html#CollectRnaSeqMetrics HaploReg v3 ( accessed May 2015 ) : http://www . broadinstitute . org/mammals/haploreg/haploreg_v3 . php 1000genomes Phase 3 ( May 2013 ) : ftp://ftp . 1000genomes . ebi . ac . uk/vol1/ftp/release/20130502/
|
Cholesterol levels in the bloodstream , in particular elevated low-density lipoprotein cholesterol ( LDL-C ) , are strong risk factors for cardiovascular disease , and LDL-C reduction reduces mortality in people at risk . One of the major determinants of plasma LDL-C levels is the low density lipoprotein receptor ( LDLR ) that acts as a scavenger for cholesterol rich lipoprotein particles . Mutations that disrupt the function of the LDLR or lead to reduction in the number of LDLR usually result in elevated LDL-C in blood . In the current study , we identified , through whole-genome sequencing and imputation into a large fraction of the Icelandic population , four LDLR gene variants that affect non-HDL-C levels ( that includes cholesterol in LDL and other pro-atherogenic lipoproteins ) and risk of coronary artery disease ( CAD ) . Two variants are known and two are novel . One of them , a splice region variant in intron 14 ( rs72658867-A ) , affects normal splicing and is predicted to generate a truncated LDLR , lacking domains essential for receptor function . Despite this , rs72658867-A lowers non-HDL-C substantially and protects against CAD in the general population , demonstrating that variants that disrupt the LDLR can result in lower cholesterol levels .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
A Splice Region Variant in LDLR Lowers Non-high Density Lipoprotein Cholesterol and Protects against Coronary Artery Disease
|
An epistatic interaction between two genes occurs when the phenotypic impact of one gene depends on another gene , often exposing a functional association between them . Due to experimental scalability and to evolutionary significance , abundant work has been focused on studying how epistasis affects cellular growth rate , most notably in yeast . However , epistasis likely influences many different phenotypes , affecting our capacity to understand cellular functions , biochemical networks adaptation , and genetic diseases . Despite its broad significance , the extent and nature of epistasis relative to different phenotypes remain fundamentally unexplored . Here we use genome-scale metabolic network modeling to investigate the extent and properties of epistatic interactions relative to multiple phenotypes . Specifically , using an experimentally refined stoichiometric model for Saccharomyces cerevisiae , we computed a three-dimensional matrix of epistatic interactions between any two enzyme gene deletions , with respect to all metabolic flux phenotypes . We found that the total number of epistatic interactions between enzymes increases rapidly as phenotypes are added , plateauing at approximately 80 phenotypes , to an overall connectivity that is roughly 8-fold larger than the one observed relative to growth alone . Looking at interactions across all phenotypes , we found that gene pairs interact incoherently relative to different phenotypes , i . e . antagonistically relative to some phenotypes and synergistically relative to others . Specific deletion-deletion-phenotype triplets can be explained metabolically , suggesting a highly informative role of multi-phenotype epistasis in mapping cellular functions . Finally , we found that genes involved in many interactions across multiple phenotypes are more highly expressed , evolve slower , and tend to be associated with diseases , indicating that the importance of genes is hidden in their total phenotypic impact . Our predictions indicate a pervasiveness of nonlinear effects in how genetic perturbations affect multiple metabolic phenotypes . The approaches and results reported could influence future efforts in understanding metabolic diseases and the role of biochemical regulation in the cell .
An epistatic interaction between two genes occurs when the phenotypic impact of at least one of the genes is dependent on the other [1] . This dependence is often a consequence of an underlying functional relationship between the two genes [2] , [3] . By extending the study of epistasis from individual interactions to networks of interactions , recent work in S . cerevisiae has demonstrated that genome-wide patterns of epistasis can be used to uncover the global organization of biological systems [4]–[7] . In such studies epistatic interactions are identified as instances where the effect of a double perturbation on growth differs from the expectation based on the observed effects of the corresponding single perturbations [8] . The choice of growth rate as a phenotype is motivated by the role of epistasis in the dynamics of selection [9] , and by the fact that growth rate , a proxy for fitness , can be accurately measured in a high-throughput manner [10] . In parallel to the experimental efforts , large-scale studies of epistasis on growth phenotypes have also been pursued computationally , especially using the approach of flux balance analysis [6] , [11] , [12] . Such computational studies have offered preliminary novel insight before the availability of corresponding experimental data , e . g . in predicting a coherence principle ( monochromaticity ) in the organization of epistatic interaction networks [6] , subsequently observed experimentally [7] , [13] . Overall , large-scale studies of epistasis have become increasingly relevant to functional genomics [4] , [7] , [14] , drug development [13] , [15] , [16] and evolutionary biology [17] , [18] . Albeit important , growth rate is just one of many possible phenotypes relative to which genes can interact epistatically with each other . In contrast with the rapidly increasing understanding of the nature and scope of epistatic interactions relative to growth , many questions remain unresolved with respect to epistasis relative to non-growth phenotypes . Are interactions relative to non-growth phenotypes as widespread as interactions with respect to growth ? Do genes tend to interact relative to more than one phenotype , and if so , is the type of epistasis consistent across phenotypes ? How much more dense can an epistatic network become upon adding new phenotypes ? Do interactions with respect to specific phenotypes provide biological insight than cannot be obtained from knowing interactions relative to growth rate ? Most importantly , does the potential presence of multi-phenotype epistasis affect the way cells operate and evolve ? While these questions have not , to our knowledge been asked before , epistasis relative to non-growth phenotypes is not in itself a new concept . Interactions between polymorphisms have been detected by using multiple mRNA transcript levels as phenotypes [19] . Another recent study searched for interactions among genes conferring resistance to a DNA-damaging agent and showed that a denser network was observed with respect to the capacity to cope with the damaging agent , than was found with respect to growth rate under standard conditions [20] . In addition , in the study of human genetic diseases , while epistasis relative to disease-related traits poses challenging technical problems , it is a potentially important component , especially in light of the relative paucity of explanatory power detected through the analysis of individual loci [21]–[23] . Hence interactions relative to diverse phenotypes are likely widespread and informative . However , the combinatorial complexity resulting from the large number of possible genetic perturbations and phenotypes has prevented so far a systematic analysis of the extent and biological implications of this phenomenon . In this work we report the computational study of epistatic interactions in a flux balance model of metabolism that is simple enough to allow an exhaustive computation of all possible perturbations relative to all possible phenotypes , but at the same time realistic enough to provide meaningful biological insight . Specifically , we use an experimentally informed variant of the method of minimization of metabolic adjustment ( MOMA ) in a genome-scale metabolic network model of Saccharomyces cerevisiae [24] to predict all steady state metabolic reaction rates ( fluxes ) in response to all possible single and double enzyme gene deletions . By comparing single and double mutant values for all fluxes and defining appropriate metrics , we construct an epistatic map for each flux phenotype ( Figure 1 ) . This multi-phenotype genetic interaction map allows us to explore for the first time the properties and significance of epistasis across a combinatorial set of perturbations and phenotypes .
Quantifying epistasis relative to multiple metabolic flux phenotypes introduces three fundamental challenges , one specific to the use of flux balance models and two broadly relevant to any study of multi-phenotype epistasis . The first issue is the reliability of flux predictions for deletion mutants . The availability of experimentally determined growth phenotypes for all gene deletion mutants in S . cerevisiae has allowed for extensive evaluation of the yeast model's capacity to predict mutant growth . These previous studies [25] , [26] , including a comparison of model predictions against experimental growth measurements for 465 gene deletion mutants under 16 metabolically diverse conditions [26] , have demonstrated that the yeast model can predict deletion mutant viability with high accuracy . Furthermore , observed discordances between model predictions and experimentally determined mutant growth phenotypes have been used in refinements of the existing yeast model , further bolstering the ability of the model to accurately mimic the effect of different gene deletions [25] , [27] . In addition to effectively predicting single mutant growth , flux balance models have also been shown to predict viabilities of double deletion mutants with high accuracy [28] . However , while model predictions of mutant growth have been evaluated extensively , comparisons between measured and predicted fluxes through the underlying metabolic reactions in different mutants are less readily available [29] . To address this need we recently evaluated the ability of the yeast model to predict the fluxes through central carbon metabolism in single gene deletion mutants by comparing model predictions to a previously released compendium of experimentally measured mutant fluxes [30] . An assessment of different approaches for mutant flux prediction revealed that an experimentally driven variant of the minimization of metabolic adjustment [31] gives the best correlation with measured fluxes ( Spearman rank correlation greater than 0 . 90 , Figure S1 ) , and hence chose it for our calculations ( See Materials and Methods ) . In essence , this method implements the hypothesis that the metabolic response to genetic perturbation will be a minimal rerouting of flux around the insult . A conceptual illustration of the methodology for predicting mutant fluxes is shown in Figure 2 , with a detailed quantitative description provided in the Materials and Methods and Text S1 . A second issue , which is critical to any multi-phenotype study of epistasis , is the choice of a metric for quantifying epistasis . The quantification of epistasis requires an assumption as to how the phenotypic effects of non-interacting mutations combine: deviations from this expectation are inferred to be indicative of epistasis . While previous work has provided both theoretical and empirical evidence for how the effects of mutations on fitness combine [8] , no comprehensive study has yet explored how the effects of mutations on metabolic fluxes combine . To this end we evaluated two standard metrics for computing epistasis ( multiplicative and additive definitions ) , in addition to a novel metric . This novel metric was designed so as to avoid making any assumption on how the phenotypic effects of two mutations combine . Avoiding such assumptions is ideal for detecting epistasis across multiple phenotypes , relative to which the effects of mutations may combine differently ( See Text S1 and Table S1 ) . However , comparing these three different quantitative definitions ( See Materials and Methods and Text S1 ) , we found that epistasis relative to metabolic fluxes is overall robustly detectable independent of the metric used ( Figure S2 ) . In the following analyses , based on this result , a multiplicative model is used , and all main conclusions were verified to be robust relative to different metrics . A third issue arising in a global analysis of epistatic effects with respect to metabolic fluxes is the partitioning of interactions into different classes of epistasis [32] . These different classes of epistasis represent different ways in which the combined effect of two mutations may defy expectation , and can be indicative of different types of underlying functional relationships between genes [2] , [6] . In moving from growth to flux phenotypes , the classification of interactions becomes more complex , due to the fact that fluxes can increase or decrease upon genetic perturbation , while the growth rate typically only decreases . While the increased complexity present in our data allows for discrimination of many different classes of interactions ( See Figure S3 and Text S1 ) , for the current analysis we consolidate all sub-classes of interactions into two groups , synergistic and antagonistic ( See Materials and Methods , Figure S3 ) . A synergistic interaction between two genes indicates that the change in the observed flux ( phenotype ) caused by the simultaneous deletion of both genes is greater than expected based on the effects of the corresponding single deletions , while an antagonistic interaction indicates a flux change in the double mutant that is less than expected . Synergistic interactions are indicative of a compensatory relationship between two genes , such that the extreme phenotype of the double mutant is a consequence of this compensation being lost . Antagonistic interactions are indicative of two genes working together towards some function , such that the reduced phenotypic effect of the double mutant occurs because the common function is compromised by the loss of either of the genes individually . These preparatory steps allowed us to compute and analyze a 3-dimensional epistatic map for the yeast metabolic network , as illustrated in Figure 1 . The complete set of synergistic and antagonistic epistatic interactions were reduced to a high-confidence set by independently applying a standard deviation cutoff to the distributions of epistasis relative to each phenotype ( See Materials and Methods and Figure S4 ) . Considering only these high confidence interactions , it was found that 100 of the 672 genes in the model interact with respect to at least one of the 293 fluxes active under the modeled minimal glucose condition . To simplify the subsequent analysis of the epistatic map , we consolidated the 100 interacting genes into the 30 metabolic processes to which they are assigned in the model , and counted an interaction between two processes if any gene from one process interacts with any gene from the other . This consolidated epistatic map is represented in Figure 3A , where the total numbers of synergistic ( red ) , antagonistic ( blue ) and mixed ( yellow ) interactions between pairs of biological processes , across all phenotypes , are displayed as a stacked histogram . Mixed interactions between two processes occur when some pairs of genes across the processes interact synergistically , while others antagonistically . Figure 3A indicates that such mixed process interactions are less frequent than process interactions that are purely synergistic or antagonistic , suggesting that the previously observed monochromaticity of epistatic interactions between biological processes [6] applies to diverse metabolic phenotypes . Monochromaticity is a consequence of the fact that genes in the same biological processes function cohesively , and hence share similar patterns of epistatic interactions [4] , [5] , [7] . Notably , however , in our multi-phenotype epistatic map , the “color” ( synergistic or antagonistic ) of the interaction between two processes depends on the phenotype observed . Figure 3B demonstrates that this dependence of process interaction colors on phenotype is due to the fact that individual gene pairs often interact synergistically relative to some phenotypes and antagonistically relative to others . This pattern reveals that the class of an epistatic interaction is not an absolute characteristic of a pair of genes , but rather a characteristic of the gene-gene-phenotype triad . This suggests that the functional relationship between two genes is not necessarily one dimensional , but may depend on the function ( the phenotype ) being probed . The intuition that different phenotypes convey complementary insight into the functional associations between genes and processes was confirmed in a quantitative manner by determining how many unique interactions each phenotype contributes to the 3D epistatic map . Figure 4 shows that the total number of interactions identified when considering all phenotypes is ∼8 times larger than can be identified relative to any individual phenotype , although the exact increase in interaction coverage is dependent on the threshold for defining a significant interaction ( See Materials and Methods ) . Figure 4 also shows that 83 of the 293 total metabolic flux phenotypes are required to identify all unique epistatic interactions in yeast metabolism . Examining the distribution of metabolic processes where these 83 phenotypes come from ( Figure S6 ) , reveals that they are spread across all metabolic processes . This suggests that a set of phenotypes that represents all metabolic functions is required to identify all epistatic interactions . Conversely , this implies that different phenotypes are providing insights into unique aspects of the functional relationships between genes . To solidify the observation that different flux phenotypes reveal unique aspects of the functional relationships between genes , we next focus on the epistatic networks relative to two secretion phenotypes ( succinate , Figure 5B , and glycerol , Figure 5C ) . We chose to focus on secretion flux phenotypes because they are the most tractable fluxes to measure experimentally , and hence potentially the most relevant for future experimental studies . Both of these secretion flux epistatic networks contain several interactions that are not detected relative to the growth phenotype ( Figure 5A ) . In particular , in the succinate secretion network , the genes that are part of complex II of the electron transport chain ( ETC II ) display synergistic interactions with several other biological processes ( Figure 5B ) . Among these interactions , which are indicative of an unexpectedly large increase in succinate secretion in the double mutant , the one between serine biosynthesis and ETCII has been reported in previous experimental efforts to overproduce succinate [33] . This interaction occurs because the predicted alternate pathway for serine biosynthesis produces succinate as a byproduct , and ETC II is the primary route through which this succinate is metabolized in the wild-type ( Figure 6A , Figure S7 ) . Thus , interactions with respect to succinate may in general probe the way in which TCA cycle intermediates are produced and consumed . In the glycerol secretion phenotype network there is enrichment for synergistic interactions between glutamate biosynthesis and respiratory processes ( Figure 5C ) . Among these interactions , the interaction between glutamate synthase and the electron transport chain is supported by experimental data gathered in the context of ethanol production optimization [34] . This epistatic interaction is a consequence of the fact that glutamate biosynthesis , the electron transport chain and glycerol biosynthesis correspond to three of the major routes for cytosolic NADH oxidation ( Figure 6B , Figure S9 ) . Thus , interactions with respect to glycerol secretion may reflect the way in which different processes contribute to cellular redox balance . These examples , and others in Text S1 and Figure S8 , further demonstrate that interactions with respect to metabolic flux phenotypes can provide detailed insights into different aspects of the functional relationships between genes . So far , we have shown that epistatic interactions between gene deletions relative to metabolic flux phenotypes are ubiquitous , and can provide an understanding of the relationships between different processes in the cell . The ubiquity of epistasis relative to metabolic flux phenotypes brought to our attention the possibility that these complex network-level functional interdependencies might impose constraints on evolutionary trajectories . We hypothesized that this phenomenon might manifest itself in the form of increased evolutionary constraints on enzymes that are involved in many epistatic links with other genes . Such a relationship between epistatic connectivity and evolutionary rate has been recently observed in the experimentally constructed global genetic interaction network with respect to growth rate in yeast [7] . Thus , we set out to explore whether predicted connectivity with respect to metabolic phenotypes other than growth rate are also correlated with evolutionary constraint . To this end we calculated the Spearman rank correlation between the number of interactions in which different genes participate and the evolutionary rates of such genes , as measured by their non-synonymous to synonymous substitution ratios . This correlation was calculated separately for synergistic and antagonistic interactions relative to each of the 293 flux phenotypes , for a total of 586 correlations . The distributions of correlation coefficients for synergistic and antagonistic interactions across all phenotypes are shown in Figure 7A . Both distributions significantly deviate from zero , with an overall bias towards negative correlations ( Sign test , p = 8 . 5×10−25 ( synergistic ) , 2 . 2×10−54 ( antagonistic ) , n = 293 ) . This trend towards negative correlations suggests that genes involved in many interactions with respect to metabolic flux phenotypes do indeed evolve slower . While the negative skew of these distributions is robustly maintained upon removal of most potential confounding factors ( see Figure S11 ) we found that it is significantly reduced when controlling for the codon bias of the genes ( Figure 7B ) . Codon bias is a proxy for gene expression level , which previous research has shown to be the dominant correlate of evolutionary rate [35] , [36] . Therefore , we cannot rule out that a portion of the apparent evolutionary importance of genes with a high degree of genetic interactions across different phenotypes may be explainable by the expression level of the genes . Yet , regardless of whether the interaction degree correlates with evolutionary rate or gene expression level , either result indicates the functional importance of these multi-phenotype hubs . The increased expression level of these hubs in fact supports their central role in metabolic function . Furthermore , in our model , epistatic interaction degree with respect to growth flux alone is not significantly anti-correlated with evolutionary rate , even without controlling for expression level ( Figure S12 ) . This indicates that the importance of genes is associated with their total phenotypic impact , not just their impact on growth . While the distributions of correlations between evolutionary rate , and both synergistic and antagonistic interaction count , shift towards zero when controlling for codon bias , the distribution remains significantly different from zero only for antagonistic interactions ( p = 0 . 07 ( synergistic ) , p = 2 . 3×10−31 ( antagonistic ) ) . We believe that this observation can be understood by considering more closely the relationships between genes that interact antagonistically , versus synergistically . An antagonistic interaction implies that the phenotypic effect of deleting a gene is reduced in the absence of its interaction partner . A possible interpretation of this is that a gene's full function , as manifested in its associated phenotypic effect , is contingent on the presence of its antagonistic interaction partner . Therefore highly antagonistic genes are phenotypic hubs , whose evolutionary changes are constrained by the dependency of other genes upon them . Conversely , the reduced constraint on synergistic hubs can be understood by considering that a synergistic interaction between two genes implies that the phenotypic impact of deleting a gene is increased in the absence of its interaction partner . This can be interpreted as a gene's function being compensated for by its synergistic interaction partners . Therefore , the reduced correlation with evolutionary rate for synergistic hubs may reflect the fact that the phenotypic effect of changes in such hubs is dampened by the presence of their interaction partners . The implications of the current analysis are not limited to yeast . In fact , multi-phenotype epistatic interactions may be relevant to the manifestation and treatment of human disease . Given the previously discussed importance of multi-phenotype hub genes , it is likely that perturbations of these genes would have major effects in a biological network . Translating this observation to humans , we hypothesize that the disruption of more highly connected genes in the human metabolic network would be more likely to result in a disease state . We sought evidence for this by evaluating whether the epistatic connectivity of genes in the yeast model was predictive of the role of their human homologs in genetic diseases . Indeed , we observe a significant difference between the connectivity of yeast homologs of human genes that have been associated with a genetic disease , versus those that have not ( Figure 8 ) . While the statistical significance is limited due to the small sample size , this result provides support for the growing sentiment that majority of human genetic disorders are a consequence of complex interactions between numerous cellular components [1] , [21] .
We described the systematic generation of epistatic interaction networks relative to all observable phenotypes in a genome-scale model of yeast metabolism . Analysis of these networks revealed that different metabolic flux phenotypes yield different sets of interactions , and that a large set of phenotypes is required to capture all interactions . The basis for these observations is that different metabolic flux phenotypes capture different aspects of gene function . This is likely a consequence of the complex wiring of metabolic networks , which include multiple branching pathways , shared pools of commonly used metabolites and a high level of interconnectedness between different metabolic processes: seemingly remote processes on the metabolic chart may nonlinearly affect a third readout process ( the phenotype ) . Furthermore , because of this complexity , the relationships between different genes and processes may not be easily captured by straightforward patterns , as indicated by the observation that the same genes can interact synergistically relative to some phenotypes and antagonistically relative to others . From a functional genomic perspective , the results imply that , in future studies of epistasis , the set of observed phenotypes could be selected so as to influence the set of interactions identified and to maximize insight into the functional organization of the biological process of interest . While the focus here has been on metabolism , this concept can be generalized to other types of biological networks . For instance , mRNA transcript levels may be the most appropriate phenotype to tease out the logic of transcriptional regulatory networks [37] and phosphorylation states the most relevant for signal transduction pathways . Furthermore , our results demonstrate that the particular mRNA levels or protein phosphorylation states monitored should depend on the particular regulatory module or transduction pathway of interest . An additional layer of complexity that has not yet been addressed here is the dependence of epistatic networks on environmental conditions . As hinted to before ( Supplementary Figure 3 in [6] ) epistatic networks will likely vary under different conditions . Hence , future extensions of the current work may explore the complexity and significance of environmental conditions as a fourth dimension in the epistatic matrix of Figure 1 . From an evolutionary perspective , we found that the number of epistatic interactions with respect to multiple metabolic flux phenotypes is strongly anti-correlated with the genes' evolutionary rates and expression levels . This anti-correlation is larger than found with the number of epistatic interactions relative to growth phenotype only . On the surface this result seems surprising , given that growth rate and fitness are often taken to be synonymous with one another , and genes that have a large impact on fitness would be expected to evolve slower . However , one must consider that growth in the model is based solely on the capacity to produce biomass components , while fitness in an organism's natural environment is assuredly more complex . An organism's success ( in other words , its fitness ) likely depends on the complex interplay of a multitude of biological properties , including the proportions and efficiency of resources utilized , the choice of secreted byproducts ( which can influence the environment and the interactions with other species ) , and how fluxes are managed in the face of varying nutrient availability . Thus , the apparently reduced importance of the growth flux , and conversely , the increased importance of all metabolic phenotypes , may simply be reflective of the relative simplicity of growth in the model , when compared to the complexity of growth in the wild . More broadly , our results raise the possibility that the apparent robustness observed in the insulated environment of the laboratory may not translate to an organism's natural environment , where additional constraints exist with respect to not just how fast one grows , but the precise manner in which this is accomplished . A potential limitation of past and present , computational and experimental studies of the evolutionary impact of epistasis may lie in the use of gene deletions as the mutation relative to which epistasis is detected . While gene deletion mutations have been effective in terms of uncovering functional dependencies and the evolutionary constraints imposed by these dependencies , left unanswered is the evolutionary impact of epistasis relative to smaller perturbations to gene function . It is these minor perturbations , such as those caused by amino acid substitutions or stochastic fluctuations in protein levels that the cell must constantly confront . If epistasis relative to these small perturbations is as ubiquitous as has been observed relative to gene deletions , this begs the question as to how the cell copes with the complexity of a large number of long-distance nonlinearities affecting virtually every metabolic function . While experimental studies have begun to address this question [38] , perhaps computational frameworks such as flux balance analysis can be used to extend these analyses to the genome-scale . For flux balance analysis , or any computational framework , to adequately address this problem much work will have to be done to more fully understand the phenotypic consequences of small genetic perturbations . While our current analysis is purely computational , we anticipate that xperimental measurements of interactions based on multiple metabolic phenotypes will be increasingly feasible and valuable in the near future . Our analysis provides predictions about the properties of multi-phenotype epistatic networks , in addition to a plethora of specific interaction predictions to which these future experiments can be compared ( data downloadable at http://prelude . bu . edu/multi-phenotype-epistasis ) . Finally , being the first genome-scale analysis of multi-phenotype epistatic networks , we hope that the groundwork we have laid with respect to quantifying , discretizing and analyzing multi-phenotype epistatic interaction networks will aid future experimental and computational studies using similar approaches to help unravel the functional complexity of biological systems .
To enable our study of multi-phenotype epistasis at a genome-scale we utilized flux balance models [24] . Specifically , to compute steady state reaction rates ( the fluxes , vi ) in deletion mutants , we used the iLL672 yeast stoichiometric reconstruction . Flux balance models take as input the stoichiometry of all known metabolic reactions in the modeled organism , along with possible constraints on flux ranges , and through a Linear Programming optimization step provide predictions of fluxes through each metabolic reaction . The complete stoichiometry of an organism is typically represented mathematically as the stoichiometric matrix , S . Each row i of the matrix S represents a metabolite , and each column j represents a reaction , with an entry Sij representing the stoichiometric coefficient of metabolite i in reaction j . The set of possible flux solutions is constrained by imposing a steady state assumption along with bounds on individual fluxes . The set of steady state solutions is described mathematically as the null space of the matrix S , and dictates that the production of each metabolite is equaled by its consumption . Bounds on individual fluxes are described by inequality constraints , and are used to model known limitations , such as nutrient availabilities , reaction reversibility and maintenance requirements . Constraints on nutrient limitation in the present study were set so as to mimic as closely as possible the media conditions from a recent study by Blank et al . [30] . This condition was selected since it allowed us to use experimental flux data from that study to perform more accurate flux predictions throughout our work ( See below ) . Upon setting the linear constraints , a particular flux solution is typically computed by searching the optimal value of a given linear combination of the fluxes . Previously utilized optimization criteria are maximal ATP production [39] , minimization of total flux [40] , and the most commonly used maximization of biomass production [41] . Formally , this can be expressed as a Linear Programming ( LP ) problem:where cr is the coefficient of flux r in the objective function ( r = 1 , … , R , with R is the total number of reactions ) , and αj and βj are the upper and lower bounds on reaction j , respectively . Because of the nature of our study , accurate predictions for all individual fluxes are desirable . Hence , we wanted our flux predictions to match available experimental data as closely as possible . To this end we evaluated several optimization criteria for predicting fluxes in deletion mutants by comparing flux predictions to a previously released set of experimentally measured fluxes in yeast single deletion mutants [30] ( see Text S1 ) . Our evaluation demonstrated that the most accurate optimization criteria utilized the previously described Minimization of Metabolic Adjustment ( MOMA ) criteria [29] , along with an experimentally constrained wild type solution [24] . In effect , this criterion assumes that upon a gene deletion , fluxes will undergo a minimal rearrangement , compatible with the flux constraints imposed by the gene deletion . We found that the performance of this approach is highly dependent on the accuracy of the wild-type solution from which the distance is minimized . Therefore the wild type fluxes were computed using the following LP optimization:where is the experimentally measured value for flux j , δ is a parameter that describes the stringency of the requirement ( currently set to 0 . 10 , as done previously [24] ) and ε represents the error associated with the experimental measurement . This approach identifies , among all states compatible with the experimentally measured fluxes , the one with minimal overall flux . Mutant fluxes were computed by constraining to zero the flux through any reaction requiring the protein product of the deleted gene ( s ) , and then identifying the flux solution with the minimal Manhattan distance from the wild type flux solution vWT . The optimization problem , which can be solved using LP , is formulated as follows:where represents the flux ( es ) requiring the protein product of gene gi . This approach for determining mutant fluxes has been described in detail elsewhere [24] . All LP calculations were performed using the software Xpress , under free academic license . From the total set of genes present in the model , only a subset was used in our analyses . First , all essential genes were excluded , as by definition they cannot participate in any epistatic interactions when considering gene deletion mutations . Second , for genes assigned to complexes in the model , only a single representative gene from the complex was used . This was done because complexes are treated in a trivial way in the model , wherein removal of any gene in the complex is assumed to have the equivalent effect of disabling the complex . Finally , only genes whose deletion affected at least one metabolic flux relative to the wild-type were included . This filter was applied in response to the trivial way in which isozymes are treated in the model , wherein complete backup is assumed . The predicted fluxes for all single and double gene deletions can be represented as a three-dimensional matrix V , whose element Vijk indicates the normalized k-th flux for the deletion of genes i and j ( single mutants being represented by the diagonal elements i = j ) . The three-dimensional matrix of epistatic interactions , E , is in turn defined by measuring how much the flux for each double mutant differs from expectation , based on the flux in the corresponding single mutants . Specifically , we define an element of E as Eijk = Vijk – F ( Viik , Vjjk ) , where the specific shape of the function F is discussed in the next section . The element Eijk represents the interaction between genes i and j relative to the phenotype k . Two genes i and j were considered to have an epistatic interaction relative to phenotype k if |Eijk|>ssk , where sk denotes the standard deviation for the distribution Eijk values across all pairs ( i , j ) . For assessing total interaction coverage and process interaction patterns a value of s = 1 . 0 was used , so as to provide a cleaner picture by retaining only the most high confidence interactions . For the evolutionary rate analysis , a value of s = 0 . 5 was used , as we deemed the inclusion of weaker interactions to be conceptually important here . Note that none of the major conclusions change with respect to varying of s ( see Figures S5 and S10 for sensitivity analysis relative to s ) . Note that flux values were rounded to 5 digits after the decimal point , in order to avoid numerical errors associated with the optimization software . To identify an appropriate metric for quantifying epistasis with respect to metabolic phenotypes , we evaluated two commonly used metrics , as well as a newly defined one . The two previously applied metrics correspond to a multiplicative definition and an additive definition of epistasis respectively . The multiplicative metric assumes that the expected wild-type-normalized phenotype ( flux ) change for the double mutant is the product of the corresponding changes for the two single mutants . The additive metric , on the other hand , assumes that the expected double mutant change is the sum of the two individual mutant changes ( see Text S1 and reference [8] for details ) . The novel metric we developed is a Z-score based metric that quantifies the difference in the effect of a mutation in the wild type and mutant backgrounds ( See Text S1 for details ) . All interactions were classified as either antagonistic or synergistic . In general , an interaction was deemed as antagonistic if the phenotype of the double mutant was less severe than expected based on a multiplicative definition and as synergistic if the double mutant phenotype was more severe than expected ( See Text S1 for details ) . For our analysis of the relationship between the number of epistatic interactions associated with perturbed genes and the ratio of the rate of non-synonymous substitutions to the rate of synonymous substitutions ( Ka/Ks ) , we computed the total number of synergistic and antagonistic interactions associated with 39 genes across 293 flux phenotypes . The 39 genes were selected on the basis of Ka/Ks data being available from a previous study [42] . Statistical tests were performed as described in figure legends . Partial correlation analysis was done to control for single mutant fitness defects , metabolic network connectivity and codon bias ( See Figure S11 ) . In order to assess the potential relevance of multi-phenotype epistasis to the manifestation of human disease the number of epistatic interactions was compared between yeast genes that have a human homolog in Online Mendelian Inheritance in Man ( OMIM ) database [43] and those genes that have a human homolog that is not in the OMIM database . Human homologs of yeast genes were determined based on the Kyoto Encyclopedia of Genes and Genomes [44] . The OMIM morbid map data set was downloaded on April 9th , 2009 .
|
An epistatic interaction between two genes occurs when the phenotypic impact of one gene is dependent on the other . While different phenotypes have been used to uncover epistasis in different contexts , little is known about how cell-scale genetic interaction networks vary across multiple phenotypes . Here we use a genome-scale mathematical model of yeast metabolism to compute a three-dimensional matrix of interactions between any two gene deletions with respect to all metabolic flux phenotypes . We find that this multi-phenotype epistasis map contains many more interactions than found relative to any single phenotype . The unique contribution of examining multiple phenotypes is further demonstrated by the fact that individual interactions may be synergistic relative to some phenotypes and antagonistic relative to others . This observation indicates that different phenotypes are indeed capturing different aspects of the functional relationships between genes . Furthermore , the observation that genes involved in many epistatic interactions across all metabolic flux phenotypes are found to be highly expressed and under strong selective pressure seems to indicate that these interactions are important to the cell and are not just the unavoidable consequence of the connectivity of biological networks . Multi-phenotype epistasis maps may help elucidate the functional organization of biological systems and the role of epistasis in the manifestation of complex genetic diseases .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genetics",
"and",
"genomics/microbial",
"evolution",
"and",
"genomics",
"evolutionary",
"biology/microbial",
"evolution",
"and",
"genomics",
"genetics",
"and",
"genomics/functional",
"genomics",
"computational",
"biology/metabolic",
"networks",
"microbiology/microbial",
"physiology",
"and",
"metabolism",
"computational",
"biology/systems",
"biology"
] |
2011
|
Epistatic Interaction Maps Relative to Multiple Metabolic Phenotypes
|
Different neuronal types within brain motor areas contribute to the generation of complex motor behaviors . A widely studied songbird forebrain nucleus ( HVC ) has been recognized as fundamental in shaping the precise timing characteristics of birdsong . This is based , among other evidence , on the stretching and the “breaking” of song structure when HVC is cooled . However , little is known about the temperature effects that take place in its neurons . To address this , we investigated the dynamics of HVC both experimentally and computationally . We developed a technique where simultaneous electrophysiological recordings were performed during temperature manipulation of HVC . We recorded spontaneous activity and found three effects: widening of the spike shape , decrease of the firing rate and change in the interspike interval distribution . All these effects could be explained with a detailed conductance based model of all the neurons present in HVC . Temperature dependence of the ionic channel time constants explained the first effect , while the second was based in the changes of the maximal conductance using single synaptic excitatory inputs . The last phenomenon , only emerged after introducing a more realistic synaptic input to the inhibitory interneurons . Two timescales were present in the interspike distributions . The behavior of one timescale was reproduced with different input balances received form the excitatory neurons , whereas the other , which disappears with cooling , could not be found assuming poissonian synaptic inputs . Furthermore , the computational model shows that the bursting of the excitatory neurons arises naturally at normal brain temperature and that they have an intrinsic delay at low temperatures . The same effect occurs at single synapses , which may explain song stretching . These findings shed light on the temperature dependence of neuronal dynamics and present a comprehensive framework to study neuronal connectivity . This study , which is based on intrinsic neuronal characteristics , may help to understand emergent behavioral changes .
Coordinated motor behaviors require a delicate sequence of gestures . Their precise timing depends on instructions coming from motor control areas in the brain . Differentiated regions or nuclei are interconnected forming a motor pathway . However , little is known about how the detailed motor program arises from the dynamics of individual neurons . A widely used model to study complex behavior is birdsong , which consists of a succession of highly stereotyped vocal gestures . Like humans with their speech , oscine birds that account for approximately forty percent of the known species , need to learn their vocalizations from a tutor [1–5] . The well characterized set of forebrain nuclei in charge of this task is called the “song system” [6–8] . It has proven to be a valuable model to study the network properties necessary to generate the precise motor patterns of song [9–12] . Motor nucleus HVC ( proper name ) appears to have a key role in song production . Important aspects such as the structure and sequence of the different syllables that compose songs are represented in the activity of its neurons [13] . The representation of time and of important events in motor gestures have been described in this nucleus [14–17] , as well as a proposed role in sensorimotor integration [12 , 18 , 19] . Continuous temporal encoding of vocal features was proposed as a key role of HVC [16 , 20] . However , this view has been open to debate recently with the proposition that activity in this nucleus only encodes relevant instances of song [17 , 21–23] , and by a complementary proposition [24 , 25] in which song timescales appear from the interaction of the HVC with the rest of the motor pathway . These have broadened the discussion regarding the possible motor coding of song in HVC , leading to the proposition of recurrent networks for song timing generation [23 , 26–29] . In addition , recently , it has been shown that HVC projecting neurons fire densely during singing behavior [30–32] , and that another nucleus in the song pathway , the subthalamic nucleus Uva , has a significant contribution to song timing [33 , 34] . It remains an open and exciting field for further research since it is still lacking a definitive consensus for the precise role of HVC on the timing of birdsong . At a smaller scale , how individual neurons in HVC contribute to generate the song sequence is also not fully understood . They project both to the forebrain nucleus RA ( robust nucleus of the arcopallium ) and to a basal ganglia nucleus , area X , which is essential for song learning and belongs to the anterior forebrain pathway [35] . For some time , the vocal motor pathway was believed to have a top-down architecture , initiating activity in HVC , projecting to RA which in turn projects to the respiratory center ( RAm and PAm , nucleus retroambigualis and parambigualis ) and to the motor nucleus innervating the vocal organ ( nXIIts ) . However , experiments show that there is also a bottom-up architecture that goes from the brainstem back to HVC that influences the activity of HVC neurons [27 , 34 , 36–38] . A novel technique that has been developed recently to manipulate activity within HVC is local temperature control [16 , 24 , 25 , 33 , 39–41] . It was shown for zebra finches ( Taeniogypia guttata ) that cooling HVC leads to a stretching of the song tempo without major changes in the acoustic structure of song [16 , 39] , and for canaries ( Serinus canaria ) an initial stretching and a restructuring of song , or “syllable breaking” for colder temperatures [24] , and the possibility of controlling syllable transitions [25 , 40] . The working hypothesis for these experiments is that a timescale represented in HVC is being manipulated with temperature , somehow making it slower at colder temperatures . Recent work in zebra finches also cooled Uva to show that song timing control is distributed in a recurrent network [33] . The individual neuronal mechanisms that take part in these physiological observations within HVC will be studied in this work . Previous in vivo studies of the relationship of neuronal functionality and temperature were mostly done in poikilothermal animals , which are cold blooded and cannot regulate efficiently their temperature . The focus on those studies is in understanding how function can be robust and resilient to large temperature fluctuations of the brain [42 , 43] . However , many neuronal properties studied in these animals provide evidence of widespread mechanisms . The main findings point to timing changes with temperature decrease: period and burst interval decrease in acoustic [44] and electric [45] communication in fish . Acoustic perception seems to be deteriorated at lower temperatures in locust [46 , 47] , as spiking variability increases with lower rates , the same effect that is present in the fly vision system [48] . Another widely studied model is the crab stomatogastric system , in which a central pattern generator , although increasing burst period and inner frequency , is able to compensate its phase over a wide range of temperatures [49 , 50] . Also compensation is found in the bullfrog respiratory rhythm , whereas some neurons paradoxically increase firing at lower temperatures [51] . Few studies were done in homeothermal animals apart from the recent birdsong focal cooling . These divide into ones that monitor the temperature of the brain and others that use slices and modify the temperature of the bath of the recording solution . Brain temperature can fluctuate up to 2°C when animals do exercise [52] , when they are presented a female [53–55] and when they are given cocaine [56] . Temperature correlates with neuronal properties like latency after electrical stimulation and dynamics of EPSPs , as was shown in rat hippocampus [52] or song tempo , as shown in HVC in zebra finches [55] . Also resting membrane voltage change is usually reported as it is believed that it can put the neuron closer or further to the spiking threshold . Although many report an increase in this value at low temperatures ( 4mV in hippocampus [57] , 15 mV in rat visual cortex [58] with a 10°C change ) , others report decreases ( 10mV in one type of hippocampus interneurons [59] ) . Finally , cooling has also been used as a tool to change perception of interval timing in rat prefrontal cortex [60] . In all these studies , there is also emphasis in the reversibility of these effects when temperature is returned to normal . The studies mentioned above highlight the increase in timescales and decrease of level of activity present in neuronal behavior with temperature decrease , and that these come from changes in rate and conductance constants . In the other hand detailed modeling of temperature effects has not been explored extensively . The work done in the stomatogastric ganglion of the crab is the most comprehensive , showing a full exploration of the parameter space of extrapolation constants for each ionic rate and conductance [43 , 61] . Just above 5% of the hundreds of thousands of models explored showed bursting activity , and only 0 . 5% showed the robust behavior measured experimentally , although being more than 500 different possible parameter combinations . In our case of study , there are three classes of neuronal populations in HVC: excitatory neurons projecting to nucleus RA and X ( HVCRA and HVCX respectively ) , and inhibitory interneurons ( HVCINT ) [62 , 63] . Intracellular recordings in vitro and in vivo helped to reveal neuronal and circuit mechanisms , and physiological properties of these neurons [63–67] , and the connectivity between them was studied by means of paired intracellular recordings and antidromic stimulation in slices or by means of retrograde labeling [28 , 64] . The recent work by Daou et al . has described extensively the ionic currents present in each neuronal type with an exquisite match between electrophysiological measurements and computational single compartment models using different stimulation protocols in brain slices [67] . Recent single cell modeling in HVC is based in these findings for excitatory neurons [68] , and also was upgraded to two compartments for HVCRA [69] . Network models have also being explored , but less is known about the specific connectivity within HVC [66 , 70 , 71] . In this work we study electrophysiologically the neuronal in vivo mechanisms that manifest in HVC neurons of canaries when the nucleus temperature is decreased locally . We took advantage of the very precise neuronal models developed by intracellular studies in vitro . Although intrinsic parameters were obtained at room temperature , we could extrapolate them to the in vivo temperatures used in this work . We found that the three most characteristic changes with temperature of the experimental neuronal dynamics can be explained with the computational model . We also make predictions about connectivity and possible experimental implementations .
All experiments were approved by the Institutional Animal Care and Use Committee of the University of Buenos Aires . Adult canaries were acquired from a local breeder . The experiment was performed on 11 male canaries ( Serinus canaria ) either on the left HVC ( N = 6 ) or the right HVC ( N = 5 ) . The morning of the experiment , birds were deprived from food and water 45 minutes prior to surgery . They were given injections in the pectoral muscle with 20% urethane ( 90–120μl total; Sigma , St . Louis , MO ) , administered in three 30–40μl doses at 30 minutes intervals . Immobilization of the bird was done with a stereotactic device . After topical application of xylocaine , the skull of the birds was exposed by sculp retraction . A craniotomy of 2mm x 1mm was made over the corresponding left or right HVC which was located stereotactically and the dura over the recording site was removed . Body temperature was maintained with an electric blanket . Two custom built cooling devices were used , one for placing over the left and the other over the right HVC . Their capabilities are based on a Peltier module , which provides a constant temperature decrease on one side while the other is kept at a constant temperature . Previously , a 2A current supply was used for obtaining a temperature decrease of 10°C [16 , 24] . Such high currents circulating continuously near the recording site interfere with the electrophysiological signal about a minute after onset . To avoid this we increased the Peltier size four times and performed a pulsed current protocol with a period of 6 seconds ( S1 Fig ) . To avoid neuronal damage due to abrupt temperature changes , currents were modified in steps of 0 . 25A and the temperature was allowed to stabilize for at least 3 minutes . We achieved almost constant fluctuating temperatures with a maximum decrease of approximately ΔT = −9°C and a maximum amplitude of fluctuation of less than 0 . 5°C . Calibrations of the devices were made on two supplementary animals not used for the electrophysiological recordings ( S1 Fig ) . We obtained between 4 and 10 different temperatures for 33 different recording sites with an average of 8 . 7 temperatures ( 30 sites with 7 or more temperatures ) . To rule out the possibility of permanent effects due to cooling , we made half the measurements increasing the current and then half going back to 0A . A metal part was designed to allow the most efficient heat transfer between HVC and the cooling device ( S1 Fig ) . It has a pad of 1mm x 2mm surface and 1mm thickness that is placed on the surface of the brain right above HVC with a 0 . 5 mm diameter hole made in its center . Through that orifice we lowered the measuring electrode . Extracellular recordings of spontaneous neural activity in HVC were performed using tungsten microelectrodes ( Parylene-C-insulated , 0 . 8-1 . 2MΩ , MicroProbes ) . The electrodes were lowered between 200 and 800 μm beneath the surface of the brain , because HVC is approximately 1mm thick and is located superficially . Electrode entrance in HVC was identified by its neuron’s characteristic spontaneous bursting patterns . Signals were preamplified 10X near the electrode via a custom built , low power , low noise and battery powered instrumental amplifier [72] . Then this signal was fed into an Instrumentation Amplifier And Signal Conditioner ( Brownlee Precision Model 440 ) , where it was further amplified 100X , bandpass filtered between 300 Hz and 3 kHz and digitized at 20 kHz . The final output was partitioned in 5 to 10 contiguous files and stored on a computer through a data acquisition device ( National Instruments M Series ) . Total recording times for each site range from 2 to 8 minutes ( 4 . 8 min in average , 31 sites above 4 min ) . We implemented single-compartment conductance-based biophysical models of neurons from the HVC following the ones developed by Daou et al . [67] . Simulations were performed in C using a Runge Kutta 4 algorithm for integration with 0 . 01ms steps , and the analysis of the model output was made in MATLAB . Models are of Hodgkin-Huxley-type with additional currents . The membrane potential of the neuron obeys the following equation: C m d V d t = - I L - I K - I N a - I C a - L - I C a - T - I A - I S K - I K N a - I h - I N a p - I s y n + I a p p + I n o i s e ( 1 ) where Cm is the membrane capacitance , IK and INa are spike-producing currents , ICa−L and ICa−T are Ca2+ high-threshold L-type , and low threshold T-type currents , ISK is a small-conductance Ca2+-activated K+ current , INap is a persistent Na+ current , IKNa is a Na+-dependent K+ current , IA is an A-type K+ current , Ih is a hyperpolarization-activated cation current , and IL is a leak current . All these ionic currents were found experimentally in slice electrophysiology experiments at room temperature in all three HVC neuronal types and detailed equations and parameters can be found in Daou et al . [67] . Here we briefly comment on the different currents whose relative weight change across neuron types , shaping their characteristic behavior at room temperature: HVCX neurons show spike adaptation due to large ISK and IKNa currents and have a sag due to Ih current . They show rebound firing or depolarization due to large ICa−T current . HVCRA neurons show lack of excitability in response to depolarizing pulses due to large ISK , IKNa and IA currents , and a delay to spiking due to large IA current . HVCINT neurons show high firing frequency with no adaptation even to small pulses due to reduced ISK , IKNa and IA , a prominent sag and hence a rebound firing due to large Ih current and also fire spontaneously with variety of patterns in slice preparations . Sodium INa and potassium IK also vary among neuron types . The most interesting and novel observation in this description is the role that ISK may play in the bursting behavior of HVCRA and HVCX in vivo , having a slow current dependent on the dynamics of the free intracellular Ca2+ concentration . Although bursting did not appear at room temperature , experiments in vivo where the temperature is around 40°C show clearly their bursting behavior [66] . A typical current is as follows: I c = g c x ( V - V c ) ( 2 ) where Ic is any of the currents described before , gc its maximal conductance , Vc its reversal potential and x is a gating variable whose kinetics are governed by: d x d t = x ∞ ( V ) - x τ x ( 3 ) where x∞ ( V ) is a defined function of voltage and τx is the time constant , with 1/τx a factor that multiplies the rate constants ( usually called k values ) of opening and closing of the ion channel which for some currents has a membrane potential dependency τ ( V ) . For different simulations we used artificially applied external currents Iapp , synaptic currents Isyn and some degree of current noise Inoise . The first is just a current clamp: constant input with a finite duration . The second needs to have a presynaptic neuron releasing neurotransmitter T . Equations used read as follows [75 , 76]: T ( t ) = T m a x 1 + e x p - ( V p r e - θ ) σ ( 4 ) d s d t = α T ( t ) ( 1 - s ) - β s ( 5 ) I s y n = g s y n s ( V - V r e v ) ( 6 ) where a T ( t ) approximates a Heaviside function that turns on with the voltage of the presynaptic neuron Vpre . Parameters θX = −10mV , θRA = −20mV and σX = 7mV σRA = 5mV were selected after evaluating the amplitude and maximum voltage achieved by HVCX and HVCRA simulations in order to have neurotransmitter release at approximately the top 10% of the simulated presynaptic spikes . The rate constants of the gating variable s , α = 2 . 2mM/ms and β = 0 . 381/ms , and the maximal neurotransmitter concentration Tmax = 1 . 5mM were selected as in Gibb et al . [70] . Typical values for Vrev are 0mv for excitatory synapses and -80mV to -100mV for inhibitory synapses , and here we modeled only excitatory synapses . gsyn = 3 . 0nS and 2 . 5nS for HVCX and HVCRA were selected to have slightly less than 100% effectiveness in eliciting a spiking response in an interneuron when noise is present and to be able to reproduce the in vivo recorded firing pattern at a single excitatory synapse as recorded by Kosche et al . [28] . We added to some simulations standard Gaussian white noise: I n o i s e = n I a m p ( 7 ) where n is a random number drawn at each time step from a normal distribution and its amplitude Iamp = 400pA . It was changed in 0 . 1ms steps . Temperature control in the model was made by modifying the electrical properties most strongly affected by temperature: the maximal conductances and the rates of channel opening and closing [77] . To add these effects to the model we changed conductances gc and time constants τx , using the usual Q10 based formalism: p = p r e f Q ( 8 ) Q = Q 10 ( T - T r e f ) / 10 ( 9 ) where the parameter p is a conductance g or a rate constant 1/τ , Tref is the temperature where the parameters were originally fitted with the model from measurements , Q is the scaling factor , T is the temperature and Q10 describes temperature sensitivity . Typical values for Q10 are Q 10 k = 3 for rate constants and Q 10 g between 1 . 5 and 3 for conductances [49 , 50 , 61 , 77] . Exploration was made in these parameters with steps of 0 . 1 from 1 . 1 to 3 . 2 to obtain the best match between experimental and computational spike traces of each HVC neuron . We assumed a value of Tref = 20°C which is mentioned by Daou et al . [67] as “room temperature” , however values up to 25°C do not modify our results . Temperature exploration with parameter T was made between 40°C and 20°C , with the former being the designated “normal” temperature of the bird’s brain . Other explorations were made in the range 40 − 30°C , the latter being almost the lowest temperature obtained experimentally . We also changed temperature in ICa−L that has an explicit T dependence in Kelvin units .
We first classified the SUs by their waveform shape . In Fig 1A we can see all the mean waveforms normalized and aligned to their lowest value . We could separate them into two groups by the mean of their peak to peak width . Widths that were less than 0 . 5ms were classified as Fast Spikers ( FS ) and the longer ones as Regular Spikers ( RS ) . Usually these shapes are related to the neuronal type , with the former classified usually as interneurons , while the latter are usually designated as excitatory neurons ( Fig 1B ) . The first easily visualized effect found in all neurons , was the spike shape widening of the extracellular potential ( Fig 1C ) as temperature decreased from normal at 40°C down to almost 31°C . Single units were isolated at each measurement site by means of spike detection and clustering techniques ( See Methods ) . Mean waveforms were obtained after aligning all the events to the minimum voltage , and spike times were used to quantify activity and inter-spike-interval ( ISI ) distribution characteristics . In Fig 1C we can appreciate five widening examples , three FS and two RS . The widening effect reached up to almost a 2-fold increase from the original shape for FS . To properly assess the effect for the regular spikers , we pulled out from the 18 cell measured the 5 highest firing ( RShf ) and the five lowest firing ( RSlf ) in an attempt to represent HVCX and HVCRA respectively . The rationale for this is the following . Since the wave shapes obtained from extracellular recordings of both types of cells are indistinguishable , we took advantage of the differences in their spike rate activity . Assuming that the probability of measuring any of both cells is equally probable , taking the lowest 5 and 5 highest firing cells gives a value p< 0 . 05 ( p = 0 . 048 ) of not having selected all from the same type . This is given by the binomial cumulative density function at a value of 5 with 18 samples . This selection will prove to be appropriate after the delicate match with the model that we show below . The widening effect was lower in RS units than in FS units , and it was more pronounced in RShf compared to RSlf , as we can see in Fig 1C . The increase was of 40% and 20% respectively ( Fig 1D , S2A Fig shows not normalized values ) . In addition to the widening effect , we could see that the activity decreased substantially in all cases with temperature with no significant difference between cell types ( Fig 1E , not normalized values in S2B Fig ) . However , the way that the activity changed differed for each firing pattern . In six of the FS units , most of the activity above an ISI of 10ms gradually disappeared with lower temperatures , retaining many occurrences below 10ms , as we can see in first column of Fig 1F ( neuron FS 1 ) . Other behavior observed in two FS units is depicted in the second column of Fig 1F ( neuron FS 2 ) . These neurons do not present bursts at normal temperature , and have ISIs which peak at around 20ms . As temperature decreases the distribution of ISIs depletes and shifts to the right . Finally , there were four neurons that did not present the longer timescale , and although decreasing the firing rate , they did not show changes in the ISI distributions ( Fig 1F , neuron FS 3 ) . The existence of different firing behaviors of HVCINT ( FS ) was previously shown by Daou et al . [67] . The quantification of these two timescales and the disappearance of the second one is made evident with the cumulative distribution functions shown in Fig 1G . Second timescale disappears in neuron FS 1 and neuron FS 2 , but not in neuron FS 3 where it was not originally present . To better quantify these timescales we did separate analyses below and above 20ms , which showed that median values in these ranges have different correlation with temperature , rising for above and decreasing for below 20ms ( S2C and S2D Fig ) . This implies that although having an almost 3 fold reduction in firing rate in the range of temperatures studied , the timescale appearing below 10ms does not disappear ( for the neurons that have it ) , and that the timescale around 20ms disappears with decreasing temperatures . In addition , only the neurons having the intermediate timescale showed differences with the cumulative distribution at normal temperature ( S2E–S2H Fig ) revealing non trivial modifications of the ISI distributions . For the two example RS units shown in Fig 1C , the evolution of the ISI distribution retains its structure , but loses it at times higher than 20ms . This is clearly seen in the third row of the last column of Fig 1F ( RS lf ) , where the second peak in the distribution disappears . In all cases we can see that the first peak or left end of the distribution moves slightly to the right , to a higher ISI . We did not find differences in all these effects between the left and right HVC . In what follows , we will study with the help of a computational model of the neurons in HVC , the possible explanations for the temperature effects in the spontaneous activity of its neurons . Computational modeling of HVC neurons has proven to be successful for in vitro approaches [67] . In those studies , single compartment conductance descriptions of currents and voltages were obtained with great detail . Further work also explored non conventional complex input currents designed to activate most ionic channels and data assimilation procedures to better evaluate parameters [68] . These works show astonishing match between experiment and modeling , making them ideal to test temperature extrapolations . As mentioned above , the works done in the crab made a full exploration of the parameter space showing that a very small percentage of models will show the expected physiological behavior , but that they are nevertheless hundreds of parameter combination possibilities . This emphasizes that many models can be equivalent to explain a behavior , and that care has to be taken in restricting constraints . Other modeling variations of HVC neurons include two compartments , soma and dendrites [66 , 69] , to better explain the absence of bursting from current injections in the soma of HVCRA neurons in slices at 35°C [66] . However , dynamics of bursting within a single compartment neuron are expected to not differ from a two compartment one apart from a small delay coming from the electric couping constant between the two compartments ( not more than ∼1ms ) . Current injections in the soma will elicit bursting since the IK Ca dependent current mostly responsible for it is located in the same compartment as the currents responsible for spikes . In our modeling we will use the parameters found by Daou et al . and a single compartment model . Our hypothesis is that there is a parameter range were the responses of our model are robust to changes . Therefore we selected the extrapolation parameters using constraints from previous in vivo studies . Our first step was then , starting from these previous results , to extrapolate them to in vivo brain temperatures . As is usual in most of the in vitro electrophysiological studies , temperature is mentioned as “room temperature” . This fact has consequences , like further studies in the field yielding results that differ from the previous ones , since they work at another “room temperature” . For the purposes of this work we did center our interest in dynamical regimes , rather than in absolute values , since they are expected as we showed in the previous section in the almost 10°C temperature range explored . Our studies focus in the global changes manifested and not the exact temperature at which they occur . Therefore we decided arbitrarily to assign a 20°C value to the temperature where the computational parameters were obtained . As explained in the Methods section , we modified the maximal conductances and the rates of channel opening and closing based on an already successful temperature model for neuronal activity [43 , 49 , 50 , 61] . Search in parameter space was done with the premise that different Q10 values act on the maximal conductances and the rate variables , and the search for the proper parameters was done independently for the three neuronal types . We aimed at reproducing previous intracellular recordings that show bursting in excitatory cells and their specific voltage evolution [63 , 66] . We decided to not separate each of the specific Q10 values for gating and conductance variables ( 7 for rates and 10 for conductances ) , since finding a combination that reproduces the expected behavior should be robust under slight changes in the parameters [43] . Therefore , we made a two dimensional parameter search with the two Q10 parameters decoupled for the three neuronal types . Q10 k ( rate ) and Q10 g ( conductance ) were allowed to vary from 1 . 1 to 3 . 2 in since 1 means no effect by temperatures , and 3 is the usually biggest reported value . We explored the temperature range from T0 = 20°C to the normal temperature of 40°C . We injected 16ms length currents with 3 varying steps . Criteria for selecting parameters was capability of bursting and number of spikes and measures of voltages between resting , maxima , minima and local minima . Finally , we assessed resemblance with voltage traces from the literature . First , criteria needed to be met at 40°C and then we checked for robustness across other temperatures . The easiest search was made in HVCRA , because it presents a narrow range of parameters where it presents bursting ( Fig 2A ) . The selected value is ( Q10 k , Q10 g ) = ( 3 . 0 , 1 . 4 ) . Second , HVCINT ( Fig 2B–2D ) presents strong spiking at all parameters , and a region with strong firing and robustness across temperature was selected . Then we looked for the difference between minimum hyperpolarization after spike and resting membrane potential , as minimum difference was found in in vivo studies [63 , 66] . As a last constraint , we selected a region were the peak to peak voltage was neither too high neither too low and about 60-80 mV . Values for the interneuron are ( Q10 k , Q10 g ) = ( 3 . 0 , 2 . 5 ) . Third , we studied the behavior of HVCX which needed further constraints to arrive to the selected values ( Fig 1E–1H ) . Higher and more robust spiking across temperatures was found first and then we assessed interburst interval to be lower than 4ms . Peak to peak voltage inside a burst was chosen to be between 40 and 60mV , because we know it spans from about half to 75% of its 80mV peak to resting membrane amplitude . As a last measure , we evaluated the peak to peak width to be bigger than 0 . 6ms , as suggested by our measurements ( see S2A Fig ) , and to be smaller than 2ms which is about the half of the interburst interval . Parameters chosen from the intersection of selected zones are ( Q10 k , Q10 g ) = ( 3 . 0 , 2 , 5 ) . We did also explore other parameters as the resting membrane potential ( S3 Fig ) . The only neuron to show a variability bigger than 1mV across all the parameter space and from 20 to 40°C was HVCX . In that case , the variation in the range of experimental temperatures studied was of almost 4mV , and it increased at low temperatures . The other neuronal types also showed this increase , although it is of less than 1mV . This behavior is compatible with the measurements on hippocampal slices mentioned above . The obtained Q 10 k = 3 factor for rate constants was the same for the three neuronal types , and the evolution of Q with temperature can be seen in Fig 3A . It represents probability of channel opening and closing effects and is the typical value measured in the bibliography and used in models [49] . On the contrary , the temperature sensitivity Q 10 g factor was different for each type of neuron . This is not surprising , as it should depend on the type of channel that predominates in each neuronal type . The evolution of the Q values are shown in Fig 3B for these parameters . The predominant conductance IA in HVCRA is bigger compared to its Ih and the contrary happens for HVCX , so relative values coincide with conductance measurements made in the crab [49] , where values for the IA conductance are lower than 2 and near to 3 for the Ih . Because our modeling does not discriminate between individual Q 10 g , we expected the relative incidence of these two conductances to dominate above the other less representative . The value of 2 . 5 for HVCINT is similar to the in vivo relative current values of HVCX . The final selection of parameters was assessed by finding the closest fit between modeled and experimental traces at in vivo temperatures ( See Fig 3C ) . Not trivially , yet not surprisingly as was expected from the recently described dynamics of Ca2+-dependent K+ current ( ISK ) in HVCRA and HVCX , we found bursting behavior at normal bird temperatures of 40°C for these two neuronal types . We used external applied currents Iapp of durations close to 10ms , which is the typical burst duration measured in vivo . Once the temperature control variables Q 10 k and Q 10 g were set , we studied the behavior between 30°C and 40°C . We found that for the same duration and strength of current stimulations , the number of spikes per burst decreased and interspike-burst-time increased , and also interspike-interval increased for HVCINT ( Fig 3C ) . For room temperatures at which the original studies were performed , the bursting behavior disappeared for HVCRA and changed drastically for HVCX , showing only two spikes . Another interesting behavior shown was the delay onset for the first spike in HVCRA bursts , that may have specific implications in explaining previous temperature manipulation experiments [16 , 24 , 25 , 33] which we will discuss in the next section . To test if continuous spiking was just a response to applied current , we used a longer stimulus of 30ms ( Fig 3D ) . Interestingly , spiking terminated at approximately the in vivo durations found experimentally . As it was reported by Daou et al . [67] we did not find bursting behavior at the lowest temperature , and we were able to find a range between 30°C and 40°C where HVCRA keeps its bursting offset and between 36°C and 40°C where HVCX does . HVCINT just decreased their interspike times . We looked for post bursting effects and applied the short current pulses described above after a brief time gap from the long pulse . The behavior of bursting neurons showed a reduced number of spikes , even with increased current . To elicit bursting we needed a 100ms refractory period for HVCRA , while only 50ms was needed for HVCX , which is compatible with in vivo bursting recordings . On the other hand , HVCINT never showed bursting and interspike interval was sensitive to current values , resulting in higher spike rates at higher currents . After fixing the model parameters of temperature control , we explored single spike features of neurons . Spike width increased with temperature decrease for the three neuronal types , as can be seen in Fig 4 . Computational traces show the intracellular voltage evolution of a spike that increases time duration . In Fig 4A we focus on a single spike for several temperatures for the three neuronal types for the current injected in Fig 3A . Since the computational model generates intracellular voltage evolution , to compare it correctly with the extracellular measurements ( S2A Fig ) , we selected a width value for each neuron type corresponding to the one measured experimentally . For excitatory neurons , we searched at 40°C peak to peak threshold % to obtain a width of 0 . 6ms . This resulted in a 60% for HVCRA and a 55% for HVCX . For HVCINT we decided to include the hyperpolarization peak which has strong influence in the extracellular potential . We then made a threshold selection in the uprise of the spike , giving a value of 30% to give a width at 40°C of 0 . 35 ms . Results are shown in Fig 4B , where the resemblance of the evolution of the widths across temperature is impressive . Following , we computed the relative increase in spike width compared to the one at 40°C ( Fig 4C ) . The increase in spike width was more than 100% for HVCINT , 35% for HVCRA and 50% for HVCX for the temperature range studied . At values above 31°C they have a remarkable match with the one that was measured in experimental extracellular traces ( Fig 1D ) . Behavior of the model matches the one present in our data and is sufficient to explain the waveform widening using a single neuronal model . This means that this effect can be explained with the temperature incidence in the temporal channel subunit dynamics , which increases their characteristic time constant with the factor 1/Q ( see Methods ) . On the other hand , the maximal channel conductance changes that are modified with the Q factor may be responsible for other effects . Since the Hodgkin Huxley type of modeling used represents an excitable system , we expect that this maximal conductance reduction will delay the neuronal synaptic inputs to situate the postsynaptic neuron voltage above the threshold , thus producing a spike delay . To analyze this effect , we explored the voltage evolution of neurons for the current injections of Fig 3C . A detailed evolution of bursts is shown in Fig 5A , where voltage traces are aligned to the start of the stimulus . Here it is easier to recognize that the interspike interval within a burst increases when lowering the temperature . Also , for HVCX and HVCINT , there is no delay in the onset of firing , which changes for HVCRA . For this reason , we made the input current to HVCRA longer , in order to have a three spike burst across all explored temperatures . This adds a distortion effect on the last spike at low temperatures , but does not alter the results . We can see that there is close to a 5ms delay in the onset of firing for the same current applied when temperature is decreased 10°C . These can directly affect the firing pattern of neurons targeted in RA , which in turn innervate the brainstem nuclei ( nXII for syringeal muscles and Ram for respiratory muscles ) in charge of sending output motor activities to elicit song , or sending back further synapses through a bottom up recurrent network . Although the type of input we are providing is not physiological , it proves that temperature is capable of changing an important timing property of at least one neuronal type , HVCRA . For the purposes of this work we will concentrate on the HVCINT temperature changes , since there is still much debate on the connectivity of HVC projecting neurons , as mentioned above in the Introduction , and because it will restrict the connectivity scenarios we can use to assess the changes in their ISI distributions , as we explain in the next section . Therefore , we used the voltage profile of the excitatory neurons HVCRA and HVCX to provide a more realistic input to HVCINT . Using a simple model of synaptic current also modified with temperature ( see Methods ) , we assessed the evolution of the HVCINT neurons . The parameter we had to explore in this single synapse model was the maximal conductance gsyn ( see Eq 6 ) . To match physiological behavior , we tried to reproduce the recently measured in vivo-like activity shown by Kosche et al . [28] of interneurons in HVC while measuring and eliciting the excitatory input activity in HVCRA neurons . The authors showed that reliable spiking occurs in interneurons when HVCRA neurons fire three consecutive spikes at 100-300Hz , at a room temperature of 23°C . The single interneuron spike occurs after the third spike in the “artificially elicited burst” , and shows a short delay . Once the conductance parameter was fixed , we computed the evolution of an interneuron voltage following the different excitatory inputs of Fig 5A . The full current and the voltage evolution are shown in Fig 5B for both cases . We can notice that the interneuron spike has a much wider delay relative to the position of the first spike in the excitatory bursts ( vertical line ) . This is due to the accumulation of time resulting from the widening of spikes and interspike intervals , and from the effect of synaptic delays . The delays relative to the first spike in each condition in Fig 5A and 5B are depicted in Fig 5C . Hamaguchi et al . have recently measured in HVC at 32°C an increase in excitatory neuron delays to spiking onset of almost 2 . 8ms compared to 40°C , using presynaptic neuronal stimulations [33] . This agrees with the HVCRA delay shown here . Following the analysis , the maximal postsynaptic currents in this model were computed , and in Fig 5D we can see that they decrease from the original absolute values of around 600-700pA to close to 300pA for the two excitatory inputs on HVCINT . This remarkable change can help us explain the decrease in spontaneous activity that we witnessed in our experiment . For this , we explored the response capabilities of the HVCINT to different input currents at 40°C , as we see in Fig 5E . Here we applied currents starting from -100pA down to -800pA with steps of 100pA . We see that the modified Hodgkin-Huxley model used for this neuron is particularly sensitive to the current applied , showing an increasing spiking frequency , and can be classified as a class 2 excitable system , since it starts spiking at low frequencies with small currents . To see the incidence of temperature and current in the ISI interval elicited in the HVCINT we constructed Fig 5F . There , we explored the full temperature range , and currents from 300pA to 2300pA . The first interesting result is that the ISI has almost no sensitivity to temperature when high currents are applied . Only below 500pA we can see a change higher than 1ms for the whole temperature range . It follows that temperature starts to play a role in ISI behavior at low input currents , which happens below 36°C , as we know from Fig 5D . For the lower current of 300pA , the ISI can change up to 5ms . In conclusion , there are two parallel effects acting on the synaptic model: one is the decrease in the maximal conductance with temperature , and the other is the sensitivity to input current of the HVCINT . These two effects may act together in reducing the spiking of these neurons at lower temperatures . Nevertheless , we cannot yet explain the different shapes in the ISI histograms of the experiment , for which we refer to the next section . To account for different neuronal activity characteristics shown by HVCINT , we decided to explore realistic synaptic inputs . A priori , we cannot rule out the possibility that the HVCINT have subpopulations with dissimilar channel conductance properties , however , the existence of substantial diversity has not been reported in the interneurons measured previously [67] . Following this line , and to simplify our exploration , we expanded the single synaptic model to multiple excitatory inputs on a single interneuron . We based our analyses in the synaptic connectivity description of HVC microcircuitry where it is known that HVCINT only receive input from HVCRA and HVCX [64] . We connected synaptically excitatory neurons to a single interneuron whose firing pattern is under study . The hypothesis we want to test is that only with Poissonian inputs into the interneurons we can account for the changes observed , ruling out any connectivity structure , at least during the non-singing anaesthetized state . We fixed the spike rate of excitatory neurons as reported previously at 0 . 6 Hz and 1 . 5 Hz for HVCRA and HVCX respectively [63] . Each input consists of a presynaptic voltage pattern eliciting a synaptic current , as shown in the previous section ( Fig 6A–6C ) . We used the patterns of Fig 5A and made two new ones , reducing the stimulus pulse , which have a single spike or a burst with two spikes ( Fig 6A–6C , waveform shapes ) . These are intended to mimic real input for canaries that reported that about 9% of a bursting cell’s activity corresponds effectively to bursts and that they average around 2 . 7 spikes per burst . We therefore used for each synaptic input a fixed ratio between number of spikes . Events with one spike ranged from 70% to 100% in 10% steps , and the remaining were assigned 2 or 3 spikes per burst , retaining the desired overall spike rate . Conductance parameters were the same as the ones studied in the section above , which means that single spike events cannot elicit a spike in the interneuron . In order to be more realistic , we added noise following Eq ( 9 ) . In addition to the temperature change of all parameters in the computational model , we changed the synaptic input spike rate of the excitatory neurons following the decrease shown in Fig 1E , which is of 5 . 6%/°C . We explored at normal temperature the number of inputs necessary to keep the spike rate of the interneurons between 2 and 8Hz ( measured experimentally ) . It resulted with the parameters used that we needed 80 excitatory units . Then we allowed the number of different excitatory types to vary in the range to see if different input characteristics can shape the quality of the response . In Fig 6A–6C we show an arrangement and the typical traces of intracellular potential showed by the HVCINT and its corresponding synaptic current , for three different arrangements ( top ) of input neurons at 40°C ( middle ) and at 32°C ( bottom ) . We can see the intracellular excitatory post synaptic events in the trace and that some events get grouped in three and two spike “burst” . This “bursting” behavior is due to the way that we fixed the synaptic input in the previous section and to the intrinsic characteristics of the interneuron . Consequently , when a presynaptic event is a burst , there is a higher chance that another input from another neuron near it will elicit more spikes . This can be seen in Fig 6D where there is a dense occurrence below 10ms . This type of pattern is reminiscent of FS 1 unit described in Fig 1F . On the other hand , when we take out the bursting input type , we cannot find the “burst” arrangement in the intracellular trace ( top of Fig 6B ) and we recover a pattern without the first peak in the ISI distribution , as shown in Fig 6E . In Fig 6C we show the trace and in Fig 6F the ISI distribution of another neuron corresponding to a different balance of HVCRA and HVCX input events . In all three cases , what determines the type of ISI distribution is not the type of input neuron , but the balance between the number of 3- , 2- 1-spike inputs . We found that the only possible way to avoid the grouping of events below 10ms was by preventing bursting input to the neuron . On the other hand , once the balance was established , changing the ratio of excitatory input neuron types only changed the spike rate , without changing qualitatively the ISI distributions . In Fig 6G–6I we show the quantification of the distributions by means of their cumulative density function . In Fig 6G we see that although the structure below 10ms is retained , significant differences exist between low temperatures and the normal temperature . This is due probably to the very deterministic nature of the bursting input imposed into the cell that makes the interneuron interspike interval to increase as the interburst interval of the input does ( S4A Fig ) . More explorations using a bigger percentage of bursting events made this differences disappear and the distribution under 10ms to be slightly broader . However , when studying the firing rate decrease across temperatures for these inputs with more bursting input , it was not possible to match the firing rate decrease found in the experiment . In Fig 6H we see a resemblance to FS 2 type from Fig 1F in that the slope at small timescales of the distribution decreases at lower temperatures , but in this case no significant differences were found with respect to 40°C at times above 20ms ( S4B Fig ) . In Fig 6I we show a third combination of inputs that matches FS 3 behavior . Again in this situation as in Fig 6G , significant differences are found for smaller ISI values ( S4C Fig ) . Then we studied the evolution across temperature of the normalized spike rate for all the measured temperatures and balance of neurons and spiking inputs . We can see in Fig 6J ( in S4D Fig we show not normalized values ) that the decreasing tendency which better resembles the measured data ( Fig 1E ) is the one corresponding to all the inputs as HVCX . We have imposed a lower spiking rate on HVCRA relative to the one of HVCX , which in the simulations with a higher ratio of HVCRA neurons led to a firing rate that was at the lower end of the range of the ones measured experimentally . We conclude that for the physiological range of firing rates we simulated , their decrease can be explained by the amount of the stochastic input received . Then we asked if it was more important the specific type of input that a cell provides or if it was the firing rate at which it excites the interneuron . For that purpose , we inverted the firing rate between HVCRA and HVCX inputs and rerun the simulations . The result can be seen in Fig 6K ( S4E Fig ) , and we obtained an exact inversion of the original result , meaning that it is more important the rate of the input than its detailed characteristics . The only difference was the case where all 80 inputs were HVCX , in this scenario at a low firing rate of 0 . 6Hz , where the firing rate of the interneuron increases at low temperatures . This can be explained by the fact that HVCX input was slightly less strong than the one from HVCRA and that at low firing rates it cannot elicit spiking easily , whereas at low temperatures it shows a bigger amplitude of spike and HVCINT a slightly bigger membrane resting potential ( S4I Fig ) . Nevertheless , firing rate of the interneuron in that case was far below any of the values measured experimentally ( Fig 1B ) . We also studied what happens if the synaptic inputs are stronger ( +0 . 5nS for each synaptic input ) and found that no matter all the combinations that we explored , the interneuron does not reach the rate decrease observed experimentally ( Fig 1L and S4F Fig ) . Finally , we took out the noise present and rerun again the simulations , and we saw that results do not change regarding the firing rate ( S4G and S4H Fig ) . The sole difference was that the spread of the ISI distribution was narrower .
In this work we studied the neuronal changes produced by the temperature manipulation of the neurons present in HVC of canaries during spontaneous activity . The main purpose of this was to understand the individual underlying mechanisms that take place in single HVC neurons when cooling , which was shown to give rise to complex behavior such us song stretching and “breaking” . We could recognize from the extracellular electrophysiological recordings two different types of single units , Regular Spikers ( RS ) and Fast Spikers ( FS ) , that correspond to excitatory HVCRA and HVCX , and inhibitory interneurons HVCINT . In order to assess separately the behavior of the two excitatory neurons , we made two groups , RShf and RSlf for high and low firing rates . This proved to be a good choice based on the significant differences encountered between the two populations and the delicate match with the computational model . All these single unit waveform shapes increased their width with temperature . However , the three groups exhibited a distinct behavior: interneurons showed an almost 2-fold increase , while excitatory ones increased only 20 and 40% of their length in the almost 9°C experimental temperature range . The second marked effect was an almost 3-fold decrease in the spike rate of every measured neuron , including in this case the Multiunits ( MU ) . Finally , we observed distinct inter-spike-interval distributions that changed across temperatures in an unexpected way . First , FS showed three families of activity patterns: two with a marked peak at low ISIs , from which one had also a timescale present at around 20ms , and the third with a spread distribution . All FS neurons retained the first peak while depleting the higher times . In the case of the one only having the second spread timescale it also shifted to the right . For the RS , some sparse patterns shifted to the right also , and depleted first at higher ISI values . In order to understand these observations , we proceeded to build a computational model with detailed equations and real conductance parameters . Since these were obtained at room temperature for HVC neurons , we introduced the Q10 formalism to be able to model temperature changes . A search in the ( Q10 k , Q10 g ) parameter space was done for the three HVC types independently using constraints from previously reported in vivo measurements and values selected displayed a robust behavior across the experimental temperatures measured . The first result obtained from this implementation was the natural occurrence of bursting behavior for both excitatory neuron types at normal temperatures . This may be the result of the slow effect of the ISK current dependent on the Ca2+ intracellular concentration which is present strongly in these neurons , but not in the inhibitory type . We then computed the evolution of the waveform shapes across the range of temperatures explored experimentally . The widening effect was present in all the simulations and the three groups showed a marked difference of widening and a delicate match with our measurements . It proved sufficient to use a single neuronal model with temperature dependence to explain this effect . We then added artificial pulses of constant synaptic input to the three neuronal types and found a marked delay in the onset for spiking in HVCRA neurons . Then we used the voltage evolution of the excitatory neurons to provide realistic synaptic input for the interneurons . Matching the experimental observation that a burst of three presynaptic spikes elicits one spike in the postsynaptic neuron , we could see that the delay onset for spiking with respect to the one at normal temperature increased up to 10 ms at the lowest temperatures . We observed that the maximum synaptic current decreased with temperature in a range from 300 to 700pA . Interestingly , the interneurons show a strong dependence with temperature in their interspike interval only for currents under 500pA . This can already explain the decrease in spike rate observed in the interneurons without any assumptions on the changes of their input . Finally , we added Poissonian synaptic inputs to see if the changes in the inter-spike-interval distributions of the interneurons could be explained . We found that at least 80 synaptic inputs are needed to reproduce the spontaneous spike rate of the HVCINT due to their low firing rate ( 1 . 3Hz in average for our measured RS units ) . To reproduce the ISI distributions with a marked peak below 10ms , the most important feature needed in the presynaptic neurons was the existence of at least 10% of the input neurons having a burst of two or three spikes . In order for this peak to be absent , 100% of the presynaptic activity had to be modeled with a single spike . This provides an interesting tool to study the type of input that this type of neuron might have , only by measuring its spontaneous activity . In addition , all the simulations showed a shift of the first peak of the distribution to the right . What could not be explained with the stochastic input proposed was a second strong timescale present in the ISI histograms , the one which is above 15ms . This can be due to the fact that connectivity in HVC is not random , and that its structure manifests even when its activity is spontaneous . We also explored the relationship between the type of input and its firing rate by interchanging the firing rates of HVCX and HVCRA . Results showed a very similar behavior of the interneuron activity , but now with the number of inputs inversed . This implies that at the network description level the most important factor is the spiking activity , and not the detailed intracellular dynamics . In terms of noise input , we found that it does not change rate results , but that it has incidence in the spread of the lower peak of the ISI distribution , as expected from a very deterministic input . Lastly , we also tested stronger synaptic inputs and found no combination that could reproduce the amount of firing rate decrease measured at low temperatures . This points out the importance of the delicate choice of the synaptic strength so that it elicits an EPSP in the interneuron just below its spiking threshold: a very fine tuned synaptic strength is needed to reproduce our experimental findings . In terms of cellular mechanisms , our work provides strong evidence that the bursting behavior of excitatory neurons is an intrinsic property arising at the normal temperature of the brain . We see the way that a lot of the research made on cellular properties of neurons has a big reportability problem regarding temperature . The use of “room temperature” should be avoided , since our computational exploration shows drastic changes . Nevertheless , we showed that the Q10 formalism proves to be a very useful tool to extrapolate the parameters measured for ion channel properties . We strongly advise that further research should report a precise value of the temperature , and ensure its stability . Regarding the slowing down of the cellular processes , we found that the bigger effect is the synaptic delay . This changes the onset spiking of neurons as late as 8ms at an 8°C temperature decrease for the proposed HVCRA to HVCINT synapse . For HVCRA the effect was of more than 3ms , a value compatible to one measured in vivo recently [33] . Spike shape widening can be as dramatic as doubling its value for HVCINT , but this represents only 0 . 5ms for a 10°C decrease . When studying the distribution of spikes during spontaneous firing , we found that only stochastic Poissonian input cannot completely shape the distribution of ISIs measured . There is a second timescale present at values bigger than 10ms that we could not reproduce . This suggests that the underlying connectivity present in HVC does shape this pattern on a timescale bigger than a single spike . Although it was not the purpose of our work to unveil the microcircuitry within HVC , we added evidence that spontaneous activity is not just random . The temperature manipulation tool could prove very useful to test different scenarios in further studies . We would also like to emphasize that our results show contributions to shaping the activity of HVCINT neurons from HVCX neurons that cannot be disregarded . We are not aware of any work about network modeling in HVC that takes HVCX neurons into account . We will not be surprised in the future if we see the inclusion of them in the modeling , just the same that happened when early models of HVC only used HVCRA neurons and now all added HVCINT neurons . Coming to the initial question that triggered our work , we could not find a striking effect in the cellular mechanisms that could be related to the “breaking” of the syllables found in canaries when cooling HVC . Our results suggest that this enigma will be resolved examining network effects either at the local network in HVC or at the global network involving the whole motor pathway . We did in fact find a local unexpected emergent property that changed with temperature: the disappearance of the second timescale in the interspike-interval-distribution of the interneurons . This happened at around 36°C , which is about the temperature when “breaking” effects start to manifest . Since our modeling demonstrates that poissonian synaptic inputs are not sufficient to reveal this second timescale , it is parsimonious to hypothesize that some serial activation of HVC excitatory neurons occurs , and that a group of them that are separated by no longer than 30 ms of synapses are connected to the same interneuron . If this is the case , the absence of the second timescale at lower temperatures could point out the impossibility to sustain reliably this activity in the excitatory neurons . The two possibilities that we foresee , apart from the expected delay from synaptic output and onset , is that some neurons continue their song related activity and that some fail to spike . From the “clock” model perspective , these defects can propagate downstream in the motor pathway to produce the “breaking” . However , this timescale could be not related to song production and may be used ( or not ) in other manner . Another possibility is that this second timescale provides robustness to the system in sending synaptic signals downstream by recruiting many cells in a short time window , and that its absence does not preclude the instruction to exit from HVC . In view of the “circular model” , the breaking effect can arise at the whole motor pathway network only from the delay of spiking and the widening of the bursts . Given the case that both models have some insight into the true mechanism , it could be that short feedforward chains are produced for the generation of small parts of the song , and that the circular network flow is continuous . It would be interesting to test this hypothesis under a paradigm where the spike timing of the cells could be related to song structure . Future work can aim to asses how the timing of phasic firing neurons that respond to song playback is modified with cooling , or how this occurs during singing . From a methodological perspective , we believe that the dynamics of the relevant behavior of the neurons at a network level are already present in their spiking pattern . Two of these behaviors are measured and reported in our study: spike rate and interspike interval distributions . In addition , Q10 values are robust enough to reveal the relevant mechanisms of the neuronal dynamics without the need of having their exact values . Finally , we want to point out , that the use of computational models allows to find a region of parameter values where a system behaves in a similar manner , although not having full knowledge of every parameter in play . The work of Daou et al . has done an absolutely great job in measuring experimentally and fitting computationally this parameters , and our incremental experimental and modeling approach makes educated extrapolations and it enlightens some aspects of temperature and network modeling that were not taken into account until now . To conclude , we provided a detailed model of single neurons , synaptic connections and temperature manipulations that proved capable to explain experimental measurements . Our work provides further insight into how intrinsic cellular mechanisms may take part in the emergence of global activity patterns , that in turn produce behavior .
|
The study of the neuronal mechanisms that give rise to the complex behavior of singing in birds has been hotly debated lately . Many models have been tested and novel tools have been developed to try to understand the role of a key brain nucleus in the song pathway: HVC . It is believed that it is highly responsible for generating the precise timing of songs , and this has been tested by manipulating it with temperature . Results showed that cooling can stretch , but that it can also restructure or “break” the song syllables . However , single neuronal mechanisms are not yet described . To better understand this , we cooled HVC in canaries and measured spontaneous activity electrophysiologically . We found three effects: spike shape widening , spike rate reduction and changes in inter-spike-interval ( ISI ) distributions . To explain them , we built a computational model with a detailed description of ion channel conductances and temperature dependency . We could explain the first effect with a single neuron model . The second , could be explained adding single synapses . Finally , we showed similar ISI modifications of one of the timescales present by means of multiple stochastic inputs . In addition , we found that excitatory neurons show natural bursting behavior at normal brain temperatures and that synaptic delays are the main candidates to explain song stretching at low temperatures .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"action",
"potentials",
"medicine",
"and",
"health",
"sciences",
"ecology",
"and",
"environmental",
"sciences",
"nervous",
"system",
"membrane",
"potential",
"vertebrates",
"electrophysiology",
"neuroscience",
"animals",
"physiological",
"parameters",
"body",
"temperature",
"computational",
"neuroscience",
"interneurons",
"ecological",
"metrics",
"birds",
"animal",
"cells",
"cellular",
"neuroscience",
"cell",
"biology",
"ecology",
"anatomy",
"synapses",
"physiology",
"neurons",
"q10",
"temperature",
"coefficient",
"single",
"neuron",
"function",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"canaries",
"computational",
"biology",
"amniotes",
"neurophysiology",
"organisms"
] |
2017
|
Temperature manipulation of neuronal dynamics in a forebrain motor control nucleus
|
Histone posttranslational modifications ( HPTMs ) are involved in chromatin-based regulation of fungal secondary metabolite biosynthesis ( SMB ) in which the corresponding genes—usually physically linked in co-regulated clusters—are silenced under optimal physiological conditions ( nutrient-rich ) but are activated when nutrients are limiting . The exact molecular mechanisms by which HPTMs influence silencing and activation , however , are still to be better understood . Here we show by a combined approach of quantitative mass spectrometry ( LC-MS/MS ) , genome-wide chromatin immunoprecipitation ( ChIP-seq ) and transcriptional network analysis ( RNA-seq ) that the core regions of silent A . nidulans SM clusters generally carry low levels of all tested chromatin modifications and that heterochromatic marks flank most of these SM clusters . During secondary metabolism , histone marks typically associated with transcriptional activity such as H3 trimethylated at lysine-4 ( H3K4me3 ) are established in some , but not all gene clusters even upon full activation . KdmB , a Jarid1-family histone H3 lysine demethylase predicted to comprise a BRIGHT domain , a zinc-finger and two PHD domains in addition to the catalytic Jumonji domain , targets and demethylates H3K4me3 in vivo and mediates transcriptional downregulation . Deletion of kdmB leads to increased transcription of about ~1750 genes across nutrient-rich ( primary metabolism ) and nutrient-limiting ( secondary metabolism ) conditions . Unexpectedly , an equally high number of genes exhibited reduced expression in the kdmB deletion strain and notably , this group was significantly enriched for genes with known or predicted functions in secondary metabolite biosynthesis . Taken together , this study extends our general knowledge about multi-domain KDM5 histone demethylases and provides new details on the chromatin-level regulation of fungal secondary metabolite production .
Chromatin is the natural substrate for all eukaryotic nuclear processes such as transcription , replication , recombination or DNA repair . Chromatin structure is necessarily dynamic and the underlying mechanisms involve remodeling of nucleosomes as well as depositing and removing posttranslational modifications on N-terminal and central residues of histones proteins ( HPTMs ) present in the nucleosome octamer [1–4] . Some of these histone marks , such as acetyl groups on lysines , profoundly influence the chromatin landscape by neutralizing the positive charge of histones thereby weakening the interaction between nucleosomes and DNA and increasing chromatin accessibility [5] . HPTMs also work indirectly by providing binding sites for chromatin-associated proteins that promote or inhibit specific genomic functions . Notably , many HPTMs recruit additional chromatin-modifying enzymes that add new or remove existing marks , enabling cells to dynamically regulate chromatin structure in response to environmental or developmental cues . Fungi have served as model systems for chromatin studies and in many basic mechanisms they are similar to higher eukaryotes but in some aspects they are quite different and this fact allows evolutionary insights into the development of chromatin regulatory systems ( reviewed in [6–8] ) . For example , there is ground-laying work from the filamentous ascomycete Neurospora crassa , where the molecular machinery relating heterochromatin formation and DNA methylation was deciphered [9–12] . Similar to animals also in N . crassa Heterochromatin Protein 1 ( HP1 ) , docks on di- or trimethylated lysine-9 on histone H3 ( H3K9me2/3 ) to promote heterochromatin formation [13 , 14] and in addition is important to maintain H3K27me3 , another repressive mark , at facultative heterochromatin [15 , 16] . This mark was found to span 6 . 8% of the fungal genome [17] corresponding to over 700 transcriptionally repressed genes , some of which are upregulated upon deletion of the H3K27 methyltransferase [16 , 17] . While H3K27 methylation and elements of Polycomb Repressive Complex 2 ( PRC2 ) responsible for depositing this mark are present in Neurospora and the Fusarium group of fungal pathogens ( see below ) this silencing mechanism has not been detected in Aspergillus species [18] . In addition , DNA methylation has not been found in the Aspergilli although a cytosine methyltransferase is functionally expressed in A . nidulans and has a role in regulating sexual development [19] . Mycotoxins , antibiotics , pigments and other low molecular weight natural products are summarized under the term of secondary metabolites ( SMs ) . The Fusarium and Aspergillus genera are large groups of fungi comprising important plant and animal pathogens and they all produce ( SMs ) at certain developmental stages or under conditions of growth restriction , nutrient limitation and environmental stress ( reviewed in [20–23] ) . It was shown initially in Aspergillus nidulans by genetic analysis that expression of the corresponding SMs biosynthetic genes , which are usually organized in gene clusters , is under chromatin control ( reviewed in [24] ) . Under conditions of active growth SMs genes are silenced by H3 deacetylation [25 , 26] as well as by the H3K9 methylation machinery of ClrD ( KMT1/ DIM-5 homolog ) and the hpo homolog HepA [27] . Interestingly , H3K4 methylation and a subunit of the COMPASS complex which are usually known to be associated with gene activation , also contribute to silencing although this has only been observed for a small subset of SM genes [28] . Several recent studies in a number of other fungi have implicated heterochromatin as a regulator of secondary metabolism and the production of virulence factors . In the plant pathogens F . graminearum ( wheat and maize pathogen ) and F . fujikuroi ( rice pathogen ) as well as in the fungal endophyte Epichloë festucae , H3K9me3 and H3K27me3 regulate expression of specific gene clusters responsible for the production of secondary metabolites [20 , 23 , 29–32] . H3K9me3 and HP1 were also shown to negatively regulate other virulence factors such as genes encoding small secreted proteins ( SSPs ) in Leptosphaeria maculans [29] . How HPTM patterns change as SM clusters switch from a repressed state to an active state is not completely understood . The requirement of histone H3 and H4 acetylation for SM gene expression is well documented in Aspergillus species through HDAC inhibitor studies and SAGA- complex mutants [33–35] . Interestingly , co-cultivation of A . nidulans cells with Streptomyces rapamycinicus led to an anomalous activation of several SM genes in the fungus [36] and this process is correlated with increased H3 acetylation of the corresponding genes and strictly dependent on GcnE , the catalytic subunit of the A . nidulans SAGA acetylation complex [37] . Also in F . fujikuroi , activation of the GA , bikaverin and fumonisin clusters was correlated with increased acetylation of H3K9 [38] . In contrast to acetylation , the role of histone methylation in fungal SM gene expression is much less clear . In F . graminearum , silent SM clusters are highly enriched for repressive H3K27me3 , whereas trimethylated H3 lysine 4 ( H3K4me3 ) , an activating mark , is apparently excluded . Upon deletion of the H3K27 methyltransferase kmt6 , the silent fusarin C and carotenoid clusters are activated , but H3K4me3 does not accumulate in these clusters [30] . A similar situation was shown in F . fujikuroi where increases in H3K4me2 were only observed in two genes of the gibberellin ( GA ) cluster . Similar to the case for H3K4me , expression of SM cluster genes in F . graminearum was not associated with increased H3K36me3 [30] . In contrast , H3K36me3 was gained for the sterigmatocystin ( ST ) and several other SM gene clusters in A . nidulans during activation [18 , 39 , 40] . H3K4me3 is an HPTM with important roles in transcription and this mark is generated by the COMPASS ( Complex associated with Set1 ) protein complex containing the Set1 methyltransferase catalytic subunit in addition to several regulatory and scaffold proteins [41] . COMPASS is not essential in A . nidulans although synthetic lethality of Set1 and Swd1 subunits was found with mutations in mitotic regulators [42] . Generally , H3K4me3 has been shown to be recognized by three different domains associated with proteins of various functions . One recognition module is the PHD domain , present for example in the “Inhibitor of Growth” ( ING ) protein , which recruits histone acetyltransferase ( HAT ) and deacetylase ( HDAC ) complexes [43 , 44] . H3K4me3 is also recognized by the double TUDOR domain of JMJD2A , a JmjC family demethylase that removes methyl groups from di- or trimethylated H3K9 [45] and by the tandem chromodomain of CHD1 , an ATP- dependent nucleosomal remodeler [46] recently shown to be necessary for inhibition of intragenic initiation or initiation from cryptic promoters and thus maintaining normal transcript elongation [47] . Accordingly , H3K4me3 plays a central role in the chromatin regulatory network . Usually , H3K4me3 peaks at the transcription start sites ( TSSs ) and its occurrence is correlated with gene expression [48] . However , the Set1 protein also displays some moonlighting activities as it recruits deacetylase activity independently from the H3K4me3 mark and subsequently promotes heterochromatin formation and transcriptional repression at distinct loci in the fission yeast genome [49] . This evidently negative role of the COMPASS was also documented for regulation of SMs production in three different Aspergillus species carrying genetically engineered COMPASS mutations [28 , 50 , 51] . Silencing specific SM gene clusters might be related to previously documented subtelomeric silencing functions of the COMPASS complex [41] and mechanistically similar to the recently identified heterochromatin-promoting role in fission yeast [49] . Dynamic demethylation of lysine residues adds additional complexity to the modulation of transcription by lysine methylation [3 , 52] . Recently we showed that KdmA , a JMJD2/JHDM3 family H3K9/36me3 demethylase [53 , 54] can , in equal measure , positively and negatively influence gene expression in A . nidulans [18] . Here , we characterize another member of the JmjC demethylase family , KdmB , which acts on H3K4me3 in vivo , thus is assigned to the Jarid group of enzymes . Jarid ( JMJ–AT-rich interacting domain-containing protein ) subfamily demethylases have been shown to target di- and trimethylated H3K4 and are therefore generally considered to be repressors of gene transcription , though they can also act as activators [55] . For example the function of mammalian RBP2 ( retinoblastoma binding protein 2 , alias JARID 1A or KDM5A according to the new nomenclature [56] ) in transcription regulation is context dependent . RBP2 represses transcription via H3K4me3 demethylation and association with an HDAC complex , however when associated with retinoblastoma protein ( pRb ) , it activates certain genes in the mammalian genome [57] . Similarly , the D . melanogaster ortholog LID can repress transcription via H3K4me3 demethylation , however when associated with the MYC transcription factor , its demethylase activity is inhibited and consequently the LID-MYC complex mediates gene activation [58 , 59] . These examples demonstrate that Jarid demethylases can act directly on their target genes in a context dependent positive or negative manner . In this work we studied the Jarid-type demethylase in A . nidulans by reverse genetics and performed genome-wide HTPM profiling by mass spectrometry of histones , by ChIP analysis of H3K4me3 , H3K9me3 , H3K36me3 and H3 acetylation on K9 and K14 ( H3Ac ) modifications in wild type and compared the results with the KdmB mutant . We recorded these HPTM changes in parallel with the transcriptome under optimal physiological conditions promoting active growth ( primary metabolism ) as well as under stationary-phase conditions that lead to SM production ( secondary metabolism ) . Comparison of ChIP-seq profiles with RNA-seq of the same cultures allowed us to correlate transcriptional changes with changes in chromatin landscapes across different conditions and genetic backgrounds . Histone proteomic analysis in wild type and the KdmB histone H3K4 demethylase mutant provided direct evidence for H3K4me3 as the dominant substrate for KdmB and confirmed that A . nidulans does not feature H3K27me3 , the canonical facultative heterochromatic mark in other eukaryotes and responsible for SM gene silencing in a number of other fungi .
Based on the domain composition of the full length KdmB ( AN8211 ) and detailed analysis of the amino acid sequences of the catalytic JmjC domains of histone demethylases from yeast to humans , KdmB was classified as a Jarid1-type histone H3 lysine 4 demethylase ( Fig 1 ) . Residues responsible for substrate recognition of Jarid demethylases are not known due to the lack of available crystallographic data , although the conserved amino acids required for substrate recognition in the JMJD2 subfamily of lysine K9 and K36 histone H3 demethylases ( marked in green in Fig 1A ) are not present in the Jarid group [60] . Domain analysis revealed that KdmB is more similar to the proteins from higher eukaryotes than from budding yeast . Specifically , we found that KdmB contains a putative ARID/Bright domain and a C5-HC2 zinc finger motif and an additional PHD domain at the C-terminus , which are both absent from the budding yeast homolog ( Fig 1B ) . To investigate the in vitro specificity of KdmB we heterologously expressed KdmB as a GST fusion protein in E . coli . KdmB has predicted molecular weight of 216 kDa but the resulting full size recombinant protein was not sufficiently soluble . Another construct producing a truncated KdmB protein without the second PHD domain , however , was readily soluble under native buffer conditions . This KdmB fusion containing residues 1 to 922 displayed an apparent mass of roughly 130 kDa ( S1A Fig ) . In vitro demethylase assays ( DeMt ) were subsequently performed with purified GST-KdmB ( 1–922 ) and calf thymus histones as a substrate . Products of the DeMt reactions were detected with modification-specific antibodies by Western blot ( S1B Fig ) . Under our assay conditions , we found a decrease in trimethylation signals for all three tested lysine residues ( H3K4me3 , H3K9me3 and H3K36me3 ) and the strongest reduction in abundance was seen in H3K9me3 . Acetylation was not reduced by the enzyme , as expected . Consistent with KdmB being a JmjC-type demethylase , the activity of the GST-KdmB ( 1–922 ) , fusion protein was dependent on the presence of the cofactors α- ketoglutarate and Fe2+ ( S1B and S1C Fig ) . In our assay conditions we observed high standard deviations between independent replicates of H3K4me3 and H3K36me3-specific Westerns . This could be due to experimental variation in enzymatic activity of different batches of the purified recombinant enzyme . The very broad substrate range of KdmB in vitro is unexpected for this Kdm5-family member because so far the identified and tested enzymes target either H3K4me2/3 ( Jarid1 group enzymes ) or H3K9/36me2/3 ( Jmjd2 group ) . However , it is possible that the absence of PHD-finger 2 , interacting proteins or the presence of the GST domain compromises substrate specificity in our assay . Although none of the KdmB orthologs identified so far demonstrated such broad substrate specificity in vitro [57 , 63–66] , the in vitro demethylase activity found in our assays suggests that this protein possesses histone demethylase activity . To determine whether KdmB can act as a histone demethylase in vivo , we performed LC- MS/MS on acidic extracted histones from actively growing A . nidulans wildtype and kdmBΔ cells ( see Materials and Methods for description of gene deletion procedure ) . In wildtype , mass spectrometry revealed that 71 . 3% of H3K4 peptides contain at least one methyl-group at the K4 position . We detected all three forms of methyl-H3K4 peptides and found that H3K4me3 is the most abundant ( 47 . 5% of total H3K4 peptides ) , followed by di-methylated ( 13 . 5% ) and mono-methylated H3K4 ( 10 . 3% ) ( Fig 2 ) . Notably , our measurements revealed an almost 20% increase in global H3K4 trimethylation in the kdmBΔ strain ( 57% H3K4me3 ) . Because the levels of H3K4me2 , H3K4me1 and unmodified H3K4 were concomitantly decreased in the mutant in roughly the same range as H3K4me3 increased we concluded that in vivo KdmB primarily acts to demethylate H3K4me3 . The MS results also revealed that in vivo KdmB does not target H3K36me3 as these levels remained constant in histones of kdmBΔ cells ( S2 Fig ) . Interestingly , the overall low marking of H3K9 by trimethylation ( 1 . 53% of the mapped peptides ) was further reduced ( to 0 . 2% of the mapped peptides ) in the mutant . This contrasts the in vitro assay results which showed a strong H3K9me3 demethylating activity of recombinant KdmB ( S1 Fig ) . The further reduction of H3K9me3 in kdmBΔ cells might be attributable , however , to an increase in the opposing , positively acting H3K4me3 mark limiting the possibility to deposit or maintain H3K9 trimethyl marks in the target regions . Strikingly , in vivo , global H3 N-terminal lysine acetylation ( H3K9ac/K14ac ) was increased almost by 20% in the kdmBΔ strain at the expense of unmodified peptides of H3 which are reduced from 22% in the wild type to 6% in the mutant ( S2A Fig ) . This more abundant histone acetylation could be the consequence of both stronger marking by acetylases and/or reduced deacetylation . The latter mechanism has already been reported in connection with KdmB homologs in mammals where RBP2 ( Jarid1a ) and PLU1 ( Jarid1b ) recruit the Rpd3S histone deacetylase complex [57 , 67] . Altogether , our data demonstrate H3K4me3 demethylation activity of KdmB in A . nidulans cells and lack of this activity in kdmB deletion cells leads to a shift in modification equilibrium with more abundant positive ( H3K4me3 , H3Ac ) and less negative ( H3K9me3 ) marks . To determine the genomic regions in which KdmB influences H3K4me3 levels we performed genome-wide ChIP analysis ( ChIP-seq ) in wild type and kdmBΔ strains with antibodies specific to H3K4me3 [28] . As our global histone analysis revealed a crosstalk of this modification to H3K9 trimethylation as well as to H3K9/K14 acetylation , we also included these marks in ChIP-seq . Although no changes occurred for H3K36 trimethylation at the level of bulk histones between WT and the kdmB mutant , we were interested if locus-specific differences occur and thus analyzed also this mark by ChIP-seq . As previous studies from our lab and by others revealed a crucial function of chromatin structure and histone modifications on the regulation of secondary metabolite biosynthesis ( SMB ) , we performed all subsequent RNA-seq and ChIP-seq experiments not only under the already described standard active growth conditions representing primary metabolism ( PM; 17h liquid shake cultures , no nutrient limitation ) but also under conditions promoting secondary metabolism ( 48h liquid shake cultures , nutrient depletion , see S3 Fig ) . To monitor the distribution of the tested chromatin modifications along A . nidulans genes , we used chromosome IV as an example and plotted the wild type distribution of H3K4me3 , H3K36me3 , and H3K9/14ac across the promoters and open reading frames ( ORFs ) of all genes on this chromosome ( Fig 3A ) . In this analysis , all genes are aligned to the predicted ATG ( position 0 ) and read counts per million of mapped reads ( CPM ) are analysed in a 2 kb window starting with 500 bp of their 5´UTR and promoter sequences ( -500 ) followed by 1500 bp of their coding region . This revealed that the pattern of modifications reflects the distribution observed in other model organisms including fungi [30 , 68–70] . H3K4me3 was enriched in characteristic peaks spanning the first three nucleosomes ( around 500 bp ) of the coding region , whereas H3K36me3 was enriched near the 3’ regions of genes . Finally , H3 acetylation was enriched in the promoter , with highest levels apparent in the first nucleosome just downstream of the predicted translation start sites . To explore the general relationship between H3K4me3 and transcription we quantified the average level of H3K4me3 in a 2 kb window around the predicted start codon of each gene ( average CPM from -500 to +1500 ) and related this value to the average expression level ( expressed as RPKM , reads per kilobase per million reads ) of the corresponding gene in both culture conditions ( PM and SM ) . In the resulting scatterplot ( Fig 3B ) two groups of genes became apparent , i . e . those that displayed high levels of H3K4me3 ( log2 RPKM>5 ) and a second group that showed low to no H3K4 trimethylation ( log2 RPKM≤5 ) . Correlation of H3K4me3 levels with transcription of the corresponding gene revealed an overall positive correlation between H3K4me3 levels and transcript abundance ( Fig 3B ) . This suggests that , similar to other well-studied models , H3K4 trimethylation is a marker for actively transcribed genes . To better characterize the function of KdmB in the context of transcriptional regulation we next compared by ChIP-seq the distributions of four histone modifications in wild type and kdmBΔ ( Fig 4 ) under active growth conditions ( PM ) and during SM . The kdmB deletion did not cause any gross phenotypic changes in the mutant strain which was rather similar to the wild type in growth rates and nutrient consumption ( S3 Fig ) . ChIP-seq combined with RNAseq analysis revealed the H3K4me3 enriched domains which coincide with transcriptional activity . In the example shown in Fig 4 we noticed , on the gross genomic scale , an overlap between the positively acting marks H3K4me3 , H3K36me3 and H3Ac . In contrast , repressing H3K9me3 marks are enriched mainly in pericentromeric and subtelomeric regions and a few isolated H3K9me3 blocks exist ( on the left arm of chromosome IV , for example ) . At the gross genomic scale the comparison of the chromatin landscape for H3K4me3 marks in chromosome IV between actively growing ( 17 h cultures ) wild type and kdmBΔ cells did not reveal any obvious changes . Moreover , at this scale , no large domains were visibly changed for the other tested modifications ( H3Ac , H3K36me3 , H3K9me3 ) . Because our mass spectrometry analyses uncovered increased H3K4me3 and H3Ac in the mutant , we reasoned that changes in the levels of these histone marks must occur at a subset of individual genes . To test this , we analyzed H3K4me3 levels in genes that were differentially expressed between wildtype and kdmBΔ . We first examined genes with low H3K4me3 levels [ ( log2 ( RPKM ) ≤ 5] and found that 301 genes displayed higher expression levels in the wildtype ( WT-up/Group 1 , Fig 5A ) suggesting that for this group KdmB is required for normal expression levels . In contrast , 501 genes had higher expression in kdmBΔ ( kdmBΔ-up/ Group 2 ) which points to a repressing function of the protein in these loci . In the gene set featuring high H3K4me3 levels [ ( log2 ( RPKM ) > 5] we again identified both up- and down-regulated genes; 455 genes were expressed at higher levels in wild type ( WT-up/Group 3 ) and 133 genes were expressed at higher levels in kdmBΔ ( kdmBΔ-up/ Group 4 ) . The analysis showed that KdmB influences transcriptomes in both directions . For around 750 genes KdmB function is necessary for normal transcription , whereas for around 630 genes KdmB has a negative function . The repressive role of KdmB was found in both categories , i . e . on genes carrying low ( kdmBΔ-up/G2 ) or high ( kdmBΔ-up/G 4 ) H3K4me3 levels . Significantly , the group with normally low H3K4me3 ( G2 ) displayed a marked increase in this histone mark in the kdmBΔ mutant concomitantly with increased transcript levels . One representative of this group is shown in Fig 5C for a gene ( locus AN6321 ) which is basically not transcribed in the wild type but which gains both positive marks and transcripts in the kdmBΔ strain . Although we have not tested this directly , the strict correlation between increased H3K4me3 levels and transcription , along with the in vitro K4me3-demethylase activity of KdmB , suggests that at least some of these loci are direct targets of KdmB . A slightly different situation was found for the second gene set highly decorated with H3K4me3 . Although a subset of these genes showed increased expression in the kdmBΔ mutant ( kdmBΔ-up/G 4 ) , this was not accompanied by an increase in H3K4me3 probably due to the already very high K4 methylation levels in the wild type . Consequently , a further increase would hardly be possible and thus the effect of kdmB deletion on H3K4 trimethylation is more subtle compared to genes generally not heavily marked by H3K4me3 . In contrast to the repressive function , KdmB also seems to have a positive role in transcription . kdmB deletion led to reduced expression of 750 genes belonging to both low ( WT-up/G1 ) or high ( WT-up/G3 ) H3K4me3 groups , accompanied by lower H3K4me3 , on average , in the mutant . Based on these correlations we can conclude that KdmB function is required for normal expression of these roughly 750 genes , but whether KdmB directly targets these loci or indirectly affects transcription via the transcriptome network remains to be determined . We also constructed metaplots of H3K4me3 distributions under SM conditions ( S4 Fig ) . Under these growth conditions a similar correlation was observed , i . e . H3K4me3 levels were reduced in genes that were downregulated in kdmBΔ , whereas the genes upregulated in the mutant showed no drastic change ( in the high H3K4me3 group ) or somewhat higher H3K4 trimethylation . However , in locus-specific analysis by RNA-seq and ChIP-seq ( see below ) , we also found some transcriptionally silent regions with high H3K4me3 as well as some highly transcribed genes with very low levels of this mark ( see analysis below ) indicating that specific genomic regions exist in which this general positive correlation between H3K4me3 and transcriptional activity does not apply . Our initial correlation analysis of H3K4me3 and transcription revealed that among genes requiring KdmB for full transcription , the category of SMB genes was significantly enriched ( p < 0 . 05 ) . In further analysis , PM and SMB genes were separated based on functional categories and this bioinformatic approach created a large group of genes ( 5676 genes ) predicted to be involved in general cellular functions and metabolism ( category “cell structure and function” abbreviated CSF ) and a smaller group of 149 genes predicted to be involved in SMB ( category “SM clusters” ) . [71 , 72] . Fig 6 shows that under PM conditions , approximately 5% of genes involved in CSF and 15% of genes assigned to SMB were affected by the kdmB deletion . The majority of A . nidulans SM cluster genes are not under/ during PM conditions , thus it is not surprising that differential expression of SM genes is largely restricted to the 48h cultures . Interestingly , several genes belonging to a gene cluster with a so far unidentified product were highly upregulated in the mutant at this 17h time point and this transcriptional pattern will certainly facilitate the future identification of the product derived from this predicted SM cluster . In contrast to the mild effect on SM gene expression during PM conditions , KdmB-deficient cells showed significantly altered patterns of gene expression when cells were collected from cultures under SM conditions . Over 50% of all predicted SM genes were misregulated in the mutant . The majority of these displayed lower expression , while approximately 10% of SM genes showed higher expression in the kdmBΔ strain ( Fig 6A , upper panel ) . In contrast , during the same culture condition only ~10% of genes not involved in SM were differentially transcribed in kdmBΔ . These data demonstrate that KdmB is required for normal induction of the majority of SM clusters in A . nidulans . It is probably relevant to note that the defect in SM cluster activation in the kdmB mutant is not due to a lack of wide-domain activator expression as laeA , veA , velB and velC are normally transcribed in the mutant ( changes between WT and kdmBΔ log2 ≤ ± 1 , 7 ) . The lower panel of Fig 6A presents the number of deregulated genes within each category and time point . During primary metabolism ( 17h ) KdmB function is required for a relatively small number of genes ( 143 genes in CSF and 10 genes in SM ) . In contrast , in the nutrient limited 48h cultures gene expression profiles are changed considerably in the mutant: 598 genes ( 401 CSF and 97 SMB genes ) require KdmB function for normal expression and 569 genes ( 547 CSF and 22 SMB genes ) are negatively influenced by the regulator . These data suggest that KdmB is primarily required during the stationary phase and obviously plays an important role for the expression of the majority ( 97 out of 149 of genes involved in SMB We also tested whether transcriptional changes in kdmBΔ were correlated with changes in SMB biosynthesis . For this we performed HPLC-MS/MS analyses of cultures grown in two different media , i . e . in conventional minimal medium used throughout the studies ( AMM ) and in a specialized SM-promoting ZM medium ( see Materials and Methods section ) . The comparison of WT and mutant culture extracts , grown in AMM medium , revealed a strongly decreased production of sterigmatocystin and emericellamides C and D ( Fig 6B , left chromatograms ) but other metabolites such as emodin and its derivatives were increased in kdmBΔ ( Fig 6B , chromatograms a and c ) . However , our RNA-seq data showed that genes encoding for enzymes involved in emodin biosynthesis embedded in the mdpL-A monodictyphenon pathway are not differentially expressed between WT and the kdmB mutant ( S13 Fig ) . To accommodate these differences , we speculate that the decreased transcription of other secondary metabolite clusters , such as the sterigmatocystin cluster , may lead to higher levels of available emodin precursors , such as acetyl-CoA and malonyl-CoA , and thereby to an increased synthesis of emodin derivatives . ZM culture extracts revealed reduced levels of orsellinic acid in kdmBΔ ( Fig 6B , right chromatograms ) , consistent with our RNA-seq data showing a decreased expression from the orsellinic acid gene cluster in the kdmB deletion ( S10 Fig ) . The complete list of identified metabolites together with LC-MS and LC-MS2 data are shown in the S3 Table . We also carried out correlation analyses between H3 acetylation and H3K4 methylation in genes which are differentially regulated in the kdmB mutant ( S5 Fig ) . For those genes where KdmB is required for full expression and which are consequently higher transcribed in the wild type ( categories WT-up/G1 and G3 ) H3 acetylation levels are also higher , independently of H3K4 trimethylation . The same is true for genes which are negatively influenced by KdmB ( kdmBΔ-up/G2 ) but only if H3K4me3 levels are low . On the contrary , genes with high H3K4me3 levels under negative KdmB influence ( kdmBΔ-up/G4 ) , acetylation levels are lower than in the wild type despite higher expression of the corresponding genes in this group . The molecular basis of this effect has not been investigated further in this study but it would certainly be interesting to determine if KdmB impacts acetylation indirectly or directly through protein interactions with HDACs or HATs . We also examined a possible influence of KdmB on the distribution of H3K36me3 in genes expressed under primary metabolic conditions ( S6 Fig ) . We have previously shown that this mark is associated with active transcription and that , at some tested loci , the trimethylated H3K36 state is removed by KdmA , another A . nidulans JmjC-containing protein belonging to the KDM4 family [18] . The vast majority of A . nidulans genes are highly decorated by this mark under PM conditions ( around 9 , 100 genes ) . We did not observe significant differences in the levels or in the distribution of this mark in the kdmBΔ strain neither in this group nor in the group carrying low H3K36me3 levels ( 1249 genes . This indicates that KdmB is not a demethylase of trimethyl-H3K36 in vivo . Around 13% of the 9 , 100 genes are de-regulated in the kdmB mutant strain but despite this differential expression there are no significant differences in the associated H3K36 trimethylation levels . This means that , at least for the gene set in which KdmB influences transcription , it does not do this via manipulating H3K36me3 levels . The genome-wide distribution pattern of H3K9me3 supports the previously reported low levels of H3K9 trimethylation in A . nidulans wild type cells where we found approximately 1 . 5% of peptides carrying this mark . [18] . In ChIP-seq , the H3K9me3 pattern correlates with AT-rich domains flanking the subtelomeric regions but also includes sites along the chromosome arms , as shown on the left arm of chromosome IV ( Fig 4 ) . Inspection of H3K9me3-associated regions revealed that many SMB gene clusters such as the penicillin ( S7 Fig ) , sterigmatocystin ( S8 Fig ) , austinol ( S9 Fig ) , orsellinic acid ( S10 Fig ) and terrequinone A ( S11 Fig ) are flanked by H3K9me3 domains at either one ( e . g . the ST and TDI clusters ) or at both sides ( e . g . the PEN cluster ) of the cluster . Whether these structures are functionally relevant for the regulation of SM gene clusters remains obscure but possible since deletion of the H3K9 methyltransferase gene clrD or of hepA , the gene coding for the protein recognizing H3K9me3 , lead to up-regulation of genes within these clusters [27] . The observation that many H3K9me3 blocks are found in close proximity to SMB gene clusters raises the possibility that higher order chromatin structures or a as yet unstudied set of modifications may be important for normal regulation of SM gene expression , consistent with prior genetic analyses [27 , 29 , 31 , 32] . However , we have also found several SM clusters such as the asperthecin ( S12 Fig ) and monodictyphenone ( S13 Fig ) cluster without such H3K9me3 borders . Interestingly , these clusters are not activated under the standard SM growth conditions used here ( 48 h cultures and nutrient deprivation ) . Instead , the MDP cluster is only expressed to detectable levels in a strain lacking the CclA regulatory subunit of the COMPASS complex which is responsible for H3K4 di- and tri methylation [28] and APT is highly expressed only in an A . nidulans mutant lacking SUMO , the small ubiquitin-related modifier protein known to profoundly regulate chromatin structure and function [73 , 74] . Hence , absence of the H3K9me3 blocks might be correlated with special requirements for activation whereas SMB gene clusters activated under standard SMB conditions feature H3K9me3-flanking domains . Correlation of H3K4me3 with transcriptional activity suggested that SMB gene clusters carry low levels of this mark even when they are strongly transcribed ( see Fig 3 ) . Inspection of ChIP-seq data from these regions confirmed that H3K4me3 is underrepresented in such clusters , as shown in the example of the well-studied sterigmatocystin gene cluster ( Fig 7 ) . When in conditions of primary metabolism , cluster genes are silent and are not associated with H3K4me3 but surprisingly , this mark is not established at most genes even when the cluster is fully activated ( Fig 7 ) . Eventually , a single strong H3K4me3 peak occurred around the 5´end of stcD , a gene coding for an unknown function but co-regulated with the sterigmatocystin biosynthesis cluster [75] . Qualitatively , the two other tested activating marks H3K9/K14 acetylation and H3K36 methylation seem to increase around 5´ and 3´ends of the ORFs , respectively , in the activated cluster . A very similar picture emerged from the analysis of other clusters ( S9–S13 Figs ) and in each case , as expected , no major differences in the H3K4me3 profiles became apparent between the kdmBΔ mutant and the wild type . To test our qualitative impression for significance we performed statistical analysis of our ChIP-seq data for differences in H3K4me3 , H3Ac and H3K36me3 marks in PM and SMB conditions in the wild type and in the kdmBΔ mutant . The bioinformatic separation into “Cell structure and function” and “SM clusters” categories applied for the transcriptome was also kept for ChIP-seq data analysis . The statistical analysis of ChIP data revealed a striking difference in H3K4me3 levels between the two categories . As seen in the box blot in Fig 8 genes involved in SM production are significantly less decorated by H3K4me3 , regardless of the culture condition or the presence of KdmB . Moreover , the pattern is not significantly changed in the kdmBΔ strain , suggesting that KdmB does not promote SM gene expression by directly regulating H3K4me3 within SM clusters . SM cluster activation leads to subtle increases in the level of H3K4me3 , H3Ac or H3K36me3 associated with SM cluster genes , and this increase is not visible in kdmBΔ ( Fig 5A ) . In summary , our ChIP-seq data revealed that A . nidulans SM clusters in comparison to genes involved in the cell structure and function have relatively low levels of activating histone marks , especially H3K4me3 and H3K36me3 .
Di- and tri-methylation of histone H3K4 is associated with transcriptionally active chromatin . Removal of this modification is accomplished by members of the KDM5-family demethylases , typically resulting in repression of the targeted locus . In fact , the first characterized H3K4 demethylases LID2 [59] and RBP2 [57] were identified as transcriptional repressors . However , these proteins and all KDM5 members are composed of multiple domains which are necessary for the diverse functions these regulators play . For example , demethylase activity of KDM5 is only one of the important functions required for Drosophila development [76] [77] . In addition , some domains have been associated with gene activation , for example , mammalian Jarid1a is recruited to the Per2 circadian gene promoter where it inhibits HDACs function and promotes transcription [78] . We also found in our study that deletion of KdmB has both activating and repressing effects . We found a genome-wide 20% increase in acetylated H3 N-termini and increased transcription of around 630 genes under standard growth conditions ( nutrient sufficiency , primary metabolism ) in strains lacking KdmB . These results provide evidence that the protein functions as a repressor that is able to remove H3K4me3 and perhaps recruit HDACs . Our data also demonstrate that KdmB is an H3K4me3 demethylase . The protein removed this modification in vitro and genes that are overexpressed in a KdmB-deficient mutant show increased H3K4me3 . Unfortunately , from our data , we cannot deduce which part of this gene set is directly targeted by KdmB and which may be indirectly silenced through transcriptome network effects . ChIP analysis of KdmB tagged versions will be able to clarify this point in future . On the other hand , for around 750 A . nidulans genes KdmB is required for full transcription . It is possible that KdmB mediates activation directly via one or more domains such as the potentially DNA-binding Zn-finger or ARID domains or the methylated histone binding PHD domain . However , deciphering how KdmB promotes transcription requires further investigation . Strikingly , the majority of the genes under positive KdmB control are related to the production of secondary metabolites . These small natural products are defense and signaling molecules of fungi produced during development , under stressful or nutrient-limiting conditions [23] and it is interesting that a general chromatin regulator such as KdmB takes up this specialized function in metabolism . We have shown that KdmB regulates ( directly or indirectly ) almost 5% of the genome during PM and over 10% during SM . The activation signal for the induction of genes involved in SMs production is transmitted via the so-called velvet activation complex containing also a protein termed LaeA that influences chromatin structure [27] . It will be interesting to determine if KdmB functions in this pathway . It is possible that KdmB could regulate SM gene expression by demethylation of SM regulatory proteins . Recently it was shown that several JmjC family demethylases can target non-histone substrates; however this function , to our best knowledge , has not been demonstrated for Jarid family demethylases [77 , 79] . This role of KdmB in SMB gene activation appears to be independent of its histone demethylase enzymatic activity . In kdmBΔ , low levels of H3K4me3 and H3 lysine acetylation in SM gene clusters under activating conditions are likely the consequence of lower transcription at these loci . One of the most striking features of silent A . nidulans SM clusters is a very low abundance or virtual absence of the four investigated histone marks within the borders of these gene clusters . At the moment we cannot exclude the possibility that other histone marks define the chromatin landscape within and around SM gene clusters . A large number of SM clusters , as exemplified here for sterigmatocystin ( ST ) , penicillin ( PEN ) , orsellinic acid ( ORS ) , teraquinone ( TDI ) , derivative of benzaldehyde 1 ( DBA ) , austinol ( AUS ) and asperthecin ( APT ) are located in regions for which H3K4me3 , H3K36m3 or H3Ac can hardly be detected . Even the monodictiphenone ( MDP , S13 Fig ) and asparthecin clusters ( APT , S12 Fig ) , which are located within euchromatic regions , display a low-abundance of HPTMs . A distinguishing feature of these two clusters , which are not activated by the conventional SMB culture conditions applied here , is the lack of flanking by H3K9me3 domains which are characteristic for the majority of the analyzed SMB gene clusters ( S7–S13 Figs ) . Although truncated KdmB demethylates H3K9me3 in vitro ( S1 Fig ) we did not see increased levels of this mark in the kdmBΔ mutant neither at specific loci nor at the genomic scale . This strongly suggests that in vivo H3K9me3 is not a target of KdmB . Moreover , no KDM5-type H3K9 demethylases have been described in other ascomycete fungi including S . pombe or N . crassa . In A . nidulans we even see genome-wide reduced levels of this mark in the mutant and it is likely that this reduced H3K9me3 is an indirect consequence of increased H3K4me3 or increased H3 acetylation . In addition to H3K9me3 , the chromatin landscape changes slightly also for the other tested marks when the silent SM gene clusters are activated . The majority of the genes in these clusters gain H3K36me3 at their 3’ region and H3Ac at their 5’ region . Marking by H3K4me3 , however , only occurs for a limited number of genes within these clusters such as some selected genes within the ST cluster ( Fig 8 ) or the orsD gene positioned within the ORS cluster ( S10 Fig ) . A similar situation was recently reported in two different Fusarium species in which the H3K4 dimethylation level ( H3K4me2 ) was compared to SMB gene transcription . In the rice pathogen F . fujikuroi , only two out of seven highly transcribed genes in the gibberellin cluster were significantly decorated with H3K4me2 [38] and also in F . graminearum , a pathogen of wheat and maize , genes in the fusarin C or the carotenoid biosynthesis clusters carried only background levels of this mark [30] . It is still remarkable that Liu and colleagues found an essential function of the H3K4 methyltransferase Set1 for the expression of the TRI gene cluster coding for deoxynivalenol biosynthesis [80] and in this latter study , H3K4me2 was clearly enriched over the background level and positively correlated with active transcription . Additionally , in contrast to our study , Connolly et al . [30] found H3K36me3 enrichment across the whole chromosome independent of transcriptional activity . These comparisons already highlight the high diversity of chromatin-based regulation in SMB gene expression within one single organism and even more between different organisms and which histone modifications are determining whether the SMB signal is transmitted to the transcriptional machinery or not . Surprisingly , the H3K4 demethylase KdmB plays an essential role in the activation process although this histone mark is not present in the targeted regions .
A . nidulans strains used in this study are listed in S1 Table . Experimental strains were obtained by transformation into an nkuAΔ strain , which reduces the frequency of non-homologous integration [81] , or by sexual crosses . Genetic analysis was carried out using techniques as described by Todd et al . [82] . DNA transformation of A . nidulans was performed according to [83] . KdmB deletion cassettes were constructed using DJ PCR [84] with the Aspergillus fumigatus riboB gene as selectable marker , riboB+ transformants were recovered after transformation into nkuA strains . Southern analysis confirming the deletion of kdmB was performed as described elsewhere [18 , 40] . AMM minimal media , complete medium , supplements and growth conditions were as described by Todd et al . [82] . ZM1/2 medium ( molasses 0 . 5% , oatmeal 0 . 5% , sucrose 0 . 4% , mannite 0 . 4% , D-glucose 0 . 15% , CaCO3 0 . 15% , edamine 0 . 05% , ( NH4 ) 2SO4 0 . 05% ) was used for promoting SM biosynthesis in the experiments analyzing metabolites by HPLC-MS/MS [85] . For LC-MS/MS , RNA-seq , ChIP-seq , DeMt assay , HPLC-MS/MS spores in concentration 4*106/mL were inoculated into 200 mL liquid AMM and incubated at 180 rpm , 37°C for 17 h and 48 h . For SM cluster gene expression , ChIP and HPLC- MS/ MS analysis 10 mM sodium nitrate otherwise ammonium tartrate at 10 mM was added as nitrogen source . kdmB cDNA was amplified using RevertAid Premium Reverse Transcriptase ( Thermo Scientific , EP0732 ) and specific primers . Full length ( 1717 aa ) and truncated versions ( residues 1–922 ) cDNAs were cloned into pGEX-4T1 expression vector , sequenced , transformed and expressed in Rosetta cells . GST-KdmB ( 1–922 ) was purified using glutathione Sepharose 4B ( GE Healthcare ) . Demethylation assay was performed as previously described [66] . Purified KdmB was incubated with calf thymus histones ( Sigma , H9250 ) in demethylase reaction buffer ( 20mM Tris-HCl pH 7 . 2 , 150 mM KCl , FeSO4 20 μM , α-ketoglutarate 500 μM , ascorbic acid 500μM , ZnCl2 1μM ) for 3 to 10 h at 37° . Reaction was stopped by boiling for 5 minutes with 100 mM DTT Laemmli buffer; changes in lysine methylation were measured by Western blot with the specific antibodies ( see ChIP section ) . The demethylation in vitro assay and Western blot were performed three times; negative controls were incubated without the cofactors for JmjC proteins ( Fe2+ and α-ketoglutarate ) . Mycelia from o/n liquid submerged cultures were harvested by filtration and frozen in liquid nitrogen . Histones were acid extracted as previously described [86] , suspended in Laemmli’s SDS sample buffer and quantified with Pierce BCA Protein Assay ( Thermo ) . 15 μg of purified histones , 1 μg of recombinant Xenopus laevis H3 as a negative control ( Milipore , 14–441 ) and 2 μg calf thymus histones ( Sigma , H9250 ) as a positive control were separated on 15% SDS-PAGE gel and subsequently transferred to nitrocellulose membrane ( GE Healthcare ) by electroblotting . Relevant histone modifications were detected with primary antibodies specific to H3K4me3 ( Abcam , 8580 ) , H3K9me3 ( Active Motif , 39161 ) , H3K36me3 ( Abcam , 9050 ) , histone H3 C-terminus ( Abcam , 1791 ) , H3Ac ( pan-acetyl ) ( Millipore , 06–599 ) and anti- rabbit ( Sigma , A0545 ) and anti- mouse ( Sigma , A9044 ) HRP conjugated secondary antibodies . Chemiluminescence was detected with Clarity ECL Western Substrate and ChemiDoc XRS ( Bio- Rad ) . Densitometric quantification of Western blot signals from demethylase reactions were performed with the ImagJ software . In total three independent blots of demethylase and the control reaction ( without the cofactors ) were quantified . Signal of respective HPTM were normalize to histone H3 C-term . Subsequently the signal of the control reaction was set to a value 1 , consequently the presented results are the fold change to the control reaction . For MS analysis relevant histone H3 protein bands were cut out and digested in gel . The proteins were S-alkylated with iodoacetamide and digested with ArgC ( Roche ) . The peptide mixture was analysed using a Dionex Ultimate 3000 system directly linked to a Q-TOF MS ( Bruker maXis 4G ETD ) equipped with the standard ESI source in the positive ion , DDA mode ( = switching to MSMS mode for eluting peaks ) . MS-scans were recorded ( range: 150–2200 Da ) and the 6 highest peaks were selected for fragmentation . Instrument calibration was performed using ESI calibration mixture ( Agilent ) . For separation of the peptides a Thermo BioBasic C18 separation column ( 5 μm particle size , 150*0 . 360 mm ) was used . A gradient from 95% solvent A and 5% solvent B ( Solvent A: 0 . 1% FA in water , 0 . 1% FA in ACCN ) to 32% B in 45 min was applied , followed by a 15min gradient from 32% B to 75% B that facilitates elution of large peptides , at a flow rate of 6 μL/min . The fungal cultures were incubated in triplicates , RNA from each technical replicate was pooled and each experiment was performed twice to obtain two biologically independent sets with two technical replicates for each strain and condition . Illumina sequencing libraries were made from RNA samples according to TruSeq RNA Sample prep kit v2 ( Illumina ) following the manufacturers protocol with 1μg total RNA input . 50 bp single end sequencing was performed using a HiSeq Illumina sequencer . Obtained sequences were de-multiplexed , quality controlled and mapped on the Aspergillus nidulans genome assembly ( A_nidulans_FGSC_A4_version_s10-m03-r07 ) . Mapping was performed using Novoalign ( NovoCraft ) and reverse transcripts were counted using python script HTSeq [87] . Normalization and statistics were done using R/Bioconductor and the limma and edgeR packages , using mean-variance weighting ( voom ) and TMM normalisation [88] . A significance cut-off of p < 0 . 01 ( adjusted for multiple testing by the false discovery rate method ) was applied for analysis . R plots used the ggplot2 package [89] . Transcription levels are log2 read counts per kilobase of exon per million library reads ( RPKM ) . For trace graphs as shown in S7–S13 Figs transcript coverage was calculated as explained for the ChIP-seq experiments to obtain counts per million reads ( CPM ) . SM clusters are annotated as described by Inglis and coworkers [72] . All data are available at NCBI GEO under the accession number GSE72126 . Chromatin immunoprecipitation was performed as described in [18] Chromatin was incubated with antibodies specific to H3K4me3 ( Abcam , 8580 ) , H3K9me3 ( Active Motif , 39161 ) , H3K36me3 ( Abcam , 9050 ) , H3Ac ( Millipore , 06–599 ) or Histone H3 C-terminus ( Abcam , 1791 ) and Dynabeads Protein A ( Invitrogen ) . Precipitated DNA from two biological and two technical replicates was quantified by real-time PCR according to protocol ( Bio-Rad ) using iQ SYBR Green Supermix and normalized to input DNA or sequenced . Primers used in quantitative PCR were HPLC purified and are shown in S2 Table . For Illumina sequencing , ChIP-seq libraries were prepared using 10 ng of immunoprecipitated DNA following the instructions supplied with Illumina Tru-seq ChIP-seq kits ( Illumina Cat# FC-121-2002 ) . Illumina sequencing was performed using an Illumina NextSeq500 Instrument at the University of Georgia Genomics Facility . Short reads were mapped using Novoalign ( NovoCraft ) to the current Aspergillus genome annotation , obtained from the Aspergillus Genome Database [90] . Read numbers were counted for 10 base pair bins using sam tools and R , and the read density was normalized for total read number and visualized using the Integrated Genome Viewer or Integrated Genome Browser [91–93] . In detail: Metaplots ( Figs 3A , 5A , 5B and S4–S6 ) were calculated from bam files using bedtools genomecov [94] . The sequencing coverage per base pair ( bp ) was calculated for the whole genome , normalized using a scaling factor ( 1000000/total mapped sequence read counts ) that accounts for the different counts of mapped reads per sample to obtain counts per million mapped reads ( CPM ) to allow comparison between samples represented as trace files in sgr format . Using R scripts the sgr files were smoothed by averaging a window of 100 bp length that was slided by 10bp , thereby reducing computation demand ( 10bp binning ) . Gene start/stop codon position , length and strand were retrieved from gff file provided by Aspergillus Genome Database [90] and 2kb of each gene ( 500bp upstream + 1500bp downstream of ATG in case of H3Ac and H3K4me3 or 1500bp upstream + 500bp downstream of stop codon for H3K36me ) taken and averaged for the specified group ( e . g . transcription strength ) . Detailed R scripts can be obtained from the authors . Per gene levels of ChIP-seq data were calculated as transcript sequences counting reads per gene and were normalized to the ORF length ( in comparison to the exon length ) to obtain reads per million reads per kb ORF length ( RPKM ) . All data are available at NCBI GEO under the accession number GSE72126 . Small pieces of kdmBΔ and the wildtype strain , grown on YMG agar , were used to inoculate 100 mL of AMM and ZM1/2 medium in 500 mL Erlenmeyer flasks . The flasks were kept on a rotary shaker at 37°C and 160 rpm until the glucose was consumed ( 64 hours for AMM and 9 days for ZM1/2 ) . pH value and glucose content of the culture fluid were monitored as described previously [95] . After harvesting , mycelium and culture fluid were separated by filtration . The culture fluid was extracted with the same volume of ethyl acetate twice , the combined organic layers were dried over sodium sulfate and the solvent was evaporated in vacuo ( 40°C ) to provide the crude extract . The wet mycelium was extracted with 100 ml of acetone for 30 min in an ultrasonic bath ( 25°C ) and the organic solvent was evaporated in vacuo ( 40°C ) . The remaining aqueous residue was diluted with 20 ml of H2O and extracted with the same volume of ethyl acetate twice . After drying over sodium sulfate , the organic solvent was removed in vacuo ( 40°C ) to yield the crude mycelial extract . The extracts were dissolved in methanol , filtered through SPE C18 cartridges and subjected to mass spectrometric analyses . All analyses were performed on Agilent 1260 Infinity Systems with diode array detector and C18 Acquity UPLC BEH column ( 2 . 1 × 50 mm , 1 . 7 μm ) from Waters . Solvent A: H2O + 0 . 1% formic acid , solvent B: AcCN + 0 . 1% formic acid , gradient system: 5% B for 0 . 5 min increasing to 100% B in 19 . 5 min , maintaining 100% B for 5 min , flowrate = 0 . 6 mL min−1 , UV detection 200–600 nm . LC-MS and MS/MS spectra were recorded on an ion trap MS ( amaZon speed , Bruker ) with an electrospray ionization source . Experiments were performed using positive and negative ionization modes . The capillary voltage of the ion source was 4500V and the nebulizer gas was set to 4 bar with drying gas flow of 12 L/min . For MS2 experiments SmartFrag was used with CID voltage of 1V and amplitude ramping of 60–180% ( fragmentation time 40 ms , Cutoff 17% ) . HR-MS spectra were recorded on a time-of-flight ( TOF ) MS ( MaXis , Bruker ) with electrospray ionization source using positive ionization mode . For data analysis and calculation of molecular formulas , including the isotopic pattern , dataAnalysis ( Version 4 . 2 ) from Bruker was used . Compound search was performed using Dictionary of Natural Products ( CRC Press ) and Antibase 2010 ( Wiley-VCH ) . ESI-MS , ESI-MS2 and UV-Vis absorption spectra of identified metabolites were compared to corresponding literature data [96–104] .
|
In this work we monitored by proteomic analysis and ChIP-seq the genome-wide distribution of several key modifications on histone H3 in the model fungus Aspergillus nidulans cultivated either under optimal physiological conditions ( active growth ) or less favourable conditions which are known to promote the production of secondary metabolites ( SM ) . When we correlated the chromatin status to transcriptional activities in actively growing cells we found that the silenced SM gene clusters are flanked by heterochromatic domains presumably contributing to silencing but that the bodies of the clusters only carry background levels of any of the investigated marks . In nutrient-depleted conditions , activating marks were invading some , but by far not all transcribed clusters , leaving open the question how activation of these regions occurs at the chromatin level . Surprisingly , a large number of these gene clusters actually depend on KdmB for normal activation and it will be interesting to see in future how this protein thought to mainly act as repressor by removing positive H3K4m3 marks switches gears to activate transcription directly or indirectly .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"aspergillus",
"fungal",
"genetics",
"gene",
"regulation",
"dna-binding",
"proteins",
"dna",
"transcription",
"aspergillus",
"nidulans",
"fungi",
"model",
"organisms",
"epigenetics",
"chromatin",
"research",
"and",
"analysis",
"methods",
"mycology",
"genomics",
"chromosome",
"biology",
"proteins",
"gene",
"expression",
"chemistry",
"molds",
"(fungi)",
"histones",
"biochemistry",
"fungal",
"genomics",
"cell",
"biology",
"post-translational",
"modification",
"acetylation",
"genetics",
"biology",
"and",
"life",
"sciences",
"yeast",
"and",
"fungal",
"models",
"physical",
"sciences",
"chemical",
"reactions",
"organisms"
] |
2016
|
KdmB, a Jumonji Histone H3 Demethylase, Regulates Genome-Wide H3K4 Trimethylation and Is Required for Normal Induction of Secondary Metabolism in Aspergillus nidulans
|
Dengue fever is an increasingly significant arthropod-borne viral disease , with at least 50 million cases per year worldwide . As with other viral pathogens , dengue virus is dependent on its host to perform the bulk of functions necessary for viral survival and replication . To be successful , dengue must manipulate host cell biological processes towards its own ends , while avoiding elimination by the immune system . Protein-protein interactions between the virus and its host are one avenue through which dengue can connect and exploit these host cellular pathways and processes . We implemented a computational approach to predict interactions between Dengue virus ( DENV ) and both of its hosts , Homo sapiens and the insect vector Aedes aegypti . Our approach is based on structural similarity between DENV and host proteins and incorporates knowledge from the literature to further support a subset of the predictions . We predict over 4 , 000 interactions between DENV and humans , as well as 176 interactions between DENV and A . aegypti . Additional filtering based on shared Gene Ontology cellular component annotation reduced the number of predictions to approximately 2 , 000 for humans and 18 for A . aegypti . Of 19 experimentally validated interactions between DENV and humans extracted from the literature , this method was able to predict nearly half ( 9 ) . Additional predictions suggest specific interactions between virus and host proteins relevant to interferon signaling , transcriptional regulation , stress , and the unfolded protein response . Dengue virus manipulates cellular processes to its advantage through specific interactions with the host's protein interaction network . The interaction networks presented here provide a set of hypothesis for further experimental investigation into the DENV life cycle as well as potential therapeutic targets .
With over 50 million cases per year , dengue virus ( DENV ) is a significant and growing threat to world-wide human health . Wide-spread among tropical and sub-tropical regions , this NIAID Category A pathogen consists of four serotypes , DENV1 to DENV4 , and is a member of the family Flaviviridae [1] . DENV causes a range of diseases in humans , from the mild Dengue Fever ( DF ) to the more deadly Dengue Hemorrhagic Fever ( DHF ) and Dengue Shock Syndrome ( DSS ) . While both the average number of cases reported to the WHO as well as the number of countries reporting cases of DENV have increased dramatically in the past five decades , relatively little is known about this important tropical pathogen that still lacks a vaccine , specific drug treatment , and relevant animal model [2] . An arbovirus , DENV is carried and spread to humans by the primary mosquito vector Aedes aegypti and to a lesser extent Aedes albopictus . Thus , DENV displays the remarkable capability to survive and replicate in two very different host organisms; accomplished by a genome encoding a mere 10 proteins [3] . To be successful , DENV must be able to manipulate each of its hosts at a molecular level . This manipulation must be accomplished , in part , through specific protein-protein interactions that allow the virus to bend existing host cellular systems to the purpose of furthering the viral lifecycle . However , understanding this host-pathogen system is particularly difficult given the complexities of host-virus dynamics as well as the lack of a useful animal model system . In light of these challenges , computational approaches provide an important tool in studies of host-pathogen systems . In particular , computational approaches for predicting host-pathogen protein interactions provide opportunities for identifying specific targets for further experimental work , understanding system behavior , and determining plausible therapeutic candidates . While common within model organism species such as S . cerevisiae , prediction of protein interactions between species are rare , and are especially so for host-pathogen interactions . Recent computational work has considered interactions between P . falciparum and human proteins based on interactions between orthologous eukaryotic proteins or using statistics about protein domains involved in within-species interactions [4] , [5] . For the HIV-human system , Tastan et al . used a data mining approach to predict host-virus interactions based on human protein features and knowledge of existing protein interactions [6] . Also focused on HIV , Evans and colleagues used short sequence motifs conserved in both HIV and humans as the basis for interaction predictions [7] . Protein structure information can also be used as a basis for protein interaction prediction [8]–[10] . Here , given a set of proteins with defined structures and known interactions , interactions can be mapped to another set of proteins possessing similar structures . This has been applied to HIV-human interactions as well as to non-viral pathogens for a number of tropical diseases [11] , [12] . Unfortunately , despite their potential value , such computational structure approaches have not been widely applied to the problem of predicting host-pathogen interactions . In particular , we are not aware of any studies focused on computational large-scale prediction of protein-protein interactions between DENV and humans and know of only one recent study related to Aedes [13] . Here , we establish a network of predicted interactions between DENV proteins and proteins from its human and insect hosts . These predictions are based on protein structural similarity , where we first determine structural similarities between pathogen and host proteins using an established method for comparing 3D structures . We refer to host proteins having a region of high structural similarity to a DENV protein as “DENV-similar . ” Next , we identify known intra-species interactions for these DENV-similar proteins , and refer to the one or more host proteins that they interact with as “targets . ” We then assume that the similar structural features seen between DENV proteins and their host DENV-similar counterparts allow the DENV protein to participate in the same interactions as DENV-similar proteins; joining the host protein network at these points ( Figure 1 ) . We prioritize the interaction map using data from recent RNAi screens , to create a smaller network of interactions having the greatest potential to be correct . These predictions include numerous novel interactions with potential functional relevance and we highlight predictions relevant to stress , the Unfolded Protein Response ( UPR ) as well as interferon pathways . This computational network approach provides an additional tool for the investigation of poorly-characterized host-pathogen systems such as DENV , as well as helping to identify potential targets in both hosts that may be used in future DENV vaccination , treatment , and control efforts .
Structures of DENV2 proteins were taken from the PDB ( downloaded on Dec . 9 , 2009 ) , and any DENV2 protein without a known structure was modeled using I-TASSER [14] , [15] . The protein sequences used to create I-TASSER models were Entrez Protein 159024813 , 159024814 , 159024817 , and 159024819 . Default settings were used , with no restraints nor selection/exclusion of any templates . Each of the structures for DENV proteins was run on the DaliLite v . 3 webserver [16] , [17] . HPRD Release 7 was used to obtain known human protein-protein interactions , while known D . melanogaster interactions were taken from DroID v5 . 0 and IntAct [18]–[20] . The A . aegypti orthologs of D . melanogaster proteins were determined using the Inparanoid ortholog annotation for the D . melanogaster genes in FlyBase v . FB2009 10 [21] , [22] . The literature sources and various databases used each have their own system of identifiers . PDB codes obtained from Dali were mapped to their corresponding taxonomy and Uniprot accessions using data from the SIFTS initiative , which aims to ease the integration of data from multiple databases ( http://www . ebi . ac . uk/msd/sifts/ ) [14] , [23] . Other identifier mappings were carried out using DAVID Gene ID Conversion or Uniprot ID mapping ( DAVID 6 . 7 , Uniprot Release 15 . 14 ) [24]–[26] . Network diagrams were created in Cytoscape [27] . Images of protein structures were created in MacPyMol [28] . We investigated protein mimicry using structural similarities from Dali . Dali compares the 3D structural coordinates of two PDB entries by alignment of alpha carbon distance matrices , allowing for differences in domain order , and produces a structural similarity score [14] , [16] , [17] . For this study , we ran each of the DENV2 protein structures , both known and predicted , through the DaliLite webserver , which searched against the entire PDB for structurally similar proteins , with a z score above 2 . 0 . Default settings of a score cutoff of 40 bits and sequence overlap cutoff of 50% were used . We then took from these results only those structures that were from the species H . sapiens and D . melanogaster . We refer to these human proteins as hDENV-similar proteins and the fly proteins as dDENV-similar proteins . To predict which human proteins may interact with DENV2 proteins , we sought those target human proteins that interact with the hDENV-similar proteins during cellular processes . To this end , we determined known interactions between hDENV-similar proteins and target human proteins , using data from the Human Protein Reference Database ( HPRD ) database , which contains literature curated interactions between pairs of human proteins [18] . For each hDENV-similar protein , we predicted that the target proteins which are known to interact with the hDENV-similar protein might also interact with that DENV protein . A similar process was used to predict interactions between DENV2 proteins and A . aegypti proteins , but with the added step of finding orthologs between A . aegypti and D . melanogaster proteins . Known interactions between the dDENV-similar proteins and other D . melanogaster proteins were taken from DroID , using a cutoff confidence value of 0 . 4 , and IntAct [19] , [20] . Then , orthologs of the D . melanogaster proteins were found for A . aegypti using FlyBase [22] . We then made the prediction that the A . aegypti target protein interacts with the DENV protein . The Gene Ontology ( GO ) provides a system of terms to consistently describe and annotate gene products [29] . GO term enrichment was performed using the DAVID Functional Annotation Chart tool [24] , [25] . The GO is organized as a tree structure , with terms becoming more specific as distance from the root increases . Therefore , to avoid very general and uninformative GO terms , we used only GO level 4 terms . The p-values were corrected for multiple testing using the Bonferroni procedure and transformed . Since there may be multiple PDB structures present in Dali to represent the same protein , there was some redundancy in the interaction predictions . In some cases , multiple PDB structures for the same DENV protein were found to be similar to multiple PDB structures for an DENV-similar protein , leading to the same interaction predictions . Therefore , the predictions were counted as unique pairs of human Uniprot accessions and DENV protein names . In addition , for ease of viewing the predicted interactome , each node representing an DENV protein is labeled with the protein name while each human protein is represented by Entrez GeneID . Support for the predicted interactions was obtained from literature . As few interactions between DENV and humans are known , we looked within the literature to see if any of them were predicted by our method [30]–[33] . In addition , recent studies using siRNA screens have found proteins that may play some role , either facilitatory or inhibitory in DENV infection , in both humans and D . melanogaster [34] , [35] . A . aegypti orthologs of these host factors were recently curated by Guo et al . [26] . We checked for the presence of these human host factors or mosquito orthologs among our predictions . Although it is not known if these proteins act through direct protein-protein interactions with DENV or indirect mechanisms , their involvement in DENV infection provides functional support for a possible interaction and gives them higher priority for further testing . GO annotations for the human and A . aegypti target proteins were obtained through DAVID 6 . 7 [24] , [25] . However , since DAVID assigns all DENV proteins the same GO terms , GO annotation for the DENV proteins was obtained using the GOanna webserver , provided through AgBase v . 2 . 00 [36] . This tool assigns GO terms to the input sequences by transitively assigning the GO terms of similar , already annotated sequences identified by BLAST . The most significant BLAST hits for the DENV protein sequences were DENV polyprotein sequences . However , there were multiple polyprotein sequences , each with their own annotations . The input sequences matched more significantly to some polyproteins than to others , and were therefore assigned different GO terms based on sequence similarity . The predicted interactions were filtered so that only those predictions for which the DENV protein and host protein shared at least one GO cellular component term were retained . Guo et al . recently generated a first draft of the mosquito interactome [13] . Because their interactome was based on the three model organisms A . aegypti , C . elegans , and S . cerevisiae , we found proteins from all three of these species that show structural similarity with DENV2 using the Dali server [14] , [16] . The A . aegypti orthologs of C . elegans and S . cerevisiae proteins were determined using InParanoid , and the D . melanogaster orthologs were taken from the InParanoid ortholog annotation for the D . melanogaster genes in FlyBase v . FB2009 10 [21] , [22] . Then , the interactions with these orthologs taken from the mosquito interactome were used to map predicted interactions between DENV2 and mosquito target proteins . For the GO term enrichment , we used only GO terms from DAVID's GO fat set , to eliminate non-specific terms with many children . The genome-wide set of orthologs between human and A . aegypti was downloaded from InParanoid 7 . 0 Refworks:161 . Since InParanoid lists human proteins by their Ensembl Protein IDs , mappings to Uniprot accessions were downloaded from Ensembl 57 using BioMart Refworks:192 . Orthologous human and A . aegypti targets that were predicted to interact with the same DENV2 protein were identified .
To develop a network of interactions between DENV and its hosts , H . sapiens and A . aegypti , we employed a method we developed previously in the prediction of protein interactions between HIV and human [12] ( see Methods for further details ) . First , we obtained 3D structures for the DENV proteins , from two sources . Experimentally determined structures were taken from the PDB and consist of 31 PDB entries representing the DENV2 proteins E , pr peptide , prM , C , NS2B , NS3 , and NS5 [37] . Since there are no experimentally determined structures for NS1 , NS2A , NS4A , and NS4B , we used the I-TASSER server to predict the structure of these proteins [15] . In this way , we investigated possible interactions for every DENV protein . To determine structurally similar host proteins , we used DaliLite to compare DENV structures against every other structure in the PDB [16] , [17] , [37] . We considered only significant structural matches with proteins from DENV's hosts . We found 300 human proteins with similarity to a DENV protein ( hDENV-similar ) . However , we found no similarities between DENV proteins and A . aegypti proteins . This is not surprising , given there are currently only 17 structures from A . aegypti in the PDB . Therefore , we looked for similarities between DENV proteins and the fly , Drosophila melanogaster , and found 15 proteins with structural similarity to DENV , which were then used as dDENV-similar proteins in downstream analyses . A particular challenge in host-pathogen studies is the general lack of interaction data . HIV is perhaps the most well-characterized virus in this regard , with over 800 direct interactions documented in NCBI's HIV-Human protein interaction database ( over 2500 interactions if indirect interactions are included ) [38] . In contrast , a recent compilation of host-pathogen interactions from public databases describes a total of 3 DENV-human interactions [39] . Through a more comprehensive search of the literature , we have found 20 documented interactions between DENV and human proteins ( Table 1 ) . Almost half of the documented protein interactions involve E protein and a receptor on the cell surface . Two of these , CD14 and HSPA5 , have been shown to function as DENV receptors , although their binary interaction with E protein was not explicitly demonstrated; it may be that some other protein in complex with these receptors is the direct interaction partner of E protein [40] , [41] . Furthermore , there is evidence that DENV receptor usage may be virus strain and cell type dependent [42] . Indeed , RPSA has been shown to be a DENV receptor , suggesting an interaction with E , but only for DENV1 [43] . Because our predictions were focused on DENV2 , this interaction was not considered for our predictions , but was included in Table 1 . Interactions not shown to be specific for a different serotype were included in our list of true positive interactions . Therefore , a total of 19 protein interactions were considered as known host-pathogen interactions between DENV2 and human . We are currently unaware of any well-characterized protein-protein interactions between DENV and A . aegypti . However , in the C6/36 cell line from Aedes albopictus , tubulin is believed to interact with DENV2 E protein [44] . In addition , one protein , likely to be HSP90 , has been put forward as a putative receptor for DENV2 in A . aegypti , having been shown to bind to the E protein [45] . However , its identity has not been conclusively demonstrated . In addition , mosquito La auto-antigen is known to interact with the 3′ end of DENV RNA and may play some role in RNA synthesis [46] . Human La auto-antigen ( SSB ) is also known to interact with the ends of the viral RNA , as well as NS3 and NS5 [47] . If the functions of the mosquito and human La proteins in DENV infection are similarly conserved , mosquito La may interact with NS3 and NS5 as well , although this has not been shown . It is likely that some of the protein interactions which enable DENV to manipulate the cellular pathways of two hosts are conserved between the species . After determining which host proteins are structurally similar to DENV proteins , we inquired into the known protein-protein interactions that each DENV-similar protein participates in . For the hDENV-similar set , we obtained known human protein interactions from the Human Protein Reference Database ( HPRD ) , which consists of over 37 , 000 interactions found in the literature [18] . We predicted that the DENV proteins could interact with the partners of their corresponding hDENV-similar proteins , under the hypothesis that proteins with highly-similar structures are likely to be involved in similar protein interactions ( Figure 1A ) . We predicted 4 , 273 potential host-pathogen interactions , involving 2 , 321 different human proteins ( Table 2 ) . Of the 19 known protein-protein interactions between DENV and human , 9 are present among our predictions [30]–[33] , [40] , [41] , [47]–[55] . This method may not predict all interactions , for example those mediated by sequence motifs rather than structural features . A table of all DENV-human protein interaction predictions is provided in Table S1 . For the dDENV-similar proteins , we used the interactions curated in IntAct for D . melanogaster , as well as potential D . melanogaster interactions suggested by the yeast-2-hybrid data sets in DroID [19] , [20] . However , rather than making direct predictions using these interactions , as we did for human proteins , we determined orthologs of the D . melanogaster proteins in A . aegypti , since this is the true host of DENV . We then predict that the A . aegypti ortholog of a D . melanogaster protein that interacts with a dDENV-similar protein may also interact with the corresponding DENV protein ( Figure 1B ) . As a result of this procedure , we predict that 158 A . aegypti proteins participate in 176 interactions with DENV proteins ( Table 2 ) . We note that this method did not predict interactions between E and mosquito tubulin , HSP90 , or La , which have been suggested as possible interactions [45]–[47] . However , 12 of the predictions involved orthologs of proteins involved in DENV2 infection in humans or fly . A table of all DENV-Aedes protein interaction predictions is provided in Table S2 . We note several links between DENV pathogenesis , stress responses , and apoptosis among our predictions and in the literature . The GO term “regulation of stress response” is enriched among hDENV-similar proteins , as well as several terms related to apoptosis among the human target proteins . Several potential DENV receptors are involved in stress responses , such as HSP90 , HSPA4 , and HSPA5 [33] , [41] . In particular , flaviviruses assemble within and bud from the ER , and are known to induce the Unfolded Protein Response ( UPR ) , which reacts to stressors of ER function . As the UPR is necessary for cell survival during infection , but also has a negative impact on viral replication , modulation of this response by DENV may be advantageous . The UPR can induce either survival or apoptosis signals depending on the strength and duration of the ER stressor [66] . Three major branches , running through PERK , ATF6 and IRE1 , regulate the UPR and ER homeostasis , and all three have been shown to be induced by DENV infection ( Figure 5 ) [67] . Activation of PKR-like ER kinase ( PERK ) , leads to phosphorylation of the eukaryotic initiation factor 2 ( eIF2 ) . This phosphorylation inhibits the formation of translation initiation complexes , leading to translation inhibition and a reduction in the number of unfolded proteins within the ER [66] . Production of ATF4 is also enhanced as a result of eIF2 phosphorylation , eventually leading to GADD34 production . GADD34 acts within a negative feedback loop , recruiting protein phosphatase 1 , leading to dephosphorylation eIF2 and the restoration of normal translation efficiency . Persistent ER stress leads to CHOP expression and promotion of apoptosis . It has been suggested that DENV may be able to compensate the UPR response by inducing dephosphorylation of eIF2 to restore translation [67] . We predict interactions between NS4B and GADD34 ( PPP1R15A ) ( Figure 5 ) . A listing of all predictions , with or without filtering , are provided in Tables S1 , S2 , S3 , S4 , S5 . ATF6 is a bZIP family transcription factor that transits from the ER to the Golgi in response to ER stress . It undergoes processing in the Golgi and transits to the nucleus , leading to upregulation of multiple apoptosis-relevant genes and eventual apoptosis . While we do not predict any direct interactions with ATF6 , we do predict interactions between NS2A , NS4B , and C with associated pathway member NFYA , which forms a complex with ATF6 in response to ER stress [68] ( Figure 5 ) . In the third branch of the UPR , the ER transmembrane protein IRE1 , containing both kinase and RNase activities , becomes autophosphorylated and activated in response to ER stress , leading to XBP-1 splicing and translation of UPR relevant genes . Both DENV2 and Japanese Encephalitis Virus infection have been shown to activate XBP1 and its downstream genes in N18 mouse neuroblastoma cells , reducing the cytopathic effect of the virus [69] . Knockdown of XBP1 expression by siRNA has also been shown to lead to greater cytotoxicity in response to infection [69] . Persistent stress leads to apoptosis through an IRE-JNK-BCL2 pathway . Our predictions suggest potential interactions between E and BCL2 . In addition , other BCL family members are also predicted to interact with DENV proteins including BCL2ll ( BIM; a facilitator of apoptosis ) with NS4B , BCL2L1 ( BCLX; both pro- and anti-apoptotic splice variants ) with E and NS3 , and BCL2L10 ( Boo; supression of apoptosis induced by BAX but not BAK ) with NS3 ( Figure 5 ) . A recent study investigating protein interactions between DENV envelope protein and host proteins described direct interactions between E and BiP ( HSPA5 ) , Calnexin ( CANX ) and Calreticulin ( CALR ) [55] . All three major ER stress transducers interact with BiP , which serves as a negative UPR regulator , and along with other ER chaperones , facilitates proper folding of proteins . Similarly , Calnexin and calreticulin are chaperones that bind to glycosylated proteins . Our methodology predicts each of these interactions with the E protein . In addition , we also predict that CALR is likely to interact with NS1 as well as NS3 . Overall , these results suggests multiple sites within the host network at which DENV proteins can potentially manipulate the Unfolded Protein Response . In addition , several studies have implicated NS3 in DENV-induced apoptosis . The ability of DENV1 to cause apoptosis in HepG2 cells differs across strains . The mouse neurovirulent strain FGA/NA d1d differs from its parental strain , FGA/89 by 4 mutations , one of which leads to a non-conservative substitution in NS3 . FGA/NA d1d was shown to have a reduced capacity to induce apoptosis , although whether this was mediated by the mutation in NS3 or by one of the other mutations , which were all in the E protein , is unclear [70] . However , Vero cells expressing DENV2 NS3 undergo apoptosis by a mechanism that is dependent on NS3 protease activity and enhanced by the presence of NS2B [71] . In addition , West Nile Virus NS3 is sufficient to induce caspase-8-depedent apoptosis , and is suggested to directly interact with , cleave , and activate caspase-8 in NIH 3T3 cells [72] . We predicted a number of interactions between NS3 and members of apoptotic pathways . A few examples of the structural similarities that led to these predictions are shown in Figure 6 . NS3 was predicted to interact with p53 based on structural similarities with RAD51 , TK1 , and DDX5 . Similarities with RAD51 also led to predicted interactions with ABL1 , BRCA1/2 , CASP3 , and CASP7 . Furthermore , NS3 has regions of similarity to APAF1 , and is therefore predicted to interact with BCL2L1 , BCL2L10 , Fas , and the caspases -3 , -4 , -8 , and -9 . These results suggest that NS3 may play a role in DENV pathogenesis by influencing apoptosis in host cells , mediated by specific interactions between NS3 and host proteins . Humans and A . aegypti are known to use conserved defenses against DENV infection , involving several signaling pathways of the innate immune system , which is consistent with our finding of enriched GO terms related to the immune system among the target proteins of both humans and A . aegypti . In particular , the JAK-STAT signaling pathway has been shown to modulate susceptibility to DENV infection , in both mosquitos and humans [73] , [74] . In humans , the JAK-STAT pathway can be activated by the interferons ( IFN ) , IFN- , IFN- , and IFN- , and mediates the antiviral response ( Figure 7A ) . When IFN- or IFN- bind their receptor , IFNAR , the tyrosine kinases JAK1 and TYK2 are activated . This results in the phosphorylation and activation of STAT2 and STAT1 , which then recruit IRF9 to form a transcription factor complex that transcribes IRF-7 and then the set of genes that are induced by IFN- . The interferon response is known to be induced upon DENV infection and high levels of IFN- are normally present in the sera of DENV patients [75] . Furthermore , pretreatment of cells with IFN has been shown to block negative strand accumulation of DENV RNA , but this inhibition was strongly attenuated if treatment occurred even 4 hours after initial infection [76] . In fact , while infection induces an interferon response , several studies have shown that DENV interferes with the signaling pathway downstream of IFN- . In dendritic cells , DENV is known to protect itself from the antiviral effects of IFN- by reducing the phosphorylation of TYK2 and preventing the activation of STAT1 and STAT3 , although the effects of IFN- are not averted [74] . In addition , IFN-- , but not IFN-- , dependent phosphorylation of STAT1 and STAT2 was found to be inhibited in A529 and HepG2 cells by the NGC strain of DENV2 , but not by the strain TSV01 , suggesting strain-dependent rather than serotype-specific differences in response [77] . Such strain-dependent differences also highlight the possibility of viral RNA sequence variations that potentially lead to changes in interaction specificity or the strength of interaction with host proteins . A few specific proteins have been identified as modulators of the IFN response . For instance , IFN- signaling was prevented by the viral proteins NS4B , and to a lesser degree by NS4A and NS2A [78] . Inhibition of signaling by NS4B was thought to occur through an observed reduction in the level of phosphorylated STAT1 . Expression of STAT2 was also observed to be repressed following infection [79] . Recently , NS5 has also been shown to bind to STAT2 , resulting in reduced IFN signaling [80] . In this same study , when expressed as a proteolytically processed precursor , NS5 was also found to target STAT2 for proteasome-mediated degradation . However , while it is clear that DENV is actively involved in modulating the host interferon response , there likely remain many specific interactions by which DENV proteins inhibit IFN signaling that are not known . Our predictions suggest many potential interactions between DENV and multiple human proteins involved in the JAK-STAT pathway ( Figure 7B ) . In particular , for NS4B , we have predicted possible interactions with IFNAR2 , IFNGR1 , JAK1 , JAK2 , TYK2 , PTPN11 , PTPN2 , PTPN6 , PKR ( EIF2AK2 ) , STAT1 , STAT2 , and STAT3 . Thus , NS4B may reduce the observed phosphorylation of STAT1 through direct interactions with the host STAT1 protein , through interactions that effect the activities of proteins upstream of STAT1 ( e . g . JAK1 or TYK2 ) , or through interactions with at least one PTP protein , which are negative regulators of STAT activity . In addition , NS4B is predicted to interact with PKR , a key component of the IFN response in blocking virus replication . A close relative of DENV , hepatitis C virus , has been shown to inhibit interferon signaling through inhibition of PKR as well as by competing with eukaryotic translation initiation factor 2 as a PKR substrate [81] , [82] . NS2A is also predicted to interact with the same members of the this signaling pathway as NS4B . However , NS4A is predicted to interact with IFNAR1 , IFNAR2 , IFNB1 , JAK2 , TYK2 , PTPN11 , and SOCS1 . In summary , we have predicted specific host-pathogen protein interactions that may enable DENV to escape the antiviral response induced by IFN and which can be tested in the future to determine the precise mechanism by which DENV manipulates this host system . To complete its lifecycle , DENV must survive in two very different hosts and must perform many of the same processes in each , such as transcription and translation . Since some proteins and essential processes are conserved between mosquitos and humans , it is possible that some of the proteins that are manipulated by DENV in one host are orthologous to the proteins used in the other host . To identify potential interactions of this type , we compared our interaction network predictions before CC filtering for human and A . aegypti to find proteins which are orthologous between the two hosts , as well as predicted to interact with the same DENV protein in each . We found 47 pairs of orthologs that were predicted to interact with the same DENV protein ( Figure 8 ) . Four of these predicted interactions represent host factor targets for A . aegypti . We note that our method depends on known interactions within species , and may miss some orthologous host-pathogen interactions if the within species interactions are not known . To examine the functional role of these conserved interactions , we performed GO term enrichment for biological process and molecular function . No molecular function terms were significantly enriched ( Bonferroni corrected p-value <0 . 01 ) . Most of the biological processes represented by orthologous interactions in both hosts were also found enriched in the predictions for a single host . For example , at least 17 of the human proteins and their orthologs in A . aegypti are involved in mRNA processing or metabolism ( ASCC3L1 , DCP1B , DCP2 , HNRPA1 , HNRPD , HNRPF , HNRPM , UPF3B , WDR77 , MAGOH , NCBP2 , PABPN1 , PAPOLA , PRMT5 , SNRPA1 , SF3B3 , SMN1 ) . Five of the interactions known to occur between DENV and human proteins involve the mRNA processing proteins HNRNPK , HNRNPC , PTBP1 , and SSB [30] , [47] , [50] , [53] . “RNA processing” and “RNA metabolic process” were highly enriched in the mosquito predictions and in the orthologous predictions . In addition , we found a number of enriched GO terms relating to the regulation or formation of synapses among the predictions conserved in both species . Previously , the term “brain development” was enriched in the set of dDENV-similar proteins . DENV2 virus particles have been found in vesicles near the presynaptic membrane in spinal cords of SCID mice , and it was suggested that fusion of these vesicles at the synapse might aid the spread of DENV2 from neuron to neuron [83] . In Culex pipiens quinquefasciatus mosquitos , West Nile Virus , another flavivirus , was also found near synapses and in synaptic vesicles [84] . Furthermore , DENV3 is known to infect the nervous system of A . aegypti , altering the mosquito's feeding behavior by prolonging feeding and possibly enhancing the spread of DENV3 by making it more likely that feeding will be interrupted , and the mosquito will have to feed on additional humans [85] . The processes enriched in the predicted interactions conserved between the two hosts are consistent with the effects of DENV infection in each host . In particular , many of the terms enriched among the orthologous predictions are similar to terms enriched for mosquito predictions . This is not necessarily surprising , given that the mosquito prediction set is much smaller than the human one , but indicates that orthologous predictions in humans were made corresponding to many of the mosquito predictions . However , we found terms involving cell morphogenesis enriched among the orthologous predictions , but not within the predictions specific to either host . Therefore , the mosquito-specific predictions do not completely overlap with the human predictions , and new processes key to DENV infection in both hosts can be revealed . We have created a map of potential protein-protein interactions between the host-pathogen triad DENV2 and its hosts H . sapiens and A . aegypti . The computational methodology employed to generate this map assumes that proteins with comparable structures will share interaction partners . Therefore , we predict that DENV2 proteins may merge into the host protein interactome at the points normally occupied by structurally similar host proteins , creating an interface for the manipulation of downstream host processes . From previous studies , a number of human and fly proteins have been suggested to play some role in DENV2 infection , although the nature of this role is unknown in most cases . Using this methodology , we are able to make predictions regarding which host proteins may impact viral infection through interactions with specific DENV2 proteins . We note that the structural-based methodology here provides a larger picture of the interaction network , while more subtle changes at the sequence level are likely to explain experimentally observed differences in strain effects . Given the paucity of both structural and interaction data for this system , we cannot determine fine differences between strains , but this may be elucidated by further study . The networks presented here may help to provide a set of hypotheses for further investigation , potential therapeutic intervention , as well as help in improving our understanding of the DENV life cycle .
|
Dengue virus ( DENV ) represents a major disease burden in tropical and subtropical regions of the world , and has shown an increase in the number of cases in recent years . DENV is transmitted to humans through the bite of an infected mosquito , typically Aedes aegypti , after which it begins the infection and replication lifecycle within human cells . To perform the molecular functions required for invasion , replication , and spread of the virus , proteins encoded by DENV must interact with and alter the behavior of protein networks in both of these hosts . In this work , we used a computational method based on protein structures to predict interactions between DENV and its human and insect hosts . We predict numerous interactions , with many involved in known cell death , stress , and immune system pathways . Further investigation of these predicted protein-protein interactions should provide targets to combat the clinical manifestations of this disease in humans as well as points of intervention focused within the mosquito vector .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"and",
"Discussion"
] |
[
"virology/immune",
"evasion",
"virology/viral",
"replication",
"and",
"gene",
"regulation",
"infectious",
"diseases/neglected",
"tropical",
"diseases",
"computational",
"biology/protein",
"homology",
"detection",
"molecular",
"biology/bioinformatics",
"infectious",
"diseases/viral",
"infections",
"computational",
"biology/protein",
"structure",
"prediction",
"computational",
"biology/signaling",
"networks",
"biochemistry/structural",
"genomics",
"infectious",
"diseases/tropical",
"and",
"travel-associated",
"diseases",
"computational",
"biology/systems",
"biology"
] |
2011
|
Mapping Protein Interactions between Dengue Virus and Its Human and Insect Hosts
|
Advances in time-lapse fluorescence microscopy have enabled us to directly observe dynamic cellular phenomena . Although the techniques themselves have promoted the understanding of dynamic cellular functions , the vast number of images acquired has generated a need for automated processing tools to extract statistical information . A problem underlying the analysis of time-lapse cell images is the lack of rigorous methods to extract morphodynamic properties . Here , we propose an algorithm called edge evolution tracking ( EET ) to quantify the relationship between local morphological changes and local fluorescence intensities around a cell edge using time-lapse microscopy images . This algorithm enables us to trace the local edge extension and contraction by defining subdivided edges and their corresponding positions in successive frames . Thus , this algorithm enables the investigation of cross-correlations between local morphological changes and local intensity of fluorescent signals by considering the time shifts . By applying EET to fluorescence resonance energy transfer images of the Rho-family GTPases Rac1 , Cdc42 , and RhoA , we examined the cross-correlation between the local area difference and GTPase activity . The calculated correlations changed with time-shifts as expected , but surprisingly , the peak of the correlation coefficients appeared with a 6–8 min time shift of morphological changes and preceded the Rac1 or Cdc42 activities . Our method enables the quantification of the dynamics of local morphological change and local protein activity and statistical investigation of the relationship between them by considering time shifts in the relationship . Thus , this algorithm extends the value of time-lapse imaging data to better understand dynamics of cellular function .
Cell morphological change is a key process in the development and homeostasis of multicellular organisms [1] , [2] . Various types of morphological change appear during migration and differentiation; essential events occurring as part of these processes usually accompany morphologically different phenotypes . Therefore , cell morphology has been used as a key indicator of cell state [3] . High-throughput analyses of cell morphodynamic properties have been used recently to discover new functions of specific proteins [4] . Moreover , the outcomes of morphological change such as the intricate shape of neuronal dendrites , remind us that morphogenesis itself plays a role in the emergence of cellular function [5] . Quantitative approaches are helping to unveil cellular morphodynamic systems , and they are generating new technical requirements . Because cellular morphological change is highly dynamic , time-lapse imaging is necessary to understand the mechanism of cell morphology regulation . Progress in the development of fluorescent probes has enabled the direct observation of cell morphological changes and/or the localization and activity of specific proteins [6]–[8] , but time-lapse imaging has highlighted the difficulty of extracting characteristic information from an immense number of images . Nevertheless , several approaches in the context of quantitative analysis have appeared recently . A series of studies using quantitative fluorescent speckle microscopy , for instance , revealed the power of computer-assisted high-throughput analysis for time-lapse microscopy images: analysis of the number of moving and blinking speckles suggested distinct regulation of actin reorganization dynamics in different intracellular regions [9] , [10] . Indeed , computational methods have been used to determine the properties of morphological dynamics , protein activity and gene expression [11]–[14] . There are two major approaches for the detailed analysis of local morphological changes of cells . One is the kymograph , which is a widely used method to describe motion with a time-position map of the morphology time course . The time course of change in intensity could also be monitored by arranging sequential images of a specific region of interest ( ROI ) [15] . Although there are drawbacks to this approach , such as restriction of the analyzed area to a narrow ROI and the need to manually define the ROI , recent studies have avoided these limitations by using polar coordinates to explore the motility dynamics of the entire peripheral region of round cells . Indeed , the polar coordinate-based approach showed isotropic and anisotropic cell expansion , and examined stochastic , transient extension periods ( named STEP ) or periodic contractions [12] , [16] . The second approach is to track cellular edge boundaries by tracing virtually defined markers . Kass and Terzopoulos introduced an active contour model known as SNAKES [17] , which is widely used to analyze moving video images in applications including biomedicine . For example , Dormann et al . used SNAKES to quantify cell motility and analyze the specific translocation of PH domain-containing proteins into the leading edge [14] . Marker-based tracking has advantages in quantifying highly motile cell morphology , because it does not require a fixed axis , which is necessary in the kymograph approach . Recently , Machacek and Danuser developed an elegant framework to trace a moving edge , using marker tracking modified by the level set method to elucidate morphodynamic modes of various motile cells such as fibroblasts , epithelial cells , and keratocytes [18] . Although previous methodologies have successfully described the specific aspects of cellular morphodynamics , there remain challenges to quantify the relationship between morphodynamics and signaling events . One representative problem is the association between regions in different frames . To scrutinize the dynamic relationship between morphological change and molecular signaling , we need to cross-correlate them in a time-dependent manner ( Figure 1A ) . A polar coordinate system does not ensure the association of time-shifted local domains ( Figure 1B ) , and is unsuitable for non-circular cell shapes . The virtual marker tracking method satisfies this requirement for cells with broadly consistent shapes , but its fixed number of markers causes unequal distribution when a dramatic shape change such as the persistent growth of neurites in neurons , occurs ( Figure 1C ) . Taking these problems into account , we perceive the need for a novel quantification method to better understand the mechanisms of morphodynamic regulation by molecular signaling . We focused on the Rho-family small GTPases , or Rho GTPases , as signaling molecules associated with cell morphodynamics . Rho GTPases , which act as binary switches by cycling between inactive and active states ( Figure 2 ) , play key roles in linking biochemical signaling with biophysical cellular behaviors [19] , [20] mainly through reorganization of the actin and microtubule cytoskeleton [21] . It is well known that RhoA , Rac1 , and Cdc42 have unique abilities to induce specific filamentous actin structures , i . e . , stress fibers , lamellipodia , and filopodia , respectively [19] . Considerable evidence , mainly obtained using constitutively-active or dominant-negative mutants , supports a promotional role of Rac1 and Cdc42 and an inhibitory role of RhoA in cell protrusion [19] , [21] . Although some researchers have challenged this widely-accepted notion in a variety of cell contexts [22]–[24] , our current study has been motivated by this predominant view . The objective of this study was to uncover the relationship between spatio-temporal activities of Rho GTPases and morphological changes of the cells . To achieve this , we needed a data analysis tool to assess the link between biochemical signaling and biophysical phenomena . However , we do not focus on unveiling the orchestration of the complete signaling pathways that regulate cell morphology . In addition , we elucidated how Rho GTPases regulate “two-dimensional” morphological changes of cells , rather than “three-dimensional” changes . These findings will however be meaningful because the results can be compared with earlier findings [25]–[28] . Therefore , we first present an algorithm called edge evolution tracking ( EET ) to quantify local morphological change . The main features of our method are that ( 1 ) identification of a local morphological change is based on an area difference between two consecutive frames; ( 2 ) cell edge is not characterized by point markers , but by line segments , which are defined by the area difference; and ( 3 ) past history and future evolution of each segment can be evaluated by connecting segments between consecutive frames . Therefore , this method enables us to trace complex cell edge extension and contraction while maintaining the consistency of the ROI during the analysis . Second , applying EET to fluorescence resonance energy transfer ( FRET ) time-lapse images of three Rho GTPases ( Rac1/Cdc42/RhoA ) , we found a significant time-shifted cross-correlation between morphological change and GTPase activity . Our study reveals the utility of detailed cellular morphodynamic profiling and spatio-temporal signal profiling to measure the time-shifted relationship between morphodynamics and protein activity .
The EET algorithm describes the time course of local cell morphological changes based on area differences of sequential images . We focused on the local area change , rather than the local structural change as a morphological property; therefore , EET analysis did not make clear distinctions between filopodia and lamellipodia . Subdivided regions along the cell edge boundaries are connected to the corresponding subdivided regions in the next frame , and movements of the subdivided regions are then defined by these connected subregions . Thus , the subdivided regions called “segments” are basic units in EET for quantification of morphological changes . EET describes the time course of local protrusion and retraction as follows: These connected anchor points indicate the spatial associations between neighboring time frames , and allow us to trace the corresponding regions along the time course by means of the graph structure , which represents the lineage of the segments along the time course . A flow chart of the EET procedure above is shown in Figure 3D . It should be noted that EET defines how the ancestral segments of a certain segment at a certain time behave along the time course ( Figure 3E ) . Because the definition of segments depends on area differences , if a cell becomes transiently immobile the subdivided regions fuse into a single , and hence integrated , edge . In such a case , integration can be avoided by an exceptional operation that maintains the anchor points during the period of immobility . This procedure keeps the spatial resolution ( number of segments ) of EET without artificial bias as far as used for immobile anchor points , because the average activity of a single segment and that of its divided segments are the same , and the area differences are always 0 . Generally , however , continuous fluctuation is observed along the whole edge , and it is therefore possible to extract a sufficient number of subdivided regions to be analyzed . Actually , this exceptional operation is not used when analyzing the data in this manuscript . Although threshold parameters for the binarization in preprocessing affect the extraction of cell boundaries and area differences , the results of EET are consistent once the threshold parameters have been determined , even if cells show highly fluctuating behavior . Local activity along a cell boundary is defined as the mean FRET ratio inside a circle , which has its center on the cell boundary and radius r . This is equivalent to using a smoothing filter with a kernel size of r . In EET , the representative activity of a segment is defined as the mean of local activities in the segment . We thus obtain a vector of activity a , composed of the representative activities within each segment in time-lapse images . In polar coordinate-based and marker-tracking based analyses , on the other hand , local activity denotes the mean activity inside a circle , whose center is located at the intersection of a cell boundary and a radial axis or a marker position , and whose radius is r . Therefore , local activity is defined in EET in a manner that is conceptually similar to that in the polar coordinate-based and the marker-tracking-based methods; however , the EET analysis is performed segment-by-segment , which is statistically more stable than the polar coordinate-based and marker-tracking-based methods . The activity profile at time N , obtained from N images of time-lapse activity data , is denoted by aN . We calculated cross-correlation coefficients between local area changes and activities based on the defined segments . Vector data {a ( t ) |t = 1 , … , N−1} and {d ( t ) |t = 1 , … , N} denote the activities and area differences of the segments extracted from the first to N−1 and the first to N frames of the same image sequence , respectively . a ( t ) and d ( t ) represent local activities at time t and local area differences between times t and t+1 , respectively . According to Pearson's product-moment correlation coefficient , the correlation function R ( {a ( t ) } , {d ( t ) } , N ) is defined aswhere i and t are indices for segments denoting positions along a cell boundary and time ( frames ) , respectively , and Mt denotes the number of segments at frame t . Note that our EET defines the activity in a segment-wise manner , and therefore a ( t ) and d ( t+1 ) have the same dimensionality Mt . Because the histogram of the activities in the segments was found to be approximated as a normal ( Gaussian ) distribution but with a heavy tail in some samples , samples whose activity exceeded 3σ ( where σ is the standard deviation ) were removed to avoid disproportionate influences of outliers on the correlation coefficients . When the data distributions diverged from the Gaussian , we also calculated Spearman's rank correlation coefficient , which is independent of the shape of the sample distributions , to verify the results of the Pearson's correlation coefficient . Spearman's rank correlation function Rs ( {a ( t ) } , {d ( t ) } , N ) is defined aswhere , pj is the number of rank j samples of {d ( t ) |t = 1 , … , N} , nd the number of ranks in {d ( t ) |t = 1 , … , N} , pa the number of rank k samples of {a ( t ) |t = 1 , … , N−1} , and na the number of ranks in {a ( t ) |t = 1 , … , N−1} . Because the EET calculates the cross-correlations based on segments , it is insensitive to the physical size of segments; that is , the cross-correlation coefficients indicate event-wise correlations between molecular activities and morphological changes over the whole cell edge . We investigated τ time-shifted cross-correlation between activities and area differences to incorporate the time lag between molecular events and morphodynamics . Because the ancestry relationship between a single segment in a focused frame and segments in another frame is not one-to-one ( Figure 3C ) , we defined the transition matrix At , t+τ so that the τ-shifted area difference d ( t , t+τ ) could be defined . Because the graph structure was obtained under the basic assumption that each local event is defined in terms of ‘segment’ , a morphological property , we calculate the τ-shifted values only for the area differences . A series of the corresponding area differences by sequential τ , for example , d ( t , t+1 ) , d ( t , t+2 ) , d ( t , t+3 ) , … , denotes the time course of edge evolution; d ( t , t+τ ) is defined below . The transition between Mt segments at time t and Mt+1 segments at time t+1 is represented by an Mt×Mt+1 matrix At , t+1 , which consists of 0 and 1 denoting unconnected and connected segments , respectively , in the ancestry graph ( Figure 3C ) . Because the column dimensionality of the transition matrix at time t and the row dimensionality of the transition matrix at time t+1 are the same as the number of segments between time t and time t−1 , the transition matrix between time t and t+u can be calculated algebraically asThis means that each component is substituted by one if the matrix calculation results in a positive value . Corresponding area changes from time t to time t+τ are then expressed as:The i-th element of d ( t , t+τ ) denotes the summation of area differences among the segments at t+τ , which are ancestral to the i-th segment at t , according to the ancestry graph . In Figure 3C , for example , d ( T ) = {la , ab , br} and d ( T+1 ) = {lc , cd , de , ef , fr} , where each element in the sets denotes an area difference ( typically , a number of pixels ) . The transition matrix is given by:Then , d ( T+1 , T ) = ( AT , T+1 ( d ( T+1 ) ) ’ ) ’ = {lc+cd+de , de+ef+fr , fr} , where the addition is applied to the area difference values . Based on these time-shifted corresponding area differences , a one-to-one relationship between the segments in different frames is constructed . The cross-correlation coefficient with a time-shift of τ is thus obtained by calculating R ( {a ( t ) } , {d ( t , t+τ ) } , N−τ ) . In this study , cell boundaries and area differences were all extracted from fluorescence time-lapse images . To emphasize the cell edges , the images were filtered with an unsharp mask ( implemented by the image-processing software MetaMorph [Universal Imaging , Sunnyvale , CA] ) , which subtracts a low-pass filtered and scaled image from its original image . The Gradient Anisotropic Diffusion filter [29] , [30] was then applied to smooth edge boundaries for complex cell shapes . After the filtering step , the intracellular and extracellular regions were segmented using the global threshold determined for the first frame . The cell boundary was extracted directly from the outline of the thresholded images . Typically , the extracted cell boundaries were distorted when edge extraction was applied to threshold regions with one-pixel width , such as thin spikes . To avoid this , each pixel in a thresholded image was divided into sub-pixels before extraction of boundaries . Boundary extraction was then executed for each binary image at a sub-pixel resolution . We did not apply spline fitting in EET or polar coordinate-based analysis to avoid spoiling steep edge structures with filopodium-like thin shapes . Area differences were also extracted from the thresholded images . Increased areas were determined by subtracting the current frame from its next frame , while decreased areas were determined by subtracting the next frame from the current one . Most of these procedures , including EET and cross-correlation analysis , were implemented by Matlab ( The MathWorks , Natick , MA ) . For this study , we used neurite outgrowth of rat pheochromocytoma PC12 cells as an example of cells displaying complex morphological dynamics , while random migration of human fibrosarcoma HT1080 cells was used for analysis of the cross-correlation between morphological changes and Rho GTPase activity . PC12 cells were plated on polyethyleneimine- and laminin-coated 35-mm glass-base dishes ( Asahi Techno Glass , Chiba , Japan ) , and then transfected with pRaichu-1011x encoding Rac1 FRET probe . One day after transfection , the cells were stimulated with 50 ng/ml NGF in phenol red-free Dulbecco's modified Eagle's medium/F12 containing 0 . 1% bovine serum albumin for 48 h to induce neurite outgrowth . HT1080 cells were transfected with pRaichu-1011x , pRaichu-1054x encoding a Cdc42 FRET probe , or pRaichu-1294x encoding RhoA FRET probe and , after 24 h , cells were plated on collagen-coated 35-mm glass-base dishes . The medium was then changed to phenol red-free Dulbecco's modified Eagle's medium/F12 containing 10% fetal bovine serum , overlaid with mineral oil to prevent evaporation , and image acquisition was started . The cells were imaged with an inverted microscope ( IX81 or IX71; Olympus , Tokyo , Japan ) equipped with a cooled charge-coupled device camera ( Cool SNAP-K4 or Cool SNAP-HQ; Roper Scientific , Duluth , GA ) , and a laser-based auto-focusing system at 37°C . The filters used for the dual-emission imaging were purchased from Omega Optical ( Brattleboro , VT ) : an XF1071 ( 440AF21 ) excitation filter , an XF2034 ( 455DRLP ) dichroic mirror , and two emission filters ( XF3075 [480AF30] for CFP and XF3079 [535AF26] for FRET ) . The cells were illuminated with a 75-W xenon lamp through a 6% , 10% or 12% ND filter and viewed through a 60× oil-immersion objective lens ( PlanApo 60×/1 . 4 ) . The exposure times for 2×2 or 3×3 binning were 400 or 500 ms for CFP and FRET images . After background subtraction , FRET/CFP ratio images were created with MetaMorph software , and the images were used to represent FRET efficiency . Further details of microscopy and sample preparation can be found in previous reports [26] , [27] . We executed a permutation test between positive ( 6 min ) , negative ( −6 min ) and non time-shifted correlations according to the following procedure . Letters/numbers in bold fonts represent vectors . For example , if we have CP = [0 . 6 0 . 4 0 . 6] and CN = [0 . 3 0 . 5 0 . 4] , then D_P_N = [0 . 3 −0 . 1 0 . 2] . Each permutated difference vector is an element of the set of possible ones: D_P_Nper∈{[0 . 3 0 . 1 0 . 2] , [0 . 3 0 . 1 0 . 2] , [0 . 3 0 . 1 −0 . 2] , [−0 . 3 0 . 1 0 . 2] , [−0 . 3 −0 . 1 0 . 2] , [−0 . 3 0 . 1 −0 . 2] , [0 . 3 −0 . 1 −0 . 2] , [−0 . 3 −0 . 1 −0 . 2]} . Owing to the independence assumption of a sign-change between elements , the permutation ( null ) distribution is simply obtained by arranging all the possible sequences , whose number is 23 = 8 in the above example , with uniform probability . The permuted difference vectors whose mean is larger than that of the original difference vector are thus {[0 . 3 0 . 1 0 . 2] , [0 . 3 −0 . 1 0 . 2]} and number 2 . In this particular example , the permutation p-value is then given as p = 2/8 = 0 . 25 . If this p-value is smaller than a specified significance level ( usually 5% ) , the difference between CP and CN is said to be significant . In the case of two-sided permutation test , the significance level is simply divided by 2 .
We applied EET to branching PC12 cells to validate its usefulness for quantifying complex cell morphological changes . As shown in Figure 4A , the PC12 cells extended their neurites with branches after treatment with NGF . A time-lapse series ( 1-min intervals ) of the images was trimmed to help maintain visual correspondence with EET profiles because large image sizes may make the visual inspection difficult . We chose the branching region to verify the utility of EET for the complex cell shape . Next , following the EET procedure , we determined the profiles of edge boundary states , as depicted in Figure 4C , in which red , blue and green colors denote protrusive , retractile and pausing states of the cell edge boundary , respectively . Black lines connect the anchor points ( see Materials and Methods ) , and represent the corresponding segments and subdivided regions . Small fragments of the segments show spatially independent and transient behaviors of the edge evolution and contraction , while long segments represent simultaneous occurrence of edge evolution and contraction in neighboring regions during the time lapse . We also monitored global changes in cell morphology using total area and complexity ( bottom of Figure 4C ) , together with the state profiles , because the state profile by itself does not illustrate the global characteristics of cellular morphodynamics . The monitored total areas and complexity represent the balance between the length of the cell edge boundary and the total area . These values will help to identify rough images of morphological changes . To visualize the dynamics of local area differences by EET , an area difference map was constructed as shown in Figure 4D . Despite the complex morphological changes , EET was successful in quantifying detailed local area changes and preserving the positional correspondence among the subdivided edges . For example , the white squared area in Figure 4A showed a slight extension until 20 min and then retraction between 30–50 min; this corresponds to the region in the state profile starting from 60–80 mm ( ordinate ) at 0 min ( abscissa ) ( Figure 4C ) . This quantification and visualization method reduces the difficulty in dealing with time-lapse image data by summarizing the morphodynamic characteristics into two-dimensional state profiles . Because previous studies have shown the localization of GTPase activities at peripheral regions [26] , we applied EET to motile HT1080 cells to further quantify the relationship between local morphological changes and local GTPase activity . First , we imaged motile HT1080 cells with a 1-min time-lapse . Figure 5A shows a series of FRET/CFP ratio images of a single motile HT1080 cell expressing Raichu-1011x ( Rac1 probe ) , and the FRET efficiency is shown in pseudo colors . Based on a previous study indicating a correlation between FRET efficiency and Rac1 activity [26] , we assumed that the Rac1 activity should be well represented by the FRET efficiency . The time-lapse images reveal the wandering behavior of the HT1080 cell and a spatio-temporal activity pattern of Rac1 within the cell . To emphasize the protruded and retracted areas in consecutive frames , each image was first transformed into a binary image by extraction of the cell and background regions . The consecutive subtracted images were then obtained frame by frame , and the protrusion and retraction areas were colored in red and blue , respectively ( Figure 5B ) . As reported previously , the coincidence of morphological changes with increases in Rac1 activity was seen by comparing the FRET and subtracted binary images ( Figure 5A and 5B ) . Next , we applied EET to precisely examine the spatio-temporal relationships between morphological changes and GTPase activities in motile HT1080 cells . As with PC12 cells ( Figure 4C and 4D ) , the state profile and local area difference map were acquired ( Figure 5C and 5D ) . Simultaneously , we acquired the local activity map ( Figure 5F ) based on segment-wise local activity ( Figure 5E , Materials and Methods ) . This time-position map of the local GTPase activity corresponds to both the state profile and the local area difference map ( Figure 5C , 5D , and 5F ) . There appeared to be similar patterns between the local area difference map and the local activity map . The area difference map revealed chunks of persistently protruding or retracting regions at the cell periphery , while the activity profile revealed spatially and temporally associated activity patches at the cell boundary , suggesting that their dynamics correlated with each other . Visual inspection of the local area difference map ( Figure 5D ) and local activity map ( Figure 5F ) helped us to detect patterns of cell morphology and GTPase activity . The upper left area of Figure 5D shows that formation of large lamellipodia ( between 6–20 min ) was preceded by the local retraction of the cell edge , and this retraction-extension pattern was also identified in other cell types ( data not shown ) . Cell edge retraction has the potential to induce tension-dependent development of molecular activities involving Rho GTPase signaling [31] . Our data are consistent with this mechanosensory function and provides a possible mechanism for interactions between morphological changes and molecular signaling . On the other hand , the large retraction between 12 and 18 min ( Figure 5D ) was preceded by a local decrease in Rac1 activity ( blue zone in Figure 5F at 10–12 min ) and similar patterns were also observed in other cells ( data not shown ) . Potentially , the local decline in Rac1 activity may contribute to the subsequent cell-edge retraction . In addition , in contrast with the morphological changes , the local activity map revealed that the GTPase activity changed moderately at the same position . This moderate change may help maintain the stability of the polarity . We further investigated this spatio-temporal cross-correlation between morphological changes and Rho-family GTPase activity . First , we summarized their statistical characteristics to examine the cross-correlation . Figure 6A shows a scatter plot of the local activity and the local area difference for all identified segments . Because there were no non-linear relationships in this plot , we considered that common statistical analyses could be applied to these data . Next , we examined the histograms of the activity and area difference and found that the activities had a Gaussian distribution ( Figure 6B ) ; heavy tails were observed in some samples , but not in the area differences ( Figure 6C ) . Although the activity histograms of a few samples exhibited one or two minor peaks in addition to the major peaks ( data not shown ) , we assumed that they could still be approximated by Gaussian distribution for simplicity; in subsequent analyses , we used both Pearson's product-moment correlation coefficient and Spearman's rank correlation to confirm the cross-correlation data . We next examined the effects of time-shifts on cross-correlation between the activity and the area difference . The graphical structures of EET profiles display local area differences in the corresponding time-shifted segments . The middle panels of Figure 6D show time-shifted local area difference maps with various time-shift values . Different patterns appeared on the area difference map depending on the time-shifts , showing that the correlation changes depend on the time-shift values . The scatter plots of activity without time-shift against time-shifted area differences show a linear relationship for negative values of the time-shift ( Figure 6D upper ) . We calculated time-shifted cross-correlations between the local activities of Cdc42/Rac1/RhoA and local morphological changes , as shown in Figure 7 . As expected , there were strong correlations between Cdc42/Rac1 activities and morphological changes , but the peaks of the correlation coefficients were slightly time-shifted . Moreover , and surprisingly , the peaks indicated that the local morphological changes preceded changes in local activity , which can be seen in Figure 6D . We confirmed statistical significance of the difference between negative ( −6 min ) , zero and positive ( +6 min ) time-shifts by performing permutation tests ( see Table S1 ) . The number of samples used to calculate the cross-correlations was sufficiently large ( see Figure S2 and Figure S3 ) . Although there are some conspicuous morphological events seen in the EET profile ( Figure 5C ) , such as the protrusion around 6–16 min and the retraction around 12–18 min , the cross-correlation based on the EET analysis was designed to be robust against such local events arising in limited sites in the cell . In this specific case of Rac1 activity in HT1080 cell , our finding that the cross-correlation profile is highly correlated with minus time-shift values is unchangeable , even when these conspicuous morphological events are replaced by normal morphological events ( see Figure S4 ) . Note that the Spearman's rank correlation also reduces the bias effect of large values ( events ) on statistical values . The results do not appear to be intuitive with regard to the causal relationship between morphological changes and molecular signaling; upstream molecular signaling should control downstream morphological changes , for example via actin reorganization , adhesion and/or retrograde flow . In the cases of both Rac1 and Cdc42 , the time-shifted correlations showed that morphological change preceded local GTPase activity . Cdc42 activity , in particular , showed large deviations when the preceding time-shifts were short , and the correlation decayed steeply when the time-shifts were longer . Rac1 activity , on the other hand , elicited small deviations and the decay of the correlation was less steep when the preceding time-shifts were longer . It should be noted that the time-shifted correlation generally approaches zero over long time-shifts owing to an increase in the number of connections between the original segment and time-shifted segments ( see Figure 3E ) . This reflects a weakened relationship , i . e . , not one-to-one but one-to-multi relationship between the original region ( segment ) and its time-shifted regions ( connected segments ) . However , this weakened relationship does not imply a decrease in the reliability of calculations of time-shifted coefficients by making vague relationships between time-shifted segments , but instead represents the natural dilution of the correspondence between an original region and its time-shifted regions . We further examined the spatial property of the relationship between GTPase activity and morphology change by comparing the original EET profile with rotated ( see Figure S5A ) and permutated ( see Figure S5B ) segments of EET profiles . EET profiles of rotated segments showed a decreased correlation with increased rotation ( see Figure S5C ) . Because the segments have a range of lengths along the cell edge , EET did not directly show an exact proximity . However , it showed the significance of the locality of morphodynamic regulation signal . The signal locality dependency was also shown by a lack of correlation of the permutated segments profile with EET . We compared EET analysis to polar coordinate-based analysis to further prove the utility of EET . We first performed polar coordinate-based analysis to the cell in Figure 5 for direct comparison with EET ( Figure 8A ) . The polar coordinate-based analysis produced time-position maps of local activities and local morphological changes that were similar to the activity map and area difference map of EET ( see Figure S6 ) . As for EET , local activity was determined as a mean value within an ROI , which was a circle of radius r . We used the same r value for EET and the polar coordinate-based method . Both analyses produced similar maps ( see Figure S6 ) , and time-shifted cross-correlations were then calculated ( Figure 8B ) . Both of the time-shifted cross-correlations showed similar patterns for the timing between local morphological changes and GTPase activities ( Rac1 ) , i . e . , a high correlation with negative time-shifts and a low correlation with positive time-shifts . However , the EET analysis showed a higher correlation than that with the polar coordinate-based analysis at the time-shifts of −3 to −20 min . A similar tendency was observed when a population of the cells in Figure 7 was analyzed by the polar coordinate-based method ( Figure 8C ) . The averaged peaks of cross-correlations obtained by the polar coordinate-based analysis were substantially lower than those obtained with EET , particularly for Cdc42 and Rac1 ( Figures 7 and 8C ) . Permutation tests revealed significant differences between the time-shifted cross-correlations by the polar coordinate-based analysis ( see Table S1 ) . This might be due to the relatively large correlation values at the time-shift of zero . However , the variances were small , and the correlations prominently decreased when the time-shift value was far from zero . Statistical tests generally showed significant differences between two groups when the variance of each group was small . Here , the small variances in the correlations are likely to be obtained by averaging a large number of samples with small values , and the small values may be due to inconsistency in position alignment between different frames . Note that the polar coordinate-based analysis acquired a large number of samples at 1-degree intervals ( i . e . , 360 samples in each image ) from a single cellular edge and that adjacent samples were likely to have similar values because of physical edge continuity . Our EET implemented the sensitivity to detect correlations between activities and morphological changes by maintaining a consistent position between consecutive frames in terms of segments . Thus , we believe that the correlation peak at the time-shift of zero , obtained by the polar coordinate-based analysis , could be an artifact stemming from position misalignment . We also compared EET analysis with simple implementation of marker-tracking-based analysis . In this marker-tracking-based analysis , virtually defined markers were aligned uniformly along the spline-fitted cellular edge in the first frame of time-lapse FRET images . Then , the movements of markers in the direction perpendicular to the cellular edge during a single time-frame were measured according to the current marker position and the intersection of the perpendicular axes of the current cellular edge and the next cellular edge ( Figure 9A and 9B ) . Figure 9A and 9B show time-lapse cellular edges of the same cell as in Figure 5 , colored from blue ( 6 min ) to red ( 11 min ) , with virtually defined markers ( black dots ) and movements of the markers ( black lines ) . Topological violations of the markers ( crossing the black lines ) are indicated in Figure 9B , which is probably due to the highly complex morphological changes in the edges . Such complex changes could affect the marker movement maps ( although the map obtained by the marker-tracking-based method was comparable to that obtained by EET and by polar coordinate-based analysis; see Figure S6 ) , but our statistical analysis was not affected . Instead , the changes in marker distribution from a uniform ( black dots on the blue line in Figure 9B ) to a non-uniform alignment ( black dots on the red line in Figure 9B ) would have non-negligible influences on the time-shifted statistical analysis ( e . g . , Figure 1C ) . As with EET and the polar coordinate-based method , the local activity was determined as a mean value within an ROI , which was a circle of radius r . We used the same r value in EET , the polar coordinate-based and the marker-tracking-based methods . All analyses produced similar maps ( see Figure S6 ) , and time-shifted cross-correlations were then calculated ( Figure 9C ) . The time-shifted cross-correlations in Figure 9C show lower correlations at the negative time-shifts compared with EET . The marker-tracking-based analysis produced similar patterns of time-shifted cross-correlations for Cdc42 and Rac1 ( Figure 9D ) and the permutation tests revealed significant differences between the correlations at zero , and the negative and positive time-shifts ( see Table S1 ) . Similar to the polar coordinate-based method , the marker-tracking-based analysis revealed weaker characteristics in the time-shifted cross-correlations . This seems to result from biased sampling by the non-uniform marker distribution caused by morphological changes , which can be seen in Figure 9B . Thus , we suggest that the marker-tracking-based analysis has undesired affects on the statistical analysis , particularly when the cellular edge has a persistent deforming property .
We have developed an algorithm called EET , which describes changes in cell morphology using time-lapse live cell imaging . Spatio-temporal area difference maps revealed morphodynamic properties as patterns of extension and retraction , and the correspondence between time-shifted segments , achieved using anchor points , ensured that the related subdivided edges were connected between time-shifted frames . Therefore , EET effectively accounts for complex morphodynamics that include persistent extension or retraction , and arborization . This property is realized by the graphical representation of edge evolution , and ensures EET is suitable for depicting changes in cell shape , such as the branching that occurs during neural development . Application of EET to the extending neurites of PC12 cells provided a clear evidence of its utility by precisely revealing the persistent protrusion and retraction patterns . Besides , a second application to motile HT1080 cells illuminated distributions of local area differences and corresponding local activity of GTPases . Although the graph structure itself potentially generates biases when correlating the one-to-multi segments between temporally distant frames , we confirmed that our results were consistent even when we obtained our result differently by associating the area change in each segment with the average molecular activities over the corresponding segments ( see Figure S7 ) . Because cellular morphological changes have probabilistic characteristics [32] , the statistical analysis approach used here is a powerful tool for exploring the nature of dynamic processes in cellular behaviors . It has been established that Rho-family GTPases ( Rac1 , Cdc42 and RhoA ) play key roles in morphological changes through cytoskeletal reorganization [19] , [33]–[35] . Furthermore , previous FRET imaging studies have shown that these GTPases are exquisitely regulated spatio-temporally [25] , [26] , [28] , [36] . In this study , we obtained additional results with EET analysis . In particular , the activities of Rac1 and Cdc42 were localized around the peripheral regions and strongly correlated with the preceding changes in the local area , while the local activity of RhoA was only weakly correlated with changes in the local area . The activity of Cdc42 immediately preceding to the activity of Rac1 is consistent with earlier finding , suggesting that Rac1 is activated by active Cdc42 [37] , while the difference in time-shifted cross-correlations between RhoA and Cdc42/Rac1 ( Figure 7 ) would supports the existence of feedback loops common to Rac1 and Cdc42 . However , the relationship between RhoA activity and morphology remains controversial [25] , [38] . Quantitative analyses in different experimental conditions will clarify this issue . Our results , however , should prompt further investigation of the role of GTPase in regulation of morphodynamics , because this challenges the hypothesis that Rac1 and Cdc42 promote extension of lamellipodia or filopodia , respectively . The precise mechanism by which local area changes precede local activity around the cell boundary remains unclear from our current analysis . However , we speculate four possible mechanisms based on our results . The first explanation is the existence of upstream signaling molecules that regulate extension in parallel with GTPase activity . If the reactions of the signaling cascades involved with extension are faster than those linked to GTPase activation , extension could precede GTPase activity . In this respect , it would be interesting to conduct a study similar to the current one for PI3K , which activates many signaling molecules including Rac1 activators [39] . The second explanation is that protrusion site-specific stimulation activates the GTPases . There are several mechanisms by which physical force can be converted into biochemical responses [40] , and a theoretical study has suggested that signaling activity might be affected by cell shape [41] . In addition , we have shown that there is a positive feedback loop from actin polymerization to Rac1/Cdc42 activation via PI3K [39] . Therefore , it is possible that the detected increase in Rac1/Cdc42 activation was , in fact , secondary to actin polymerization at the protruding regions . The third possibility is that signaling crosstalk regulates the timing of extension and retraction [42] . If the GTPase activity induces extension and also activates factors that promote edge retraction , the peak GTPase activity appears to be delayed with morphological changes by balancing with activated retraction promoter . The fourth possibility is the existence of different mechanisms for cell edge extension . EGF-stimulated initial protrusion in MTLn3 rat adenocarcinoma cells is caused by cofilin activation and severing of F-actin , which is coincident with actin polymerization and formation of lamellipodia [43] . On the other hand , Rac1-dependent edge expansion is followed by stabilization of the protrusions [44] . Further investigations will enable us to determine which hypothesis ( including coexistence ) is most likely with the observed phenomena . In addition , the effects of the dynamics in the perpendicular axis such as changes in cell thickness and volume should be determines , because our results are restricted to the horizontal dynamics . Probe-related mechanisms should also be considered carefully; for example , the difference in the expression levels between the FRET probes and the endogenous Rho GTPases might affect the timing and dynamics of activation of GTPases . Quantitative analysis of live cell microscopy images is invaluable for better understanding of the dynamic properties of processes such as chemotaxis and development . Such quantitative data can go beyond descriptions of the dynamic features of cellular behavior to serve as a scaffold for theoretical study and to enhance system-level understanding . Based on quantitative data acquired by polar coordinate-based analysis of neurons , for example , Betz et al . discussed a bistable stochastic process derived from velocity histograms and calculated potential distribution [32] . Therefore , connecting modeling studies with quantitative experimental studies has the potential to yield breakthroughs in system-level understanding of cellular functions [45]–[47] . The EET method allows us to quantify details of morphological dynamics of cells . Moreover , it also enables to investigate the spatio-temporal relationship between morphological dynamics and local molecular signaling dynamics . Further application of EET to other signals , e . g . , different species of GTPases such as Ras and upstream signals of Rho GTPases such as PI3K , and also to localization of actin should shed light on some of the dynamic and complex properties of regulation of the morphological/migratory systems in cells .
|
Morphological change is a key indicator of various cellular functions such as migration and construction of specific structures . Time-lapse image microscopy permits the visualization of changes in morphology and spatio-temporal protein activity related to dynamic cellular functions . However , an unsolved problem is the development of an automated analytical method to handle the vast amount of associated image data . This article describes a novel approach for analysis of time-lapse microscopy data . We automated the quantification of morphological change and cell edge protein activity and then performed statistical analysis to explore the relationship between local morphological change and spatio-temporal protein activity . Our results reveal that morphological change precedes specific protein activity by 6–8 min , which prompts a new hypothesis for cellular morphodynamics regulated by molecular signaling . Use of our method thus allows for detailed analysis of time-lapse images emphasizing the value of computer-assisted high-throughput analysis for time-lapse microscopy images and statistical analysis of morphological properties .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"developmental",
"biology/morphogenesis",
"and",
"cell",
"biology",
"cell",
"biology/cell",
"signaling",
"molecular",
"biology/bioinformatics",
"biotechnology/bioengineering",
"computational",
"biology/molecular",
"dynamics",
"cell",
"biology/cytoskeleton"
] |
2008
|
Quantification of Local Morphodynamics and Local GTPase Activity by Edge Evolution Tracking
|
Peroxisome proliferator-activated receptor ( PPAR ) γ is a global transcriptional regulator associated with anti-inflammatory actions . It is highly expressed in alveolar macrophages ( AMs ) , which are unable to clear the intracellular pathogen Mycobacterium tuberculosis ( M . tb ) . Although M . tb infection induces PPARγ in human macrophages , which contributes to M . tb growth , the mechanisms underlying this are largely unknown . We undertook NanoString gene expression analysis to identify novel PPARγ effectors that condition macrophages to be more susceptible to M . tb infection . This revealed several genes that are differentially regulated in response to PPARγ silencing during M . tb infection , including the Bcl-2 family members Bax ( pro-apoptotic ) and Mcl-1 ( pro-survival ) . Apoptosis is an important defense mechanism that prevents the growth of intracellular microbes , including M . tb , but is limited by virulent M . tb . This suggested that M . tb differentially regulates Mcl-1 and Bax expression through PPARγ to limit apoptosis . In support of this , gene and protein expression analysis revealed that Mcl-1 expression is driven by PPARγ during M . tb infection in human macrophages . Further , 15-lipoxygenase ( 15-LOX ) is critical for PPARγ activity and Mcl-1 expression . We also determined that PPARγ and 15-LOX regulate macrophage apoptosis during M . tb infection , and that pre-clinical therapeutics that inhibit Mcl-1 activity significantly limit M . tb intracellular growth in both human macrophages and an in vitro TB granuloma model . In conclusion , identification of the novel PPARγ effector Mcl-1 has determined PPARγ and 15-LOX are critical regulators of apoptosis during M . tb infection and new potential targets for host-directed therapy for M . tb .
Nuclear receptors are a large family of structurally conserved , ligand activated transcription factors , which have a range of functions related to development , homeostasis , metabolism and immunity . Nuclear receptors include receptors for fatty acids such as peroxisome proliferator-activated receptors ( PPARs ) [1] . PPARs regulate expression of genes involved in fatty acid metabolism and inflammation ( pro- and anti- ) and are implicated in diabetes , cancer , and infectious diseases , including tuberculosis ( TB ) [2–6] . Drugs targeting PPARs and other nuclear receptors account for 13% of drugs approved for sale in the US and generated $27 billion in sales in 2009 [7] , highlighting their important impact on human health . There are three PPAR isoforms , α , β/δ , and γ , which are differentially distributed and activated by different ligands . PPARγ is highly expressed in AMs and is important for AM differentiation [8] . PPARγ agonists include oxLDL-derivatives , 15-deoxy-Δ12 , 14-prostaglandin J2 ( 15d-PGJ2 ) , 13-hydroxyoctadecadienoic acid ( 13-HODE ) , 15-hydroxyeicosatetraenoic acid ( 15-HETE ) , and the synthetic thiazolidinediones ( TZDs ) , which are used to treat type 2 diabetes mellitus [9] . PPARs regulate gene expression through multiple mechanisms , including heterodimerizing with the nuclear receptor retinoid X receptor ( RXR ) , and binding to PPAR response elements ( PPREs ) in the promoter to regulate gene expression [10] . TB is a global threat and leading cause of death worldwide [11] . The increasing incidence of multidrug-resistant ( MDR ) and extensively drug-resistant ( XDR ) TB highlights the need for new therapies . There has been an increasing drive towards host-directed therapies ( HDTs , [12] ) , but better understanding of how M . tuberculosis ( M . tb ) interacts with the human host is required for successful design of HDTs [13] . We have demonstrated that M . tb , and the M . tb cell wall component mannosylated lipoarabinomannan ( ManLAM ) enhance PPARγ activity through the mannose receptor ( MR ) in human macrophages , the host cell niche [4] . PPARγ is critical for M . tb survival , since knockdown or inhibition of PPARγ reduces M . tb growth in macrophages [4 , 5] . Although it is clear that PPARγ contributes to M . tb survival in human macrophages , the mechanism ( s ) behind this are incompletely understood . Apoptosis of infected cells is one important host defense mechanism that prevents the growth of intracellular bacteria and viruses , including M . tb [14] . Induction of apoptosis is linked to mycobacteria virulence ( with less virulent mycobacteria inducing more apoptosis ) [15] , and induction of macrophage apoptosis leads to reduced M . tb growth and increased mouse survival [16–18] . However , M . tb limits apoptosis through mechanisms that are not well understood [19 , 20] . Host factors that regulate apoptosis include members of the Bcl-2 protein family , which consists of pro-survival ( e . g . Bcl-2 and Mcl-1 ) , pro-apoptotic effector ( e . g . Bax and Bak ) , and initiator proteins . Bax and Bak activation is tightly regulated by the initiator and pro-survival proteins , including Mcl-1 [21] . Mcl-1 plays a critical role in regulation of apoptosis and , as such , its activity is tightly regulated at the transcriptional , post-transcriptional , and post-translational levels [22] . Here , we identify new PPARγ effectors through NanoString gene expression analysis , which led to the novel observation of altered expression of genes involved in apoptosis , including the ones encoding Bcl-2 proteins Mcl-1 and Bax . We provide the first evidence that PPARγ regulates expression of Mcl-1 and that this occurs during M . tb infection and requires 15-lipoxygenase ( 15-LOX ) activity . Further , that PPARγ and 15-LOX are important for M . tb-mediated limiting of apoptosis . Finally , we determined that pre-clinical therapeutics that target Mcl-1 significantly reduce M . tb growth in human macrophages and an in vitro granuloma model . Thus , herein we identify a novel PAPRγ effector that is a promising HDT target .
PPARγ knockdown or inhibition significantly reduces M . tb growth in macrophages [4 , 5] through unclear mechanisms . Since PPARγ is associated with anti-inflammatory actions [1 , 2 , 9] , we hypothesized that PPARγ down-regulates a host protective antimicrobial response to facilitate M . tb growth . To identify novel PPARγ effectors that facilitate M . tb growth in macrophages , we undertook expression profiling of genes involved in the immune response using the NanoString nCounter GX Human Immunology Panel . Monocyte-derived macrophages ( MDMs ) were transfected with scrambled control or PPARγ siRNA ( 81 . 7 ± 5 . 5% PPARγ knockdown , Fig 1A ) , then infected with M . tb for 6 and 24 h before cells were lysed and RNA collected . This approach identified multiple immunology-related genes that are significantly changed following PPARγ knockdown and M . tb infection . At 6 h post infection , 4 genes were significantly increased , and 7 decreased ( Fig 1B ) . More genes were altered at 24 h , with 31 genes significantly increased and 36 genes significantly decreased following PPARγ knockdown ( Fig 1C ) . This included IL-8 , which we had previously shown to be regulated by PPARγ during M . tb infection [4] . Other genes altered following PPARγ knockdown include those involved in IL-1 signaling: IL-1β , IL-1RAP , IL-1α ( Fig 1B and 1C ) . Genes involved in cell death were also differentially expressed following PPARγ knockdown , a conclusion that was corroborated with STRING analysis ( S1 Fig ) . In particular , the pro-apoptotic Bax was increased 1 . 51-fold and the anti-apoptotic Mcl-1 was reduced 1 . 65-fold at 24 h . These results suggested that M . tb induces Mcl-1 expression and represses Bax expression through PPARγ , to limit apoptosis and enhance M . tb infection . In support of this , Mcl-1 is induced following M . tb infection , Mcl-1 is critical for M . tb limiting of apoptosis and M . tb survival in human macrophages [17 , 23] , and polymorphisms in the Mcl-1 promoter are linked to TB risk [24] . In contrast , M . tb infection is correlated with reduced Bax expression [25] . It is unknown if PPARγ regulates Mcl-1 or Bax expression in immune cells . It is also unknown if PPARγ contributes to expression of any of the Bcl-2 family proteins during M . tb infection . Our NanoString results indicate that M . tb , which is well established to limit apoptosis [19] , may do so through PPARγ-mediated repression of the pro-apoptotic Bax and induction of the anti-apoptotic Mcl-1 , novel findings for PPARγ . We validated the Bax and Mcl-1 NanoString results with qRT-PCR , and confirmed that PPARγ knockdown led to a significant increase in Bax and decrease in Mcl-1 gene expression during M . tb infection . This was observed as early as 6 h after M . tb infection for both Bax and Mcl-1 ( Fig 1D and 1E ) . Besides Mcl-1 , M . tb induces expression of other anti-apoptotic Bcl-2 family members including Bcl-2 and Bcl-xL [26 , 27] . Since PPARγ induces expression of the anti-apoptotic Bcl-2 in cardiomyocytes and neurons [28 , 29] , we next determined whether PPARγ regulates expression of Bcl-2 or Bcl-xL in human macrophages during M . tb infection , similarly to Mcl-1 . The NanoString results indicated that Bcl-2 expression does not significantly change after PPARγ knockdown ( mean 1 . 03- and 1 . 26-fold change at 6 and 24 h , respectively ) . We confirmed this with qRT-PCR , and determined if Bcl-xL is regulated by PPARγ ( this gene was not included in the NanoString ) . Of the three genes , Mcl-1 exhibited the strongest increase with M . tb infection ( S2 Fig ) . Unlike Mcl-1 , Bcl-2 and Bcl-xL expression was not altered following PPARγ knockdown ( S2 Fig ) . These data indicate that PPARγ specifically regulates Mcl-1 expression during M . tb infection . Mcl-1 plays a critical role in regulation of apoptosis , and due to its short protein half-life ( 30 min ) , is tightly regulated at the transcriptional level [22] . In contrast , Bax is highly regulated through post-translational modifications , and exerts redundant activities with Bak [21] . Thus , we chose to focus on PPARγ regulation of Mcl-1 . We determined if the PPARγ agonist rosiglitazone is sufficient to induce Mcl-1 expression in MDMs and found that rosiglitazone significantly increased Mcl-1 gene expression ( Fig 2A ) . The effect of rosiglitazone was significantly reduced in a dose-dependent manner if MDMs were treated with the PPARγ inhibitor GW9662 ( Fig 2B ) , confirming that rosiglitazone induces Mcl-1 expression through PPARγ and not through off-target effects . Promoter analysis indicated that the Mcl-1 promoter contains six putative PPREs , indicating that PPARγ may directly regulate Mcl-1 expression . To confirm this , Mcl-1 promoter reporter assays were undertaken in RAW264 . 7 cells , which express very low levels of PPARγ ( Fig 2C; [30] ) . RAW cells were transfected with a luciferase reporter construct containing the entire Mcl-1 promoter region , with or without PPARγ expression plasmids , then treated with rosiglitazone . Expression of PPARγ resulted in a significant increase in luciferase activity ( Fig 2D ) . To confirm that this is mediated through PPARγ , we mutated the six putative PPREs in the Mcl-1 promoter region ( Fig 2E ) , and repeated the promoter assays . Mutation of any of the six putative PPREs substantially reduced promoter activity , to levels comparable to cells that do not express PPARγ ( Fig 2F ) . This was further reduced if all six PPREs were mutated . In contrast , mutations in any of the six PPREs had no effect on NFκB-driven Mcl-1 expression following LPS treatment ( Fig 2G ) . Together , these data provide evidence that PPARγ enhances expression of Mcl-1 , and identifies binding sites for PPARγ in the Mcl-1 promoter . We determined the kinetics of Mcl-1 gene expression in human macrophages during M . tb infection . We found that M . tb significantly increased Mcl-1 gene expression in MDMs as early as 6 h post infection , and that this was maintained through at least 48 h ( Fig 3A ) . Similarly , M . tb infection of HAMs significantly increased Mcl-1 expression ( Fig 3B ) . We next determined the kinetics of Mcl-1 protein production in human macrophages during M . tb infection . Similar to gene expression , Mcl-1 protein was induced as early as 6 h after infection with M . tb , and was maintained for at least 72 h ( Fig 3C and 3D ) . These kinetics are similar to previous reports in human and murine macrophages [17 , 23 , 31] . M . tb at different MOIs ( 5 or 10 ) significantly increased Mcl-1 protein 24 h after infection ( Fig 3E and 3F ) . M . tb also significantly induced Mcl-1 protein in HAMs ( Fig 3G and 3H ) . These data provide the first indication that M . tb induces Mcl-1 expression in HAMs , which are an important niche for M . tb during infection . We next determined if PPARγ-regulation of Mcl-1 gene expression ( Fig 1 ) would correspond to altered protein levels . MDMs were transfected with PPARγ and scrambled siRNA , then infected with M . tb for 24 h and protein lysates assessed . As expected , M . tb induction of Mcl-1 protein production was significantly reduced following PPARγ knockdown ( Fig 4A and 4B ) . The endogenous ligand for PPARγ during M . tb infection is unknown . However , previous work by our laboratory indicated that the cytosolic phospholipase A2 cPLA2 is important for PPARγ activity during M . tb infection [4] . cPLA2 releases arachidonic acid ( AA ) from the plasma membrane , which can then be converted to the PPARγ agonist 15d-PGJ2 through COX-2 or the PPARγ agonists 13-HODE and 15-HETE through 15-lipoxygenase ( 15-LOX ) [9 , 32] , but it is unknown which of these pathways is critical for PPARγ activity in human macrophages . Since M . tb infection induces COX-2 production [4] and COX-2 has been linked to regulation of Mcl-1 production in human lung adenocarcinoma cells [33] , we first determined if COX-2 is important for PPARγ activity and the production of Mcl-1 during M . tb infection of human macrophages . MDMs were treated with the COX-2 specific inhibitors NS-398 and CAY10404 and then infected with M . tb . We found that COX-2 inhibition did not reduce Mcl-1 production during M . tb infection ( S3A and S3B Fig ) . COX-2 is required for PGE2 release , and both NS-398 and CAY10404 significantly reduced M . tb- and LPS-stimulated PGE2 release ( S3C and S3D Fig ) , confirming COX-2 inhibition . These results indicate that although M . tb induces COX-2 , this is not important for PPARγ-mediated Mcl-1 production during M . tb infection of human macrophages and supports the notion that although 15d-PGJ2 is a PPARγ agonist , 15d-PGJ2 concentrations inside the cell may not reach the levels required to activate PPARγ , a much raised topic in the field [9 , 34] . Previous work indicated that LOXs are important for PPARγ activation , likely through generation of the PPARγ agonists 13-HODE and 15-HETE [35] . We next determined if LOXs were important for PPARγ activity in human macrophages during M . tb infection . MDMs were treated with the general LOX inhibitor nordihydroguaiaretic acid ( NDGA ) , and then infected with M . tb . M . tb-induced Mcl-1 production was significantly inhibited by NDGA ( Fig 4C and 4D ) . Human macrophages produce two LOX isoforms , 5-LOX and 15-LOX [32 , 36] . To determine if 5- and/or 15-LOX were critical for PPARγ activity , MDMs were treated with the 5-LOX specific inhibitor zileutin and the 15-LOX specific inhibitors PD146176 and ML351 , then infected with M . tb . 5-LOX inhibition had no effect on Mcl-1 production ( Fig 4C and 4D ) while Mcl-1 production was significantly reduced following 15-LOX inhibition ( Fig 4E and 4F ) . Together , these data provide evidence that 15-LOX , not 5-LOX or COX-2 , is important for PPARγ-mediated Mcl-1 production in human macrophages during M . tb infection . To confirm this , we inhibited 15-LOX using PD146176 , then stimulated PPARγ by treating cells with the 15-LOX product 13-HODE and probed for Mcl-1 production during M . tb infection . As expected , we found that 13-HODE ameliorated the reduction of Mcl-1 production seen when 15-LOX is inhibited ( Fig 4G and 4H ) . Taken together , these data provide evidence that 15-LOX , through production of PPARγ ligands , is important for Mcl-1 production in human macrophages during M . tb infection . Knockdown of Mcl-1 significantly increases apoptosis during M . tb infection ( Figs 5A , 5B and S4A ) [17 , 23] . Since Mcl-1 is induced by PPARγ ( Figs 1 , 2 , and 4 ) , we hypothesized that PPARγ would also be important for M . tb limiting of apoptosis . To test this , MDMs were transfected with PPARγ or scrambled control siRNA , then infected with M . tb , and TUNEL labeling and CellTiter Glo Assays were performed to enumerate apoptotic cells . As a positive control , MDMs were stimulated with the known apoptosis inducer staurosporine [20] , which substantially increased MDM apoptosis ( Fig 5C , 5G and 5H ) . Supporting our hypothesis , PPARγ knockdown significantly increased apoptosis during M . tb infection ( Figs 5D , 5E , 5F and S4B ) . Enumeration of apoptotic cells indicated that > 85% of apoptotic cells were infected with M . tb . PPARγ inhibition with GW9662 similarly increased cell death during M . tb infection ( Fig 5G ) . These results indicate that PPARγ , likely through induction of Mcl-1 , contributes to M . tb limiting of apoptosis . We next determined if 15-LOX , which is critical for PPARγ activity and Mcl-1 expression , contributes to M . tb limiting of apoptosis . Indeed , we found that the 15-LOX specific inhibitor PD146176 significantly increased cell death during M . tb infection ( Fig 5H ) . Together , these results indicate that M . tb-induced Mcl-1 expression , through 15-LOX-dependent PPARγ activity , limits apoptosis in human macrophages . During M . tb infection , Mcl-1 knockdown increased apoptosis ( Fig 5 ) [17] , which is thought to limit M . tb infection [19] . Similar to previous reports [4 , 17 , 23] , we found that Mcl-1 knockdown significantly reduces M . tb survival in human macrophages ( S5A Fig ) . Thus , we hypothesized that Mcl-1 could serve as a target for HDT to limit M . tb growth . Drugs targeting Mcl-1 , and other Bcl-2 pro-survival proteins , are being developed as cancer therapies , and some of these inhibitors have advanced to clinical trials [37] . We queried whether these drugs could be repurposed for TB therapy , by limiting M . tb growth in macrophages , which has not been studied . MDMs were infected with M . tb , then treated with the Mcl-1 inhibitors sabutoclax ( which targets Mcl-1 and the other pro-survival Bcl-2 proteins ) , TW-37 ( which targets the pro-survival Bcl-2 proteins , and has higher affinity for Mcl-1 ) , A-1210477 and MIM-1 ( the latter two are specific for Mcl-1 ) at concentrations that induce apoptosis [38 , 39 , 40 , 41 , 42] . All of the tested Mcl-1 inhibitors significantly reduced M . tb survival in human macrophages , as determined by CFU enumeration ( Fig 6A ) . The general inhibitors sabutoclax and TW-37 were more potent than the Mcl-1 specific inhibitors , likely due to inhibition of Mcl-1 , Bcl-2 and Bcl-xL , which are also induced during M . tb infection [26 , 27] . At 30 μM , the Mcl-1 specific inhibitors A-1210477 and MIM-1 reduced M . tb growth by 89% and 72% , respectively ( Fig 6A ) . Although all of the Mcl-1 inhibitors induce apoptosis , the MDM monolayer remained intact for the time period examined , even at the highest concentration of inhibitor ( S5B Fig ) . We next determined the kinetics of M . tb growth inhibition for the Mcl-1 specific inhibitors , using a luciferase-expressing M . tb strain . We noted that by 4 d , both A-1210477 and MIM-1 substantially reduced M . tb growth with this readout also , and that A-1210477 , but not MIM-1 , inhibition of M . tb growth was maintained for at least 7 d ( Fig 6B ) . This is likely due to the binding affinity for Mcl-1 , which is in the nM range for A-1210477 and μM range for MIM-1 [37] . We confirmed that the Mcl-1 inhibitors target host responses , since none of the inhibitors reduced M . tb growth in the absence of macrophages ( S5C Fig ) . To determine if the Mcl-1 inhibitors limited M . tb growth through apoptosis of host cells , we treated cells with the potent caspase inhibitor Q-VD-OPH to inhibit apoptosis . We observed that caspase inhibition during A-1210477 treatment led to a partial restoration of M . tb growth ( 1 . 52 ± 0 . 285 fold increase , mean ± SEM , n = 2 , Fig 6C ) , consistent with the notion that Mcl-1 regulates M . tb growth through apoptosis . These promising results indicate the feasibility of Mcl-1 and other anti-apoptotic Bcl-2 proteins as viable targets for HDT for TB . A characteristic of TB is formation of granulomas containing macrophages , multinucleated giant cells , lymphocytes , and fibroblasts around M . tb [13] . These granulomas are thought to help the host contain the bacterium , but also provide a niche for M . tb that is recalcitrant to antibiotics [13] . We were interested in determining if the Mcl-1 inhibitors could enter the multicellular granuloma structures and limit M . tb growth in this microenvironment . To assay for this , we used a human in vitro granuloma model that we previously characterized , which contains macrophages , multinucleated giant cells , T cells , and B cells [43 , 44] and provides a unique model system for the study of drugs in a complex human granuloma . We found that the Mcl-1 inhibitors Sabutoclax and TW-37 , at concentrations that inhibited growth in human macrophages , significantly reduced M . tb growth in human in vitro granulomas ( Figs 6D and S5D ) , indicating that these drugs can penetrate the multicellular granuloma complex . This was observed at both 3 and 6 days after treatment ( Figs 6D and S5D ) . Similar to the MDMs , the broad-spectrum inhibitors Sabutoclax and TW-37 were most effective in the granulomas , and in contrast to the MDM results , inhibitors A-1210477 and MIM-1 did not significantly reduce M . tb growth in the granuloma structures . This may indicate that A-1210477 and MIM-1 did not efficiently enter the granuloma complexes ( and thus higher concentrations might be required than what we tested ) while Sabutoclax and TW-37 did , or that inhibition of the combination of Mcl-1 , Bcl-2 and Bcl-xL is required when the bacteria are in multicellular complexes . Either way , these promising results indicate the feasibility of Mcl-1 and other anti-apoptotic Bcl-2 proteins as viable targets for HDT for TB . Since 15-LOX is important for PPARγ activity and Mcl-1 expression , we wondered if 15-LOX would serve as an additional target for HDT for TB . Interestingly , we found that the 15-LOX specific inhibitor PD146176 significantly limited M . tb growth in human macrophages ( Fig 6E ) . These results indicate that targeting Mcl-1 expression and activity are both promising options for HDT for TB .
PPARγ is critical for M . tb intramacrophage growth [4 , 5] , yet the mechanisms behind this are incompletely understood ( rev in: [3] ) . PPARγ is important for lipid body formation , and limiting TNFα and IL-6 , and increasing IL-8 and IL-10 secretion during BCG and/or M . tb infection [4 , 5 , 45] . However , it is unclear what other pathways are regulated by PPARγ during M . tb infection , and their role in regulating M . tb infection . Gene expression analysis ( Fig 1 ) reveals several new and unexpected potential PPARγ effector proteins in regulating human macrophage responses during M . tb infection . We show that PPARγ selectively regulates expression of the anti-apoptotic factor Mcl-1 . We extend our previous work showing that cPLA2 is important for PPARγ activity [4] , by showing that the down-stream enzyme 15-LOX ( which contributes to production of the PPARγ agonists 13-HODE and 15-HETE ) , is important for PPARγ activity and Mcl-1 expression in human macrophages . Mcl-1 limits apoptosis , we show that PPARγ and 15-LOX are also important for M . tb limiting of apoptosis . Treating macrophages with 15-LOX or Mcl-1 inhibitors significantly reduces M . tb growth in macrophages and our data showing partial restoration of this phenotype with caspase inhibition is consistent with the notion that Mcl-1 regulates M . tb growth though apoptosis , although it is possible that other mechanisms also contribute ( Fig 7 ) . Different virulent M . tb strains ( H37Rv , Xinjiang , and the clinical K-strain ) and the M . tb cell wall component ManLAM increase Mcl-1 gene expression [17 , 27 , 46 , 47] . We have previously shown that M . tb and ManLAM induce PPARγ through the MR [4] . Our current work showing that PPARγ induces Mcl-1 gene expression provides a mechanism for M . tb- and ManLAM-induced Mcl-1 expression which was previously unknown ( activation of MR and PPARγ ) . The importance of PPARγ in Mcl-1 expression may also explain why heat killed M . tb H37Rv does not induce Mcl-1 expression [17] , since our lab has previously shown that heat killed M . tb has significantly reduced ManLAM on the surface and is not recognized by the MR , which is required for PPARγ activation during M . tb infection [4 , 48] . Also , the clinical K-strain is more efficient at inducing Mcl-1 expression than M . tb strain H37Rv [27] . We have previously shown that ManLAM exposure is highly variable amongst clinical M . tb isolates [49] , suggesting that the M . tb K-strain has more exposed ManLAM and thus leads to more efficient MR recognition and PPARγ activity , and thus more Mcl-1 induction . In contrast , BCG induces PPARγ less efficiently than M . tb H37Rv [4] , and does not induce Mcl-1 expression [46 , 50] . These data all support the notion that M . tb and ManLAM , through MR and PPARγ , induce Mcl-1 expression , and that mycobacteria that do not efficiently activate PPARγ do not induce Mcl-1 expression . Although PPARγ plays a critical role in regulation of Mcl-1 gene expression during M . tb infection of human macrophages , M . tb can regulate Mcl-1 expression through other mechanisms in murine macrophages . For example , in murine macrophages M . tb activates PKCδ and STAT3 , which induce Mcl-1 expression , and M . tb represses expression of miR-17 , which targets Mcl-1 [51] . The Mcl-1 promoter region is highly variable between mouse and man [52] and it is unclear if PKCδ and STAT3 contribute to Mcl-1 induction by M . tb in human cells . Here , we identified six PPREs in the human Mcl-1 promoter , and the murine Mcl-1 promoter similarly contains putative PPREs , suggesting that although Mcl-1 expression may be regulated differently in murine and human cells , this is potentially driven by PPARγ in both species . Previous work by our lab indicated that the phospholipase cPLA2 is important for PPARγ activity [4] . cPLA2 mediates release of AA from the plasma membrane , which is utilized to generate various eicosanoids [32] . Altered eicosanoid levels are associated with TB disease progression [53] and enzymes involved in eicosanoid generation contribute to resistance ( 5-LOX ) and susceptibility ( COX-2 ) to M . tb in animal models [54 , 55 , 56] . It is largely unknown how these eicosanoids regulate macrophage immune function during M . tb infection , although some of this is likely through PPARγ since various eicosanoids are suggested to activate PPARγ in vivo , including 15d-PGJ2 ( which requires COX-2 ) and 13-HODE ( which requires 15-LOX ) [9 , 32] . Our current work shows that PPARγ activity and Mcl-1 expression are independent of COX-2 ( S3 Fig ) , but requires 15-LOX ( Fig 4 ) , providing evidence that 15-HETE and/or 13-HODE could serve as endogenous PPARγ agonists . This expands our previous work , and highlights that cPLA2 and 15-LOX are both critical for PPARγ activity during M . tb infection of human macrophages . Mcl-1 plays a critical role in limiting apoptosis , including during M . tb infection [17] . Mcl-1 also limits autophagy during M . tb infection , and reduces M . tb co-localization with Beclin-1 and lysotracker [31 , 57] . PPARγ has been linked to induction of autophagy , through increased Beclin-1 expression [58] but the role of PPARγ in regulation of apoptosis is less clear , with studies indicating that PPARγ agonists induce [59–62] and repress [28 , 29 , 63 , 64] apoptosis . This controversy may be due to cell-type specific responses ( immortalized cell lines vs primary cells ) , off-target effects of the PPARγ agonists [59 , 60] and concentration of agonist used [63] . For example , many studies reporting PPARγ induction of apoptosis use PPARγ agonists at high concentrations , which are known to have off-target effects ( or over-express PPARγ , which can also lead to off-target effects ) , and/or do not confirm that the effect seen with the agonist occurs through PPARγ ( using knockdown or inhibition approaches ) [61 , 62] . Indeed , other studies have shown that TZD induction of apoptosis occurs independently of PPARγ , since this occurs in cells that do not express PPARγ or in cells where PPARγ has been inhibited [59 , 60] . Another study showed high concentrations of PPARγ agonists induce T cell death , while lower concentrations enhance T cell survival; the latter occurred in a PPARγ dependent manner [63] . Here , using PPARγ- and 15-LOX-specific siRNA and/or inhibitors , we show that PPARγ is important for M . tb limiting of human macrophage apoptosis ( Fig 5 ) , and provide a mechanism for this: regulation of Mcl-1 expression . The pro-survival Bcl-2 proteins , including Bcl-2 and Mcl-1 , are highly expressed in various cancers , including follicular lymphoma and chronic lymphocytic leukemia ( CLL ) , and expression levels of multiple pro-survival Bcl-2 proteins , including Bcl-2 and Mcl-1 , are correlated with survival outcomes , and resistance to chemotherapeutic agents [65] . As such , targeting these proteins has been a focus of drug development , and clinical trials are underway with promising results for the Bcl-2 specific inhibitor ABT-199—survival of > 2 years for 59% of CLL patients—leading to advancement to Phase III trials [37] . Development and optimization of drugs that target the pro-survival Bcl-2 proteins , including drugs that specifically target Mcl-1 are underway . We capitalized on this active area of research and tested two pan pro-survival Bcl-2 inhibitors that also inhibit Mcl-1 ( sabutoclax and TW-37 ) and two Mcl-1 specific inhibitors ( MIM-1 and A-1210477 ) for their ability to limit M . tb growth in macrophages and in vitro TB granulomas [37] . We show that the Mcl-1 specific inhibitors MIM-1 and A-1210477 significantly limit M . tb growth in macrophages and that the general Bcl-2 inhibitors sabutoclax and TW-37 also significantly limit M . tb growth in macrophages and granulomas ( Fig 6 ) . Sabutoclax and TW-37 have been successful in cancer xenograft animal models at significantly limiting tumor growth with minimal to no animal toxicity [38 , 66 , 67] , highlighting the potential use of these inhibitors as therapies for M . tb infection . Since laboratory and clinical M . tb strains induce Mcl-1 expression , inhibiting Mcl-1 activity in human cells is expected to limit M . tb growth , regardless of strain . Targeting these molecules may be a viable HDT option for the various infections that are restricted with apoptosis ( e . g . Streptococcus pneumonia and Legionella pneumophila ) [14] . We further show that inhibiting Mcl-1 expression by inhibiting 15-LOX activity significantly reduces M . tb growth , indicating that targeting Mcl-1 expression and activity represent potential therapeutic routes . In summary , we have identified a novel PPARγ effector , Mcl-1 , and demonstrated upstream mediators of Mcl-1 expression in human macrophages . 15-LOX , PPARγ and Mcl-1 contribute to reducing apoptosis during M . tb infection , highlighting the importance of the nuclear receptor PPARγ in regulating mediators of apoptosis , and providing a potential mechanism for the critical role of 15-LOX and PPARγ during M . tb intracellular growth . Importantly , decreased Mcl-1 expression or activity limits M . tb intracellular growth , opening the door to a new potential HDT target . In this regard , repurposing cancer therapeutics that target Mcl-1 becomes a viable strategy for limiting M . tb growth in human macrophages and TB granulomas .
Peripheral blood mononuclear cells ( PBMCs ) were isolated from human peripheral blood collected from healthy donors , following OSU and Texas Biomed approved IRB protocols . HAMs were isolated from bronchoalveolar lavage of healthy human donors [68] , following OSU approved IRB protocols . All donors for these studies provided informed , written consent . MDMs were prepared as described elsewhere [69 , 70] . Briefly , heparinized blood was layered on a Ficoll-Paque cushion ( GE Healthcare , Uppsala , Sweden ) to allow for collection of PBMCs . PBMCs were cultured in RPMI ( Life Technologies , Carlsbad , CA ) with 20% autologous serum in Teflon wells ( Savillex , Eden Prairie , MN ) for 5 days at 37°C/5% CO2 . MDMs were harvested and adhered to tissue culture dishes for 2–3 h in RPMI with 10% autologous serum , lymphocytes were washed away , and MDMs were incubated overnight in RPMI with 10% autologous serum . Such MDM monolayers are 99% pure and viable . In vitro TB granulomas were generated as described elsewhere [43] . Briefly , human peripheral blood was collected from healthy Mantoux tuberculin skin test ( TST ) and/or IFNγ release assay ( IGRA ) -positive individuals . PBMCs were isolated by the published protocol as above , and were immediately infected with M . tb at MOI 1 in RPMI with 10% autologous serum , then incubated at 37°C/5% CO2 . After 1 day , Mcl-1 inhibitors were added , and additional serum was added after 4 days . Cells were incubated for a total of 4 or 7 days before M . tb intracellular growth was enumerated with colony forming units ( CFUs ) , as described below . Lyophilized M . tb H37Rv ( 27294 ) was obtained from the American Tissue Culture Collection ( ATCC , Manassas , VA ) . M . tb H37Rv lux was created and used as described [71] . This bacterial strain contains the entire bacterial Lux operon cloned in a mycobacterial integrative expression vector . mCherry M . tb was a kind gift from Dr . Sarah Fortune ( Harvard University , Boston , MA ) . Single cell suspensions of bacteria were prepared as previously described [72 , 73] . The bacteria concentration and degree of clumping ( <10% ) were determined with a Petroff-Hausser Chamber . This method results in ≥90% viable bacteria , as determined by CFU assay . Single cell suspensions of M . tb in RHH [10mM HEPES ( Life Technologies ) and 0 . 1% human serum albumin ( CSL Behring , King of Prussia , PA ) in RPMI] were added to the macrophages at various MOIs and cells were incubated for 2 h at 37°C , with the first 30 min on a platform shaker . Macrophages were then washed and incubated in RPMI with 2% autologous serum for the indicated times . For luciferase-based M . tb growth assays , MDMs were infected with M . tb-lux , and bacterial bioluminescence was measured every 24 h for up to 7 days with a GloMax Multi Detection System ( Promega , Madison , WI ) [71] . Where indicated , MDMs were pre-treated with solvent controls ( DMSO ) , PPARγ ( 1 h ) , LOX ( 1 h ) , or COX-2 ( 30 min ) inhibitors , with or without the PPARγ agonist 13-HODE ( 1 h ) prior to infection . The PPARγ inhibitor GW9662 , the COX-2 inhibitors CAY10404 and NS-398 , the 15-LOX inhibitor PD146176 , and 13-HODE were purchased from Cayman chemical ( Ann Arbor , MI ) , the 5-LOX inhibitor zileuton was purchased from Sigma ( St . Louis , MO ) , and the LOX inhibitors nordihydroguaiaretic acid ( NDGA; pan inhibitor ) and ML351 ( 12 , 15-LOX inhibitor ) were purchased from Calbiochem ( Billerica , MA ) . The Mcl-1 and caspase inhibitors , or solvent control ( DMSO ) , were added 2 h after infection . For caspase inhibition , cells were pre-treated with 100nM Q-VD-OPH ( MP Biomedicals , Santa Ana , CA ) 1 h prior to addition of 30 μm A-1210477 and cells were pulsed with 100nM Q-VD-OPH every 24 h . The Mcl-1 inhibitor MIM-1 was purchased from APExBIO ( Houston , TX ) . The Mcl-1 inhibitors A-1210477 , TW-37 , and sabutoclax were purchased from Selleckchem ( Houston , TX ) . All inhibitors were maintained throughout the course of infection . MDMs were transfected with 50 nM Accell PPARγ siRNA ( GAUUGAAGCUUAUCUAUGA ) , 50 nM SMARTpool Mcl-1 siRNA ( GGUUUGGCAUAUCUAAUAA , GAAGGUGGCAUCAGGAAUG , GAUUAUCUCUCGGUACCUU , CGAAGGAAGUAUCGAAUUU ) , or the same concentration of scrambled control siRNA ( UGGUUUACAUGUCGACUAA , UGGUUUACAUGUUGUGUGA , UGGUUUACAUGUUUUCUGA , UGGUUUACAUGUUUUCCUA ) with TransIT-X2 Transfection reagent ( Mirus , Madison , WI ) , following the manufacturer’s recommendations . All siRNAs were purchased from Dharmacon ( Lafayette , CO ) . MDMs were incubated 24 h before use . MDMs from three different donors were transfected with scrambled control or PPARγ siRNA and then infected with M . tb at MOI 5 for 6 and 24 h . MDMs were lysed with TRIzol ( Invitrogen , Carlsbad , CA ) and total RNA was isolated according to the manufacturer’s recommendations . RNA was analyzed with the NanoString nCounter Human Immunology v2 panel ( NanoString , Seattle , WA ) , which contains primers for 15 housekeeping genes and 579 different immunology-related genes . NanoString processing was performed by the Ohio State University Comprehensive Cancer Center Genomics Shared Resource Core Facility . Data normalization and analysis were performed by the Ohio State University Center for Biostatistics according to the manufacturer’s guidelines using SAS 9 . 3 and R . Technical normalization was performed using spiked controls , and background was based on the included negative controls . Genes that had counts higher than background were normalized to the housekeeping controls , and fold change in gene expression following PPARγ knockdown was calculated for each donor . Genes that showed a mean fold change of 1 . 5 , with p < 0 . 05 were considered significantly changed . STRING Analysis [74] was performed on genes that were significantly changed after 24 h M . tb infection . Macrophages were stimulated with 100 nM rosiglitazone ( Abcam , Cambridge , MA ) in RPMI with 2% autologous serum overnight . When indicated , macrophages were pre-treated with 0 . 1–10 μM PPARγ antagonist GW9662 ( Cayman chemical ) for 1 h prior to rosiglitazone stimulation . GW9662 was maintained during the rosiglitazone stimulation . Cells were washed with PBS , then lysed with TN1 lysis buffer ( 125 mM NaCl , 50 mM Tris , 10 mM EDTA , 1% Triton X-100 , 10 mM Na4PO7 , 10 mM NaF with 10 mM Na3VO4 , 10 μg/ml aprotinin , and 10 μg/ml leupeptin ) at 4°C . Lysates were centrifuged ( 10 , 000g , 4°C , 10 min ) to remove cell debris , then a Pierce BCA assay ( Thermo Scientific , Waltham , MA ) was performed to determine protein concentration . Equivalent amounts of denatured and reduced protein were separated by SDS-PAGE and analyzed by Western blot using antibodies against PPARγ ( C26H12 Cell Signaling , Danvers , MA ) , Mcl-1 ( Santa-Cruz , Dallas , TX ) , and β-actin ( Santa Cruz ) . Protein band intensities were determined with ImageJ , for each sample background values were subtracted and then values were normalized to the β-actin loading control . Macrophages in triplicate wells were lysed with TRIzol ( Invitrogen ) and total RNA was isolated according to the manufacturer’s recommendations . The NanoDrop 1000 was used to determine quantity and quality of RNA . cDNA was reverse transcribed from RNA with SuperScript III Reverse Transcriptase ( Invitrogen ) . Gene expression was determined by quantitative real-time RT-PCR ( qRT-PCR ) using TaqMan Gene Expression Assays ( Applied Biosystems , Foster City , CA ) and a CFX96 Real-Time System ( Bio-Rad , Hercules , CA ) . Values were normalized to β-actin , which was used as a housekeeping gene with the Bio-Rad CFX Manager using the ΔΔCq method . pSV Sport PPARγ2 was a gift from Bruce Spiegelman ( Addgene plasmid #8862 ) [75] . The pGL3-Basic vector containing the full length Mcl-1 promoter region was kindly provided by Dr . Steven W . Edwards ( [52]; University of Liverpool , Liverpool , UK ) and Dr . Daqing Wu ( [76]; Augusta University , Augusta , GA ) . Six PPRE consensus sequences were identified using Genomatix software [77] , and submission to http://www . classicrus . com/PPRE/ and http://www . cbrc . kaust . edu . sa/ppre/ . The high complexity of the Mcl-1 region with several repeats , homopolymeric stretches and high GC content demanded a variety of strategies be employed to mutate the various PPRE to non-consensus sequences ( summarized in Fig 2E ) . Sites #4–6 were replaced with gBlocks ( IDT , Coralville , IA ) ; site #3 was replaced by SOE-PCR; site #2 was mutated with QuikChange mutagenesis ( Agilent Technologies , Santa Clara , CA ) ; site #1 employed limited inverse PCR; all other combinations were subsequently assembled by subcloning . Mutants were fully sequenced through the Mcl-1 region and , where whole plasmid mutagenesis was used , the mutant region was further back cloned to pGL3Mcl-1 . RAW 264 . 7 cells ( ATCC TIB-71 ) were maintained in 10% HI-FBS/0 . 1% penicillin-streptomycin/DMEM ( Life Technologies ) and co-transfected with the above PPARγ and Mcl-1 constructs using Lipofectamine 2000 ( Life Technologies ) according to the manufacturer’s instructions . Cells were stimulated with 0 . 1 μM rosiglitazone 24 h after transfection , and luciferase activity assayed after an additional 24 h using the Promega Luciferase Assay system . For LPS stimulations , cells were transfected with Mcl-1 constructs as above , then treated with 1 μg/ml LPS for 24 h before luciferase activity was assessed . Protein concentration in the lysates was determined with a Bradford Assay ( Bio-Rad ) and luciferase activity was normalized to protein concentration for each sample . MDMs were incubated with the COX-2 inhibitors for 30 min prior to addition of M . tb at MOI 5 or 1 μg/ml LPS ( Sigma ) . After 24 h , cell free supernatants were collected and the amount of PGE2 in the supernatant was analyzed with a PGE2 EIA kit according to the manufacturer’s instructions ( Cayman Chemical ) . TUNEL staining: Transfected MDMs on coverslips were infected with mCherry M . tb at MOI 5 and 50 for 24 and 48 h . Cells were then fixed with 4% PFA ( Affymetrix , Santa Clara , CA ) and labeled using the Click-iT TUNEL Alexa Fluor Imaging Assay ( Invitrogen ) following the manufacturer’s instructions . Cells were imaged with an Olympus FV1000 confocal microscope ( Olympus , Shinjuku , Japan ) . Using Olympus Fluoview Viewer , at least 100 MDMs were manually counted to quantify % MDMs that stained with TUNEL . CellTiter Glo Assays: Transfected or inhibitor treated MDMs in 96 well plates were infected with M . tb at MOI 5 for 24 h and cell death was assayed in triplicate with the CellTiter Glo Assay ( Promega ) following the manufacturer’s instructions . Intracellular growth was assayed with two approaches . For CFU assays , infected MDMs were lysed as described previously [78] . Lysates were diluted , and plated on 7H11 agar ( BD , Franklin Lakes , NJ ) . The number of CFUs was enumerated after growth for 3–4 weeks at 37°C . For luciferase growth assays , MDMs were infected with M . tb-lux , and bacterial bioluminescence was measured in relative luminescence units ( RLUs ) every 24 h for up to 7 days with a GloMax Multi Detection System ( Promega ) [71] . For measurement of M . tb growth in the absence of macrophages , M . tb was incubated in 7H9 broth ( BD ) with 30 μM of the Mcl-1 inhibitors . After 4 d at 37°C , M . tb was diluted and plated on 7H11 agar and CFUs were enumerated after growth for 3–4 weeks at 37°C . To assess monolayer integrity during the course of experiments , three images per condition were acquired under 40x magnification with phase microscopy ( Olympus DP71 microscope digital camera ) . The total number of cells per field of view was enumerated with ImageJ and then averaged together to calculate relative MDM counts . Macrophages from at least three different donors were used for each assay , unless indicated otherwise . Although the trend was the same for each donor , the magnitude of change differed among donors . Consequently , results from each experiment were normalized to an internal control and an unpaired one-tailed Student’s t-test or ANOVA were performed on the normalized data , with P < 0 . 05 considered significant .
|
The bacterium Mycobacterium tuberculosis is the causative agent of the disease tuberculosis ( TB ) , which is a global health problem and a leading cause of death world-wide . There is a clear need for better therapies for this disease , the design of which is predicated on better understanding of how M . tuberculosis interacts with the host . Alveolar macrophages ( AMs ) are sentinels in the lung which clear inhaled agents , including pollutants , allergens and microbes; yet do not efficiently clear M . tuberculosis . Here we show that a transcription factor highly expressed in AMs ( PPARγ ) that is critical for M . tuberculosis growth inside human macrophages , regulates the anti-cell death protein Mcl-1 . We also characterize upstream molecules required for Mcl-1 production and show that PPARγ is important for M . tuberculosis regulation of cell death . Excitingly , we also show that pre-clinical Mcl-1 inhibitors significantly inhibit M . tuberculosis growth in human macrophages and multicellular granuloma structures . In summary , here we identify a novel effector of the global regulator PPARγ , determine a role for PPARγ in limiting cell death , and show that the anti-cell death protein Mcl-1 is a promising host-directed target for TB therapy .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"blood",
"cells",
"cell",
"death",
"medicine",
"and",
"health",
"sciences",
"luciferase",
"immune",
"cells",
"gene",
"regulation",
"granulomas",
"enzymes",
"immunology",
"cell",
"processes",
"enzymology",
"tropical",
"diseases",
"bacterial",
"diseases",
"bacteria",
"small",
"interfering",
"rnas",
"infectious",
"diseases",
"white",
"blood",
"cells",
"animal",
"cells",
"proteins",
"tuberculosis",
"gene",
"expression",
"oxidoreductases",
"actinobacteria",
"biochemistry",
"rna",
"cell",
"biology",
"nucleic",
"acids",
"mycobacterium",
"tuberculosis",
"apoptosis",
"genetics",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"macrophages",
"non-coding",
"rna",
"organisms"
] |
2018
|
PPARγ is critical for Mycobacterium tuberculosis induction of Mcl-1 and limitation of human macrophage apoptosis
|
In order to analyse large complex stochastic dynamical models such as those studied in systems biology there is currently a great need for both analytical tools and also algorithms for accurate and fast simulation and estimation . We present a new stochastic approximation of biological oscillators that addresses these needs . Our method , called phase-corrected LNA ( pcLNA ) overcomes the main limitations of the standard Linear Noise Approximation ( LNA ) to remain uniformly accurate for long times , still maintaining the speed and analytically tractability of the LNA . As part of this , we develop analytical expressions for key probability distributions and associated quantities , such as the Fisher Information Matrix and Kullback-Leibler divergence and we introduce a new approach to system-global sensitivity analysis . We also present algorithms for statistical inference and for long-term simulation of oscillating systems that are shown to be as accurate but much faster than leaping algorithms and algorithms for integration of diffusion equations . Stochastic versions of published models of the circadian clock and NF-κB system are used to illustrate our results .
Dynamic cellular oscillating systems such as the cell cycle , circadian clock and other signaling and regulatory systems have complex structures , highly nonlinear dynamics and are subject to both intrinsic and extrinsic stochasticity . Moreover , current models of these systems have high-dimensional phase spaces and many parameters . Modelling and analysing them is therefore a challenge , particularly if one wants to both take account of stochasticity and develop an analytical approach enabling quantification of various aspects of the system in a more controlled way than is possible by simulation alone . The stochastic kinetics that arise due to random births , deaths and interactions of individual species give rise to Markov jump processes that , in principle , can be analyzed by means of master equations . However , these are rarely tractable and although an exact numerical simulation algorithm is available [1] , for the large systems we are interested in , this is very slow . It is therefore important to develop accurate approximation methods that enable a more analytical approach as well as offering faster simulation and better algorithms for data fitting and parameter estimation . A number of approximation methods aimed at accelerating simulation are currently available . This includes leaping algorithms [2 , 3] and algorithms for integration of diffusion equations ( or chemical Langevin equations ( CLE ) ) [4] that provide faster simulation . However , these methods do not provide analytical tools for studying the dynamics of the system and they can also be extremely slow for data fitting and parameter estimation . One obvious candidate for overcoming these limitations is the Linear Noise Approximation ( LNA ) . The LNA is based on a systematic approximation of the master equation by means of van Kampen’s Ω-expansion [5] which uses the system size parameter Ω that controls the number of molecules present in the system . The large system size validity of the LNA has been shown in [6] , in the sense that the distribution of the Markov jump process at a fixed finite time converges , as the system size Ω tends to ∞ , to the LNA probability distribution . The latter distribution is analytically tractable allowing for fast estimation and simulation algorithms . However , the LNA has significant limitations , particularly in approximating long-term behaviour of oscillatory systems . We show below that for the oscillatory systems that we study , the LNA approximation of the distribution Pt = P ( Y ( t ) |Y ( 0 ) ) , of the state Y ( t ) of the system at some time t becomes inaccurate when the time t is greater than a few periods of the oscillation . However , if we rather consider a similar system which in the Ω → ∞ limit instead of a limit cycle has an equilibrium point that is linearly stable , then the LNA approximation of Pt remains accurate for a much longer time-scale . For example , in Fig C in S3 Appendix we give an example where the LNA fails in a matter of a period or two for the oscillatory system , but for the corresponding equilibrium system it is very accurate for over a hundred times as long ( and probably much longer ) . Similar behaviour is also observed in other systems and using different measures in [8] . The observation that non-degenerate limit cycles have such linearised stability in the directions transversal to the limit cycle suggests the way forward for oscillatory systems . Our approach exploits the fact that , because of this transversal linearised stability , the distributions Pt for a general class of systems with a stable attracting limit cycle in the Ω → ∞ limit are , like the above fixed point systems , similarly well-behaved on long time-scales provided one conditions Pt on appropriate transversal sections to the limit cycle . We introduce a modified LNA , called the phase-corrected LNA , or pcLNA , that exploits the above observations to overcome the most important shortcomings of the LNA and we develop methods for analysis , simulation and inference of oscillatory systems that are accurate for much larger times . We build on previous work of Boland et al . [9] which uses the 2-dimensional Brusselator system as an exemplar to investigate the failure of the LNA in approximating long-term behaviour of oscillatory systems and presents a method for computing power spectra and comparing exact simulations with LNA predictions of the same phase rather than time . Using various low-dimensional oscillatory systems for illustration , a related analysis has been employed to study the temporal variability of oscillatory systems in the tangental direction of the Ω → ∞ limit cycle [10] and/or the amplitude variability in the transversal direction of the limit cycle [11–13] . Other papers derive related descriptions of the asymptotic phase of stochastic oscillators [14 , 15] . We extend these results in a number of ways including the following: ( i ) we develop a theory that treats the general case and provide analytical arguments that justify our approximations and enable computation of trajectory distributions , ( ii ) we show that the approach is practicable for larger nonlinear systems , ( iii ) we present a new powerful system-global sensitivity theory for such systems using measures such as the Fisher Information Matrix and the Kullback-Leibler divergence that are analytically computed , ( iv ) we present a simulation algorithm and show it is as accurate but faster than leaping and integration of diffusion equation algorithms , and ( v ) we derive the Kalman filter associated with the pcLNA in order to provide a practical way to accurately approximate the likelihood function thus facilitating estimation of system parameters θ and predictive algorithms . The approach in [9] uses transversal sections which are normal to the limit cycle . We follow this but in the supplementary information ( S1 Sects . 8 . 2 & 8 . 3 ) we show that for most considerations one can use any transversal to the limit cycle , including those defined in [14 , 15] . To illustrate and validate our approach we apply it to a relatively large published stochastic model of the Drosophila circadian clock due to Gonze et al . [16] ( see S2 Sect . 1 ) . This model involves 10 state variables and 30 reactions and its structure is discussed in S2 . The large system limit is given by the differential equation system of 10 kinetic equations that are listed in the supplementary information ( S2 Sect . 1 ) along with the reaction scheme of the system . The stochastic version of the Brusselator system and a stochastic version of a well-studied model of the NF-κB signalling system [17] are also used to illustrate our methods and the results can be found in S3 Appendix . These systems are free-running oscillators in the sense that they correspond to a limit cycle of an autonomous differential equation in the the Ω → ∞ limit . However , our results also apply to the equally important classes of entrained forced oscillators and damped oscillations . We therefore consider two such systems in S2 and S3: the light-entrained Drosophila circadian clock model of [18] which is an example of a forced oscillator and the NF-κB system model [17] . The latter has the extra feature that the analysis is not concerned with a limit cycle but of a transient solution that converges to the limit cycle as time increases . This solution is the biologically interesting one that describes how the system responds to being stimulated by TNFα . The supplementary information S1 includes technical derivations and S2 and S3 contain further illustrative figures that we refer to in this paper .
The convergence of the transversal distributions to approximately normal distributions naturally raises the question of whether asymptotic approximation methods such as the LNA , which provide multivariate normally distributed approximation of the stochastic system , can be used to accurately approximate these transversal distributions . The LNA as formulated by [6] is derived directly from the underlying Markov jump process and is valid for any time interval of finite fixed length . It is based on the ansatz X ( t ) = Y ( t ) Ω = x ( t ) + ξ ( t ) Ω ( 3 ) where x ( t ) is a solution of the limiting ( Ω → ∞ ) deterministic system Eq ( 1 ) and ξ ( t ) / Ω describes the stochastic variations . In our case we always take x ( t ) to be the periodic solution g ( t ) . A key aspect of this ansatz is that ξ ( t ) satisfies a linear stochastic differential equation that is independent of Ω , with drift and diffusion matrices that are functions of the deterministic solution g ( t ) . Details are given in S1 Sect . 4 & 5 . Given an initial time t0 and an initial condition ξ ( t0 ) for ξ , the LNA determines the distribution , of ξ ( t ) , t > t0 , and hence X ( t ) = g ( t ) + ξ ( t ) / Ω that we respectively denote by PLNA ( ξ ( t ) |t0 , ξ ( t0 ) ) and PLNA ( X ( t ) |t0 , ξ ( t0 ) ) . If ξ ( t0 ) is only known up to a multivariate normal ( MVN ) distribution P0 then we denote these distributions , respectively , by PLNA ( ξ ( t ) |t0 , ξ ( t0 ) ∼P0 ) and PLNA ( X ( t ) |t0 , ξ ( t0 ) ∼P0 ) . Details of how to calculate these distributions are given in S1 Sect . 4 . Each of the above distributions is MVN enabling analytical approaches , for example in analysing the stochastic sensitivities of the system . If we fix t > t0 then as Ω → ∞ the true distribution of ξ converges to the distribution PLNA ( ξ ( t ) |t0 , ξ ( t0 ) ) ( see e . g . [5] ) . However , one most certainly cannot reverse the limits i . e . for a fixed Ω one cannot expect the approximation to hold for large time t → ∞ . As we now show , this is certainly the case for oscillators and we aim to overcome this limitation by developing methods that remain accurate for much larger times than the LNA . We first consider the distribution P ( X ( t ) |X ( 0 ) = x0 ) and compare this for SSA simulated samples and the LNA at a sequence of times t = τ , 2τ , … , 8τ and for an arbitrary ( fixed ) initial state x0 ∈ γ . As we can see in Fig 3 , the LNA fits the SSA simulations relatively well in the short run ( t ≤ τ ) , but as time progresses the Kolmogorov-Smirnov ( KS ) distance between the two distributions for each state variable for the LNA and the SSA increases substantially beyond the threshold level ( see Fig 3 ( B ) ) . The LNA predictions spread along the tangental direction and therefore fail to accurately reflect the SSA samples that have instead spread along the curved limit cycle . On the other hand , as we saw earlier , the transversal distributions P x 1 ( r ) = P ( Q x 1 ( r ) | X ( t 0 ) = x 0 ) of the Drosophila circadian clock system are approximately normal ( Fig 3 ( B ) ) . We next derive an approximation of P x 1 ( r ) under the LNA and show that it accurately approximates P x 1 ( r ) for the Drosophila circadian clock , Brusselator and NF-κB systems . We now generalise slightly and consider the set of stochastic trajectories X where the initial conditions X ( t0 ) have a MVN distribution P0 that is supported on the normal transversal section S g ( t 0 ) ( denoted X ( t0 ) ∼P0 ) . We consider how to approximate the distribution P x 1 ( r , P 0 ) of the intersection points Q g ( t ) ( r , X ) of these trajectories with the normal transversal section S g ( t 1 ) , t1 > t0 . As an approximation we take the conditional distribution P LNA , t 1 ( r ) = PLNA ( X ( t 1 ( r ) ) | X ( t 1 ( r ) ) ∈ S g ( t 1 ) , ξ ( t 0 ) ∼ P 0 ) , t 1 ( r ) = t 1 + ( r - 1 ) τ ( 4 ) given by conditioning P LNA , t 1 ( r , free ) = PLNA ( X ( t 1 ( r ) ) | ξ ( t 0 ) ∼ P 0 ) on X ( t1 ( r ) ) ∈ Sg ( t1 ) . It gives a MVN distribution supported on S g ( t 1 ) . In S1 we show that , although , for free-running oscillators , P LNA , t 1 ( r , free ) diverges as r → ∞ , the mean and covariance of the MVN transversal distribution P LNA , x 1 ( r ) converge exponentially fast to those of a MVN distribution P LNA , x 1 ( ∞ ) ( S1 Sect . 8 , cf . Fig 4 ) . The distribution P LNA , x 1 ( ∞ ) is a fixed point in the sense that if the distribution of X ( t1 + τ ) is given by the LNA using as initial condition ξ ( t 1 ) ∼ P LNA , x 1 ( ∞ ) then conditioning on X ( t1 + τ ) ∈ S g ( t 1 ) gives ( X ( t 1 + τ ) | X ( t 1 + τ ) ∈ S g ( t 1 ) ) ∼ P LNA , x 1 ( ∞ ) . Using this fact enables us to calculate P LNA , x 1 ( ∞ ) directly because we show in S1 that its mean and covariance matrix satisfy a simple fixed point equation that is easily solved numerically ( S1 Sect . 9 . 1 ) . The reader will note that in Eq ( 4 ) we approximate by conditioning on X ( t 1 ( r ) ) ∈ S g ( t 1 ) whereas we should have conditioned on X ( t ) ∈ S g ( t 1 ) for arbitrary t corresponding to the rth round . In S1 Sect . 7 we argue that the error in the mean and variance of the distribution due to taking t = t 1 ( r ) is O ( Ω−1 ) . The question remains as to how well these distributions capture the exact simulation transversal distribution P x 1 ( r ) . This is addressed in Fig 5 where it is shown that the fit is excellent even for Ω as low as 300 . The fit is even better for higher system sizes ( Fig E in S2 Appendix ) . In S3 we also show similar low Ω results for the Brusselator ( Fig B in S3 Appendix ) and the NF-κB system ( Fig E in S3 Appendix ) . The result is also true for the light-entrained Drosophila circadian clock system ( see Fig J in S2 Appendix ) and the transient oscillations of the NF-κB system ( see Fig H in S3 Appendix ) . Thus we note that although the LNA cannot be used directly to accurately compute P ( X ( t ) |X ( 0 ) ) for a fixed Ω and increasing t , using it to compute the transversal distributions provides accurate estimates of P x 1 ( r ) for much larger times t1 + rτ . Moreover , in S1 Sects . 8 . 2 & 8 . 3 we also explain why the convergence of the distribution on normal hyperplanes implies convergence on other transversal sections to γ . In S1 Sect . 8 . 4 . we explain that in contradistinction to free-running oscillators , for entrained forced oscillators , Pr = PLNA ( X ( ( r − 1 ) τ + t1 ) |x0 , ξ0 ∼ P0 ) converges as r → ∞ so that , under the LNA , the phase fluctuation have a variance that is bounded independently of r . The corresponding conditional distribution is therefore a correspondingly good approximation to the transversal distribution P x 1 ( r , P 0 ) ( see Fig J in S2 Appendix ) . However , it does not mean that P LNA , t 1 ( r , free ) is a good approximation to the corresponding distribution P ( X ( t 1 ( r ) ) | X ( 0 ) ) for an exact simulation . In fact , we show in Fig I in S2 Appendix that P LNA , t 1 ( r , free ) is a poor approximation of the empirical distribution P ( X ( t 1 ( r ) ) | X ( 0 ) ) derived from exact simulations for the light-entrained Drosophila circadian clock ( Ω = 300 ) . The bounded variance of the phase fluctuations as r → ∞ for forced oscillators is the basic mechanism behind the population-level entrainment of stochastic oscillators introduced in [21] . We now consider the fluctuations δt in the time taken for the lifted phase of a stochastic trajectory to go from a given phase φ1 to a greater one φ2 . If φ2 − φ1 = 2rπ then this corresponds to the time taken to perform r cycles . In Fig 6 we give an example using the Drosophila circadian clock model where we take φ1 to be 0 ( with a fixed initial condition x0 ∈ γ ) and φ2 = 2πr for r = 1 , … , 8 . The distributions of δt appear to be very close to normal and the variance appears to grow linearly with r . We also consider the case where φ1 = 2 ( r − 1 ) π and φ2 = 2rπ for r = 1 , … , 8 . Again the distributions are approximately normal but the variances are approximately constant ( Fig 6 ( C ) ) . Because for a given r the trajectory has done r − 1 cycles before reaching the lifted phase φ1 the distribution of the state at this phase is changing with r . We expect that this distribution is converging with increasing r and this result ( Fig 6 ( C ) ) is in accordance with this . In S1 Sect . 6 we approximate the statistics of δt using the LNA and show that as a random variable δt is approximately normal with mean that is O ( Ω−3/2 ) and we also calculate its variance up to terms that are O ( Ω−3/2 ) and the extent of its divergence from normality . All points with a given lifted phase φ1 lie in a particular transversal S g ( t 1 ) with 0 ≤ t1 < τ . If t2 = t1 + ( φ2 − φ1 ) τ/2π , then , the mean and variance of δt can be calculated in terms of t1 and t2 . If the initial conditions ξ ( t1 ) are MVN distributed on S g ( t 1 ) with mean 0 and covariance V1 , this variance is ( 1 / α 2 Ω ) V ˇ 11 + O ( Ω - 3 / 2 ) where V ˇ 11 = V ˇ ( t 1 , t 2 ) 11 , V ˇ ( t 1 , t 2 ) = C ( t 1 , t 2 ) V 1 C ( t 1 , t 2 ) T + V ( t 1 , t 2 ) . written in adapted coordinates at g ( t2 ) ( see S1 Sect . 1 ) . All terms on the right hand side of this equation are defined in S1 Sect . 4 . The above exact simulations of the Drosophila circadian clock agree with these theoretical predictions . It is easy to see ( cf . S1 Sect . 8 ) that V ˇ 11 grows roughly linearly with t2 − t1 . Given the ability to accurately approximate the transversal distributions and the results in [9] we realised it should be possible to use this to construct a rapid simulation algorithm . The linear increase of the variance of the deviations δt , or equivalently , the linear growth in the variance of the deviation of the lifted phase φ ¯ X ( t i ) from 2πti/τ , indicates the reasons for the long-time failure of the standard LNA . It is unable to cope with the increasing phase deviations . This motivates the phase correction approach used in the simulation algorithm we now define . The approach is to amend the LNA Ansatz X ( t ) = g ( t ) + Ω−1/2 ξ ( t ) to X ( t ) = g ( s ) + Ω−1/2 κ ( s ) where g ( s ) = GN ( X ( t ) ) and to use resetting of t to s to cope with the growth in the variance of φ ¯ X ( t i ) - 2 π t i / τ keeping the LNA fluctuation κ ( s ) normal to γ . While for free-running oscillators the variance of ξ ( t ) grows without bound as t increases , κ ( s ) has uniformly bounded variance . The pcLNA simulation algorithm iteratively uses standard LNA steps of length Δτ to move from a state X ( si − 1 ) to a new state X ( si − 1 + Δτ ) = Xi , i = 1 , 2 , … . After each LNA step , the phase of the system is reset or “corrected” such that g ( si ) = GN ( Xi ) and the ( global ) fluctuations ξ ( si − 1 + Δτ ) = Ω1/2 ( Xi − g ( si − 1 + Δτ ) ) are replaced by the normally transversal fluctuation κ ( si ) = Ω1/2 ( Xi − g ( si ) ) which are MVN distributed and , as we showed in the previous section , approximate well the transversal fluctuations under the exact Markov Jump process . The steps of the pcLNA simulation algorithm are described next in more detail ( see also Fig 7 ) . In the for loop Ci = C ( si − 1 , si − 1 + Δτ ) and Vi = V ( si − 1 , si − 1 + Δτ ) are the drift and diffusion matrices in the linear SDE describing the evolution of the noise process ξ ( t ) under the LNA ( see S1 Sect . 4 ) . The simulated sample Xi corresponds to time ti = t0 + iΔτ , i = 1 , 2 , … , where t0 is the initial time . The time ti is not necessarily equal to the phase si , defined by the relation g ( si ) = GN ( Xi ) , which is stochastic and has variance linearly increasing with the time step Δτ . If one wants to record simulated trajectories at a finer time-scale than Δτ then one can run the algorithm with Δτ replaced by Δτ/M for some integer M > 1 and only carry out the phase correction in step 3 ( c ) every Mth step and at all the other steps just proceeding as in the standard LNA ( ignoring step 3 ( c ) ) . This gives the same distribution as if the intermediate points had not been calculated because of the transitive nature of the LNA i . e . the distribution PLNA ( X ( s + t ) |X ( 0 ) ) is equal to the distribution PLNA ( X ( t ) |X ( s ) ∼PLNA ( X ( s ) |X ( 0 ) ) ) . In the simulation results described below the time-step Δτ = 6 hours and M = 3 so that there are τ/6 ≈ 4 . 5 corrections in every round of the limit cycle . The effect of less frequent correction is studied in S2 Sect . 5 . The derivation of analytical expressions of the transversal distributions allows us to analyse various aspects of the stochastic behaviour of these systems that can possibly involve a large number of variables and parameters . Here we illustrate the use of pcLNA transversal distributions to perform such an analysis . We begin by describing the pcLNA joint distribution of multiple intersections to possibly different transversal sections on the limit cycle and then discuss Fisher information , sensitivity analysis and estimation by Kalman filtering . Fisher Information quantifies the information that an observable random variable carries about an unknown parameter θ . If P ( X , θ ) is a probability distribution depending on parameters θ , the Fisher Information Matrix ( FIM ) I = IP has entries I i j = E [ ∂ ℓ ∂ θ i ∂ ℓ ∂ θ j ] = - E [ ∂ 2 ℓ ( θ ; X ) ∂ θ i ∂ θ j ] , ( 6 ) where ℓ = log P , and θi and θj are the ith and jth components of the parameter θ . If P is MVN with mean and covariance μ = μ ( θ ) and Σ = Σ ( θ ) then I i j = ∂ μ ∂ θ i T Σ - 1 ∂ μ ∂ θ j + 1 2 tr ( Σ - 1 ∂ Σ ∂ θ i Σ - 1 ∂ Σ ∂ θ j ) . ( 7 ) The FIM measures the sensitivity of P to a change in parameters in the sense that D K L ( P ( · , θ + δ θ ) , P ( · , θ ) ) = 1 2 δ θ T I P δ θ + O ( ‖ δ θ ‖ 3 ) where DKL is the Kullback-Leibler divergence . The significance of the FIM for sensitivity and experimental design follows from its role in Eq ( 6 ) as an approximation to the Hessian of the log-likelihood function at a maximum . Assuming non-degeneracy , if θ* is a parameter value of maximum likelihood there is a s × s orthogonal matrix V such that , in the new parameters θ′ = V ⋅ ( θ − θ* ) , ℓ ( θ ) ≈ ℓ ( θ * ) - ∑ i σ i 2 θ i ′ 2 . for θ near θ* . From these facts it follows that the σ i 2 are the eigenvalues of the FIM and that the matrix V diagonalises it . If we assume that the σi are ordered so that σ 1 2 ≥ … ≥ σ s 2 then it follows that near the maximum the likelihood is most sensitive when θ 1 ′ is varied and least sensitive when θ s ′ is . Moreover , σi is a measure of this sensitivity . The theory of optimal experimental design is based on the idea of trying to make the σi decrease as slowly as possible so that the likelihood is as peaked as possible around the maximum , thus maximising the information content of the experimental sampling methods . Various criteria for experimental design have been proposed including D-optimality that maximises the determinant of the FIM and A-optimality that minimises the trace of the inverse of the FIM [6] . Diagonal elements of the inverse of FIM constitute a lower-bound for variance of any unbiased estimator of elements of θ ( Cramer-Rao inequality ) . However , for the systems we consider here the σi typically decrease very fast and there are many of them . Thus , in general , criteria based on a single number are more likely to be of less use than consideration of the set of σi as a whole . Calculation of the FIM for stochastic systems using the LNA has been carried out in [22] but only for small systems and short times where the LNA is accurate . It is notable that the pcLNA enables one to do such sensitivity analysis for large systems and large times . As an example , we analyse the stochastic behaviour of the Drosophila circadian clock based on the limit distribution P ( Q _ | Q 0 ) when Q _ = Q x 0 ( 1 ) , Q x 1 ( 1 ) , Q x 0 ( 2 ) , Q x 1 ( 2 ) , … , Q x 0 ( m ) , Q x 1 ( m ) where x0 = g ( t0 ) and x1 = g ( t1 ) are chosen so that t0 is the time of the peak of per mRNA MP , and t1 is the peak of the nuclear complex of PER and TIM proteins CN . We compute the Fisher Information of the distribution P ( Q _ | Q 0 ) using the closed form expression ( S1 Sect . 9 . 3 ) for this distribution . As we can see in Fig 10 ( A ) the eigenvalues of the Fisher Information matrix decay exponentially , with a sharp decline followed by a slower decrease . This reveals that the influential directions in the parameter space of the system are much less than its total dimension and that only a few parameters appear to be most influential . The eigenvectors associated with the two largest eigenvalues of the Fisher Information matrix ( see Fig 10 ( B ) ) have large entries only for the parameters kdn ( PER-TIM complex nuclear degradation ) , kd ( per mRNA linear degradation ) , k2 ( PER-TIM complex transportation to cytosol ) , vst ( tim mRNA transcription ) , kip ( per mRNA Hill coefficient ) and kit ( tim mRNA Hill coefficient ) . The exponential decrease of the eigenvalues is typical of tightly coupled deterministic systems [25–31] , but has to our knowledge not been demonstrated before for stochastic systems . It has important consequences . For example , it tells us that only a few parameters can be estimated efficiently from time-series data unless the system is perturbed in some way to get complementary data and that there will be identifiability problems that can be analysed using the FIM . It can also be used to design experiments by considering the FIM of a combination of models including one for the proposed new experiment , choosing the new experiment so as to optimally alleviate the decline of the eigenvalues . The fact that we can calculate the Fisher Information allows a new approach to sensitivity analysis for stochastic systems . Anderson [23] and Srivastava et al . [24] also perform sensitivity analysis for small stochastic systems ( up to 4 species and 8 reactions ) in which they calculate the dependence of certain summary functions or statistics at one or more times to individual parameters . Our approach is different in that we use the fact that our distributions of interest are MVN and measure the change in the distribution of the system state at any given set of phases without recourse to any summary function and , moreover , this change is calculated for any combination of parameter variations . A major difference is our use of SVD below to find a basis of mean-covariance space using the principal components that enables us to decompose these changes into different orthogonal directions that pick out the important and unimportant directions . The approach can also be formulated for the wider class of exponential families i . e . distributions that admit a representation of the form P ( x ) = exp { C ( x ) + ∑ i θ i F i ( x ) - φ ( θ ) } in terms of functions C , F1 , …Fm of the state variable x and a function φ of the parameters θ . We consider a family of probability distributions P ( X , θ ) which we assume are MVN with mean μ ( θ ) and covariance matrix Σ ( θ ) depending on the parameters θ . We show that there is a natural matrix of sensitivities Sij associated with such a system . These are system-global in that they look at how all components of the systems change with parameters . They also have an intimate relationship with Fisher information . Note that these results are not restricted to the transversal distributions derived in previous sections but apply more generally to any MVN distribution with mean μ = μ ( θ ) and covariance matrix Σ = Σ ( θ ) parameterised by a s-dimensional vector θ , s ≥ 1 . As is well-known in Information Geometry , the set of multivariate normal distributions MVNn on ℝn can be given the structure of a Riemannian manifold in which the Riemannian metric is given by the line element d s 2 = d μ T Σ - 1 d μ + ( 1 / 2 ) tr { ( Σ - 1 d Σ ) 2 } . Points in MVNn are denoted by Θ = ( μ , Σ ) where μ is the mean and Σ the covariance matrix . The corresponding inner product in the tangent space at Θ0 = ( μ , Σ ) is given by ⟨ δ Θ 1 , δ Θ 2 ⟩ Θ 0 = δ μ 1 Σ - 1 δ μ 2 + 1 2 tr ( Σ - 1 δ Σ 1 Σ - 1 δ Σ 2 ) ( 8 ) where δΘj = ( δμj , δΣj ) , j = 1 , 2 . In calculating the FIM we have to determine the partial derivatives ∂μ/∂θi and ∂Σ/∂θi . The derivative M of the mapping θ → ( μ ( θ ) , Σ ( θ ) ) at a parameter value θ0 is given by M · δ θ = ( ∑ i ∂ μ ∂ θ i δ θ i , ∑ i ∂ Σ ∂ θ i δ θ i ) where the derivatives are calculated at θ0 . We can regard M as a linear mapping between the parameter space ℝs and MVNn with the inner product given in Eq ( 8 ) . We can then prove ( S1 Sect . 11 ) that we can find s orthonormal vectors Vi spanning the parameter space ℝs , s orthonormal vectors Ui in the space MVNn and numbers σ1 ≥ ⋯ ≥ σs ≥ 0 such that M V i = σ i U i , i = 1 , … , s . ( 9 ) Note that the orthonormality of the Ui is with respect to the inner product 〈⋅ , ⋅〉Θ0 . The eigenvalues of the FIM F are the squares of the σi because with respect to the standard inner product on θ-space and 〈⋅ , ⋅〉Θ0 on MVNn the adjoint M* satisfies M* M = F ( S1 Sect . 11 ) . If we let U i = ( U i μ , U i Σ ) denote the decomposition of Ui into μ and Σ components , then the following key property follows from Eq ( 9 ) : if δθ is any change of parameters , the change in μ and Σ is given by δ μ = ∑ i U i μ ( ∑ j S i j δ θ j ) + O ( ‖ δ θ ‖ 2 ) δ Σ = ∑ i U i Σ ( ∑ j S i j δ θ j ) + O ( ‖ δ θ ‖ 2 ) ( 10 ) where Sij = σi Vji . One can define other sensitivities in a similar way but using a different orthogonal basis of MVNn , but the above Sij satisfy an important optimality condition explained in S1 Sect . 11 which asserts that the basis Ui and the corresponding sensitivities Sij are optimal for capturing as much sensitivity as possible in the low order principal components Ui . In view of this we call the Sij the principal control coefficients . Note that the role of the Sij as sensitivities is seen from the following relation which follows from Eq ( 10 ) ( where S = ( Sij ) ) , ∥ δ Θ ∥ = ∥ S · δ θ ∥ + O ( ∥ δ θ ∥ 2 ) . ( 11 ) These sensitivities are relatively easy to calculate using the information in S1 Sect . 11 . In Fig 10 ( C ) we show the Sij for the transversal distribution of the Drosophila circadian clock at the times of the peak of per mRNA and the peak of the nuclear complex of PER and TIM proteins . As we can see , because Sij = σi Vji the coefficients rapidly decrease with the singular values σi , while a few parameters , similar to those with large eigenvector entries , have high coefficients . The likelihood function of a set of time-series observations of a system can be used for parameter estimation , hypothesis testing and other forms of statistical inference . For example , one may wishes to use the likelihood function to estimate parameters of a biological system . Although there is no elegant formula for P ( X _ | X ( t 0 ) ) = P ( X ( t 1 ) , … , X ( t m ) | X ( t 0 ) ) similar to that for P ( Q _ | X ( t 0 ) ) above , we can efficiently calculate it . To do this we derive a Kalman Filter for the pcLNA that is a modification of the Kalman Filter associated with the LNA [35] . This can be used to compute the likelihood function L ( θ ; X _ ^ ) of the system parameters θ with respect to observations X _ ^ = ( X ^ ( t 0 ) , X ^ ( t 1 ) , … , X ^ ( t N ) ) recorded at N times t0 , t1 , … , tN that are noisy linear functions of the species concentrations . This is slightly more general than just calculating P ( X _ | X ( t 0 ) ) because we allow for a measurement equation . The Kalman filter can also be used for forward prediction . We assume the measurement equation , X ^ ( t ) = B X ( t ) + ϵ , ( 12 ) relating the observations X ^ ( t ) to the state variables , X ( t ) . Here B is a transformation matrix ( often simply removing unobserved species or introducing unknown scalings ) and ϵ = ( ϵ1 , … , ϵn ) ∼MVN ( 0 , Σϵ ) the observational error . The pcLNA likelihood can be decomposed as the product L ( θ ; X _ ^ ) = P ( X ^ ( t 0 ) ; θ ) ∏ i = 1 n P ( X ^ ( t i ) | X ^ ( t i - 1 ) ; θ ) . The pcLNA Kalman Filter algorithm , which we describe in more detail in S1 Sect . 15 , uses a recursive algorithm for computing the terms in L ( θ ; X _ ^ ) . The algorithm proceeds by iteration i = 1 , 2 , … and uses Bayes rule to derive the posterior distributions of ( X ( t i - 1 ) | X ^ ( t i - 1 ) ) and a phase correction to obtain g ( si − 1 ) = GN ( μ* ( ti − 1 ) ) and the corrected noise distribution of ( κ ( s i - 1 ) | X ^ ( t i - 1 ) ) . The LNA transition equation ( S1 Eq . ( 4 . 2 ) ) is then used to derive the distribution of ( ξ ( t i ) | X ^ ( t i - 1 ) ) and the LNA ansantz Eq ( 3 ) to obtain ( X ( t i ) | X ^ ( t i - 1 ) ) . The measurement equation Eq ( 12 ) is finally used to obtain the ( i + 1 ) th term of the likelihood function P ( X ^ ( t i ) | X ^ ( t i - 1 ) ) before proceeding to the next iteration . All the distributions obtained in this way are MVN with easily computable parameter values . If the observations are recorded in short time intervals , the phase correction can be omitted in some steps , in which case the algorithm proceeds as in [35] . Computational methods such as those described in [32] and [35] can then be used to perform likelihood-based statistical inference .
All computations have been carried out using MATLAB Release 2016b , The MathWorks , Inc . , Natick , MA , USA . In particular , the empirical CDF plots , ( q-q ) plots , histograms , smooth probability density functions and KS distances are derived using the ecdf , qqplot , histogram , ksdensity , kstest functions of MATLAB and Statistics Toolbox . The computations for the SSA , tau-leap , integration of diffusion and pcLNA simulation algorithms , and the computation of Fisher Information and principal control coefficients for the sensitivity analysis were performed using the PeTTSy software which is discussed in S1 Sect . 14 and is freely available at http://www2 . warwick . ac . uk/fac/sci/systemsbiology/research/software/ . Further details concerning methods are given in S1 Sect . 16 .
We present a comprehensive treatment of stochastic modelling for large stochastic oscillatory systems . Practical algorithms for fast long-term simulation and likelihood-based statistical inference are provided along with the essential tools for a more analytical study of such systems . There is considerable scope for future work in various directions . We expect that these results can be extended to a broader class of systems including those that are chaotic in the Ω → ∞ limit . Our approach should provide the opportunity to develop new methodology for parameter estimation , likelihood-based inference and experimental design in such systems . Finally , there is currently much interest in information transfer and decision-making in signaling systems and our methods provide new tools with which to tackle problems in this area . If system biologists are to reliably use complex stochastic models to provide robust understanding it is crucial that there are analytical tools to enable a rigorous assessment of the quality and selection of these models and their fit to current biological knowledge and data . Our aim in this paper is to contribute to that but the results should be of much broader interest .
|
Many cellular and molecular systems such as the circadian clock and the cell cycle are oscillators that are modelled using nonlinear dynamical systems . Moreover , oscillatory systems are ubiquitous elsewhere in science . There is an extensive theory for perfectly noise-free dynamical systems and very effective algorithms for simulating their temporal behaviour . On the other hand , biological systems are inherently stochastic and the presence of stochastic noise can play a crucial role . Unfortunately , there are far fewer analytical tools and much less understanding for stochastic models especially when they are nonlinear and have lots of state variables and parameters . Moreover simulation is not so effective and can be very slow if the system is large . In this article we describe how to accurately approximate such systems in a way that facilitates fast simulation , parameter estimation and new approaches to analysis , such as calculating probability distributions that describe the system’s stochastic behaviour and describing how these distributions change when the parameters of the system are varied .
|
[
"Abstract",
"Introduction",
"Results",
"Methods",
"Discussion"
] |
[
"invertebrates",
"applied",
"mathematics",
"animals",
"circadian",
"oscillators",
"simulation",
"and",
"modeling",
"algorithms",
"animal",
"models",
"drosophila",
"melanogaster",
"model",
"organisms",
"probability",
"distribution",
"mathematics",
"experimental",
"organism",
"systems",
"chronobiology",
"drosophila",
"research",
"and",
"analysis",
"methods",
"insects",
"probability",
"theory",
"arthropoda",
"approximation",
"methods",
"biochemistry",
"circadian",
"rhythms",
"normal",
"distribution",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"organisms"
] |
2017
|
Long-time analytic approximation of large stochastic oscillators: Simulation, analysis and inference
|
Plants are associated with a complex microbiota that contributes to nutrient acquisition , plant growth , and plant defense . Nitrogen-fixing microbial associations are efficient and well characterized in legumes but are limited in cereals , including maize . We studied an indigenous landrace of maize grown in nitrogen-depleted soils in the Sierra Mixe region of Oaxaca , Mexico . This landrace is characterized by the extensive development of aerial roots that secrete a carbohydrate-rich mucilage . Analysis of the mucilage microbiota indicated that it was enriched in taxa for which many known species are diazotrophic , was enriched for homologs of genes encoding nitrogenase subunits , and harbored active nitrogenase activity as assessed by acetylene reduction and 15N2 incorporation assays . Field experiments in Sierra Mixe using 15N natural abundance or 15N-enrichment assessments over 5 years indicated that atmospheric nitrogen fixation contributed 29%–82% of the nitrogen nutrition of Sierra Mixe maize .
Plants grow in close association with microbial communities that influence plant traits related to nutrient acquisition , plant development , plant defenses , and abiotic stress responses . The root-associated microbiota of plants has been characterized and shown to be much less complex than the microbiota of the surrounding soil , being enriched in Proteobacteria , Bacteroidetes , and Actinobacteria . These microbes are selected in part by plant cell wall features and metabolic cues from host cells [1 , 2] . Characterization of the rhizosphere microbiome associated with 27 modern maize ( Z . mays ) inbred lines also indicated substantial differences in relative abundance of microbial taxa between bulk soil and the rhizosphere , with the maize genotype contributing a small but significant influence on rhizosphere selectivity [3] . Nitrogen-fixing microbial associations with nonlegumes , especially cereals , have been a topic of intense interest for more than a century [4–7] . Nitrogen-fixing endophytes contribute to the nitrogen nutrition of sugarcane in some environments [8–10] , but there is less evidence for the occurrence of efficient diazotrophic associations in other cereals . A 1-year study based on 15N dilution experiments in Miscanthus × giganteus suggested that this perennial bioenergy feedstock can acquire about 16% of its nitrogen from the air [11] . It has also been demonstrated that the model cereal , Setaria viridis , as well as Setaria italica ( foxtail millet ) can acquire a significant amount of fixed nitrogen from associations with Azospirillum brasilense[12 , 13] . Other examples of fixed atmospheric N2 being transferred to cereals include associations between Azoarcus sp . strain BH72 and Kallar grass [14] , Herbaspirillum seropedicae and rice [15 , 16] , and Klebsiella pneumoniae and wheat [17] . Because of its economic importance , the search for diazotrophic associations with maize ( Z . mays ) [18] has been a “holy grail” for decades , and several studies examined the contribution of nitrogen fixation by H . seropedicae and Azospirillum sp . to various maize accessions [19 , 20] . However , it is often difficult in these studies to distinguish the general plant growth–promoting benefits of these diazotrophic bacteria on yield from an actual transfer of fixed nitrogen to host plants . Five techniques are commonly used to evaluate nitrogen fixation: acetylene reduction assays ( ARAs ) , 15N natural abundance , 15N enrichment , 15N2 gas enrichment , and nitrogen balance experiments [21 , 22] . All of these approaches have potential pitfalls , yet very few studies have compared different techniques or conducted assessments over multiple years to evaluate nitrogen fixation in nonlegumes . Triplett suggested that it may be interesting to survey primitive maize landraces from the areas of maize origin to identify maize diazotrophic endophytes [5] . Estrada and colleagues [23] followed this suggestion and examined a landrace of maize in the Sierra Mixe region of Oaxaca , Mexico , and isolated a nitrogen-fixing endophyte from the resident maize landrace . The isolate was tentatively identified as a new species of Burkholderia , but the contribution of atmospheric dinitrogen ( N2 ) to the nitrogen economy of the plant was not tested . This group also reported the isolation of a similar endophyte from field-grown teosinte plants and speculated that the Burkholderia strain might have formed a primitive symbiosis with teosinte that persisted during domestication of maize . We also learned of isolated indigenous landraces of maize in the Sierra Mixe region of Oaxaca that were reportedly grown using traditional practices with little or no fertilizer and speculated that unique microbial community associations , not found in cultivated maize , might have evolved . This indigenous maize landrace is characterized by the extensive development of aerial roots that produce large amounts of mucilage . Mucilage associated with maize underground roots has been previously described [24 , 25] , and it has been suggested that root exudates play a significant role in structuring rhizosphere microbial communities [26 , 27] . Indeed , it has been shown that pea root mucilage can serve as a sole carbon source for some rhizosphere bacteria , including Rhizobium sp . , Burkholderia sp . , and Pseudomonas sp . [28] . Aerial roots at the base of the maize shoot , also known as brace roots or nodal adventitious roots , can often reach the ground and are thought to provide anchorage to prevent lodging but may also contribute to nutrient and water uptake as well as gas exchange [29–31] . However , very little is known about the role of aerial roots that do not reach the ground and the mucilage that they produce . Here , we demonstrate that a Mexican maize landrace can acquire 29%–82% of its nitrogen from the air and that at least some of this N is fixed by diazotrophic bacteria present in the mucilage of aerial roots .
The Sierra Mixe maize varieties cultured locally—referred to as Rojo , Piedra Blanca , and Llano—share similar plant morphologies , growing to a height of over 5 meters and exhibiting extensive aerial root formation at each node . We compared the development of aerial roots in the Sierra Mixe maize with another tall maize variety , Hickory King ( Fig 1A ) . Unlike most modern maize varieties in which aerial root formation ceases after the juvenile-to-adult transition ( arrow , Fig 1A ) , aerial root formation in Sierra Mixe maize continued well after this transition , resulting in a 3- to 4-fold greater number of aerial roots ( Fig 1B ) . Approximately midway through development ( July to September ) , these maize aerial roots secrete significant amounts of mucilage that is rich in arabinose , fucose , and galactose when moisture is available ( Fig 2 ) . The sugars comprise a complex polysaccharide that presumably contributes to the viscosity of the mucilage and may be disassembled to provide monosaccharides to support microbial growth and metabolism . Mucilage produced by underground maize roots also contain high levels of fucose and arabinose [32] , although at approximately one-half the concentration found in aerial root mucilage . The microbiota associated with the underground and aerial roots , stems , and aerial root mucilage of Sierra Mixe maize grown in Sierra Mixe was investigated by amplifying and sequencing of 16S rRNA genes and by shotgun metagenome sequencing . The rhizosphere samples were the most diverse , and among plant samples , the aerial root mucilage had the highest diversity ( S1 Fig ) and a higher relative abundance of bacteroidetes and proteobacteria ( beta and gamma ) compared to other parts of the plants ( S2 Fig ) . Many of the lineages that are overrepresented in mucilage include known plant-associated nitrogen fixers . A comparison of samples based on the total community composition showed a clustering of mucilage samples that were statistically distinct from the rest of the plant and rhizosphere samples ( Fig 3A ) . Clustering of samples based upon the variance-stabilized abundance of sequence variants again indicates that while the rhizosphere samples were distinct and diverse , the mucilage samples were distant from the other plant tissues ( S3 Fig ) . The complete list of all sequence variants identified in the Sierra Mixe maize and soil samples can be found at ( DOI: 10 . 6084/m9 . figshare . 4789759 ) . Additionally , the metagenomic data was searched for homologs of the 6 core nif genes as described by Dos Santos and colleagues [33] . This search revealed the presence of all 6 core nif genes in the metagenomes from mucilage and rhizosphere and only a subset of the 6 core nif genes from stem tissue ( Fig 3B ) . A higher normalized abundance of most nif genes ( except nifD ) in mucilage than stem tissues suggests that the mucilage may be enriched in nitrogen-fixing microbes . To assess the possibility that mucilage harbored a diazotrophic microbial community , the mucilage was tested for nitrogenase activity using ARAs [34] and by incorporation of 15N2 gas . ARA was used to assess leaves , stems , underground roots , aerial roots ( with and without mucilage ) , and mucilage collected from Sierra Mixe maize plants grown either in Sierra Mixe , Mexico , or Madison , USA . No ARA activity was detected in underground roots , leaves , stems , or even aerial roots before mucilage production ( S4 Fig ) . In contrast , significant ARA activity was detected in aerial roots harboring mucilage ( S4 Fig ) and in isolated mucilage from plants grown in either Sierra Mixe or Madison ( Fig 4A ) . The ARA activity in mucilage isolated from Sierra Mixe maize that was grown in Madison suggests either that the Sierra Mixe maize seeds carry an endogenous inoculum of nitrogen-fixing bacteria or that Sierra Mixe maize can recruit adequate nitrogen-fixing bacteria from local environments . Nitrogen fixation by mucilage samples was also measured by direct incorporation of 15N2 in mucilage samples collected from Sierra Mixe maize ( Fig 4B ) . As with Sierra Mixe maize , a wild relative , Z . mays ssp . mexicana ( teosinte ) , also produced extensive aerial roots ( S5 Fig ) but much smaller amounts of secreted mucilage . To test the ability of the teosinte mucilage to support nitrogen fixation , we collected mucilage from several plants of teosinte and measured endogenous nitrogenase activity using ARA . Acetylene reduction was readily observed in teosinte mucilage ( Fig 4A ) , suggesting that production of mucilage that supports nitrogen fixation by an associated nitrogen-fixing microbiota may be an ancient trait of maize and potentially introgressed from Z . mays ssp . mexicana into the Sierra Mixe landrace postdomestication . To assess the mucilage characteristics that support nitrogen fixation , we tested the ability of 2 phylogenetically distinct nitrogen-fixing bacteria to reduce acetylene when inoculated in the mucilage collected from aerial roots of Sierra Mixe maize . Before the experiment , the mucilage was frozen for 2 weeks at –80 °C and thawed , which abolished endogenous nitrogenase activity ( Fig 4C ) . Two nitrogen-fixing bacteria , H . seropedicae , and A . brasilense , showed readily detectable ARA activity when added to the mucilage ( Fig 4C and 4D ) . Bacterial nitrogenase is O2-sensitive and needs to be protected by a low-oxygen ( <5% ) environment or physiological protective mechanisms , as well as an abundant carbon source to derive energy for this process [35] . To determine if mucilage could fulfill these requirements , we measured the free-oxygen concentration in the mucilage of Sierra Mixe maize and teosinte at the depth of 8 mm and found it to be <5% , indicating that the mucilage can provide a microaerobic environment compatible with nitrogen fixation for these bacteria [36] ( Fig 4D ) . Mucilage is also comprised of complex sugars that may be catabolized to provide free sugars—mainly arabinose , fucose , and galactose—capable of supporting bacterial growth and nitrogen fixation . To determine whether these properties of the mucilage are sufficient to support nitrogen fixation , we created an artificial medium mimicking these mucilage properties by using a low-N medium , solidified with 0 . 2% agar , that reduced oxygen concentration to levels almost as low as those found in the mucilage ( S6A Fig ) and supplemented with a mix of free sugars corresponding to the composition of the fully hydrolyzed mucilage carbohydrates . H . seropedicae , A . brasilense , and Burkholderia unamae ( S6B , S6C and S6D Fig ) showed significant ARA activity in this reconstructed mucilage , indicating that the low O2 and free sugars provided by the aerial root mucilage are sufficient to support nitrogen fixation by these diazotrophs . Based on ARA , we can conclude that mucilage from Sierra Mixe maize harbors native diazotrophs and can also support the N2-fixing activity of the exogenously inoculated diazotrophs , H . seropedicae , A . brasilense , and B . unamae . However , these data did not demonstrate that the aerial roots had the capacity to take up and assimilate the fixed N . To test whether atmospheric N2 that was fixed by mucilage-associated diazotrophs could be transferred to and utilized by the Sierra Mixe maize , a more direct 15N2 gas–enrichment experiment was used . Aerial roots , along with their generated mucilage , inoculated with A . brasilense Sp7 exhibited significant 15N2 gas incorporation in comparison to the 14N2 gas-treated roots ( Fig 5 ) , and isotope-ratio mass spectrometry ( IRMS ) analysis confirmed significant enrichment of 15N in chlorophyll ( converted to pheophytin for analysis ) of these roots compared to the negative controls ( Fig 5 ) . The transfer of 15N2 from mucilage to the aerial root tissue and chlorophyll demonstrated the potential of this diazotrophic community to contribute to the nitrogen nutrition of the plant , but a major question of this study is whether the mucilage-associated diazotrophic microbiota served to transfer fixed nitrogen to fulfill , at least in part , the reduced nitrogen requirements of Sierra Mixe maize under field conditions . The contribution of atmospheric nitrogen fixation to Sierra Mixe maize was first estimated in the field using natural abundance 15N measurements [37 , 38] . This method relies on the relative abundance of the stable isotope 15N in the atmosphere and soil , with 15N abundance being more abundant in the soil than in the air . As a consequence , plants that derive N from the atmosphere will exhibit reduced δ15N levels when compared to reference nonfixing plants . In 2006 , samples of Sierra Mixe maize and reference plants from the Asteraceae and Ranunculaceae ( families with no known nitrogen-fixing members ) growing near each other were collected from each of 2 fields . In this preliminary experiment , Sierra Mixe maize δ15N was significantly lower than the reference plants , indicating the assimilation of atmospheric nitrogen ( S1 Table ) . In 2010 , 2011 , and 2012 , the methods from [38 , 39] were used in experiments in Sierra Mixe by sampling reference species of non-nitrogen-fixing plants growing near the Sierra Mixe maize plants and a conventional maize variety , Maiz Blanco Conasupo . In addition to determining δ15N from each maize and reference plant sample , leaf samples of each reference plant were used for 18S rRNA sequence analysis to identify the reference species ( Table 1A ) . The δ15N values of Sierra Mixe maize grown in the field in Sierra Mixe were determined at a single developmental time point in 2010 and at 5 developmental time points in 2011 and 2012 . In 2010 , the δ15N values for the Sierra Mixe maize were significantly lower than those of the reference plant species and of the conventional maize variety , Maiz Blanco Conasupo ( Table 1A ) , indicating that the Sierra Mixe maize was able to derive a significant part of its tissue nitrogen from atmospheric dinitrogen . Similar δ15N values from root and leaf samples also showed that the leaf samples analyzed were representative of the whole plant ( S2 Table ) . In 2011 and 2012 , the δ15N values of Sierra Mixe maize were significantly lower than the reference plants at 4 of the 5 developmental time points , suggesting that Sierra Mixe maize derived a portion of its tissue nitrogen from atmospheric nitrogen ( Table 1B ) . The calculated percent of nitrogen derived from the atmosphere ( %Ndfa ) from the δ15N values in Table 1 ranged between 30% and 80% ( S7 Fig ) . The method of using natural abundance 15N and other species as reference plants to calculate %Ndfa is potentially limiting because of differences in root and shoot growth and phenology of reference and test plants and the limited range of δ15N found in soils . An alternative method , 15N enrichment , is similar to the natural abundance methods but enriches soil 15N by the addition of 15N fertilizer , thereby increasing the difference between soil and atmospheric δ15N and assay sensitivity . Several direct comparisons have indicated that both methods can give comparable but slightly different results [38 , 40] . Another method of estimating N2 fixation is the “N Difference” method , which determines the difference between the total N content of an N2-fixing plant and the total N content of a reference nonfixing plant . Total N is calculated by multiplying total N content ( % ) in a specific plant sample and the total biomass ( kg/ha or kg ) produced by the plant . The %Ndiff is calculated as described in Materials and methods . In 2016 and 2017 , we assessed atmospheric nitrogen fixation using the 15N-isotope-enrichment method ( 1%–10% enrichment ) at 3 vegetative growth stages—V9 , V12 , and Tassel—[41] in a random complete block design ( 5 replicates ) in 3 low-N Sierra Mixe fields and at Tassel stage in 2017 in the same fields using the same design ( Table 2 ) . Field 3 ( with a history of 0–1 year of maize production ) yielded significant differences in Atom% 15N only for SM2 at the Tassel stage in 2016 , but shoot N was significantly different in both 2016 and 2017 . In 2016 , Sierra Mixe maize landrace varieties exhibited significantly lower Atom%15N levels than the reference plants in Field 4 , ( with a history of 1–2 years of maize production ) at Tassel and at V9 in 2016 , and at Tassel in both years for shoot N . In both 2016 and 2017 , Sierra Mixe maize landraces in Field 5 ( over 3 years of continuous maize production ) exhibited significantly lower Atom%15N and shoot N at all stages sampled , indicating a significant level of atmospheric nitrogen fixation in those experiments . The calculated %Ndfa ranged between 31% and 55% , and Ndiff ranged from 29%–82% . The correlation between Ndfa and Ndiff was 0 . 55 and 0 . 44 ( P < 0 . 01 ) across locations in 2016 and 2017 , respectively . Significant measures of N2 fixation were detected in 4 of 6 experiments by Atom% δ15N N determinations ( Ndfa ) and in 6 of 6 experiments by Ndiff determinations . Root and shoots exhibited significant differences in biomass , height , and stem diameter between control hybrids and local landraces ( S4 Table ) . Soil analyses ( S5 Table ) showed that all sites were depleted in soil nitrogen yet produced a crop greater than 2 , 000 kg/ha . It is possible that differences between microbiota associated with the 3 fields in different stages of crop rotation ( 0 to >4 years of continuous maize production ) account for differences observed among fields and between years , but further research is needed to answer this question .
We have demonstrated that the mucilage associated with the aerial roots of Sierra Mixe maize can support a complex diazotrophic microbiota enriched for homologs of genes encoding nitrogenase subunits that harbor active nitrogenase activity , and that nitrogen is transferred efficiently from the nitrogen-fixing bacteria to the host plant tissues . Collectively , over several years and locations , the 15N natural abundance and 15N enrichment results of 2 selections of a Sierra Mixe indigenous maize landrace suggest that its nitrogen nutrition when grown in its native environment is partially fulfilled by fixation of atmospheric nitrogen . Nitrogen fixation is a particularly difficult phenotype to evaluate , as all the techniques available are prone to artifacts and can give different estimates for nitrogen fixation [42 , 43] . In this study , we used several classical techniques with Sierra Mixe maize and have shown over multiple locations and years that Sierra Mixe maize can fix nitrogen at rates ( 29%–82% ) not previously reported , to our knowledge , in maize . This study also revealed a new and important function for aerial roots and the mucilage they produce besides preventing lodging or water uptake [30] . This role in nitrogen fixation is probably the most important one for aerial roots that do not reach the ground . It will be interesting to explore if aerial roots produced by other cereals such as sorghum can perform a similar function [44] . We cannot rule out that diazotrophic activity in other parts of Sierra Mixe maize may also contribute to the acquisition of reduced nitrogen from the atmosphere . The developmental timing of the appearance of fixed atmospheric N2 in Sierra Mixe maize plants ( Tables 1 and 2 ) before the extensive development of aerial roots suggests that there may , indeed , be additional sites of nitrogen fixation . However , we have not detected any significant nitrogenase activity outside of the aerial root mucilage . The genetic basis of the trait or the source of the microbial inoculum , which may be either environmental or seed-borne , are unresolved . The observation that a teosinte species ( Z . mays ssp . mexicana ) also exhibits a similar diazotrophic activity in aerial root–associated mucilage suggests that this is an ancient trait that may have been introgressed and amplified in the Sierra Mixe landrace . It will be important , in the future , to determine the genetic basis of the trait , the identity of associated microbial diazotrophs , and the mechanisms of microbial recruitment . This research , together with other published research [5 , 18 , 23] , suggests new avenues for research into potentially novel mechanisms of biological N2 fixation in maize . This could have a significant impact on maize crop productivity and nitrogen use efficiency , particularly in regions of the world where agriculture is characterized by poor soil nutrition .
Sierra Mixe maize seeds were obtained in Sierra Mixe region of Oaxaca , Mexico , from an open pollinated population . Z . mays ssp . mexicana ( teosinte ) , LH123HT , Mo17 , PHG39 , LH82 , PH207 , and B73 seeds were obtained from USDA National Plant Germplasm System ( NPGS ) ( accessions Ames 8083 , PI601079 , PI558532 , PI600981 , PI601170 , PI601005 , and PI550473 , respectively ) . Maize line Hickory King was obtained from Victory Seeds ( accession 3140041 ) . Tornado-F21 and H-377 were obtained from Semillas Ceres in Oaxaca . SM1 and SM2 are different selections from the Sierra Mixe landrace . SM1 is a uniform population based on kernel size , shape , and color , and SM2 is a heterogeneous population representing the landrace . Biological materials were accessed and utilized under an Access and Benefit Sharing Agreement between the Sierra Mixe community and BioN2 , Inc . , and with permission from the Mexican government . An internationally recognized certificate of compliance under the Nagoya Protocol ( ABSCH-IRCC-MX-207343-3 ) has been issued for such activities . A . brasilense Sp7 and B . unamae MTI-641 were kindly provided by Dr . G . Alexandre ( University of Tennessee , Knoxville , USA ) and Dr . A . Hirsch ( University of California , Los Angeles , USA ) , respectively . H . seropedicae Z152 ( ATCC 35894 ) was provided by the ATCC ( http://www . atcc . org/ ) . Bacteria were grown in liquid BSE medium . The rhizosphere and plant tissues that include stem , leaf , aerial roots , underground roots , and mucilage of Sierra Mixe maize were sampled during seasons 2010 , 2011 , and 2012 from Fields 1 and 2 in Sierra Mixe . For plant endophyte analysis , tissues were surface sterilized by rinsing with mqH2O , shaken gently in 70% ethanol for 5 minutes , placed into 1% hypochlorite bleach , gently stirred for 10 minutes , rinsed 3 times in mqH20 , and dried in a laminar flow cabinet . Roots and stems were also dissected to remove epidermal tissues before extraction . For seed endophyte analysis , embryo and endosperm of Sierra Mixe , Hickory King , and B73 were withdrawn from the seeds by hand using a razor blade in a laminar flow cabinet and were collected in 1 . 5 ml sterile microcentrifuge tubes . For this study , the rhizosphere was defined as a layer of soil covering the outer surface of the root system that could be washed from roots in a buffer/detergent solution . Roots were first separated and shaken to remove loosely adhering soil . All soils and plant material samples were used immediately for DNA extraction . Soil fertility analysis , which included physical parameters , soil reaction , and salinity , was performed in AgroLab ( Pachuma , Mexico ) . The number of nodes with aerial roots was monitored weekly ( greenhouse ) or after 14 weeks ( field ) . The total number of aerial roots was quantified after 14 weeks . The disappearance of leaf wax and appearance of trichomes were monitored weekly to determine the transition between juvenile stage and adult stage in Sierra Mixe and Hickory King maize . For experiments in the greenhouse , seeds of Sierra Mixe of Fields 1 and 2 and Hickory King were surface sterilized and germinated as described previously . After 1 week , the seedlings are transplanted in 40-liter pots filled with a mix of sand and perlite ( v:v ) and grown in a high-ceiling greenhouse at the Biotron facility ( University of Wisconsin , Madison , USA ) . Plants were watered twice a day for 2 minutes with half-strength of Hoagland solution . For experiments in the field , 3 independent plots of 20 plants per genotype were planted , with 3 border rows ( B73 ) between each genotype . Sierra Mixe and Teosinte plants that were grown in Madison for the ARA were planted in the same field at the same time . This experiment was replicated in 3 different field plots . Glycosyl composition analysis was performed by combined gas chromatography/mass spectrometry ( GC/MS ) of the per-O-trimethylsilyl ( TMS ) derivatives of the monosaccharide methyl glycosides produced from the sample by acidic methanolysis . Methyl glycosides were first prepared from dry mucilage samples by methanolysis in 1 M HCl in methanol at 80 °C ( 18–22 hours ) , followed by re-N-acetylation with pyridine and acetic anhydride in methanol ( for detection of amino sugars ) . The samples were then per-O-trimethylsilylated by treatment with Tri-Sil ( Pierce ) at 80 °C ( 0 . 5 hours ) . GC/MS analysis of the TMS methyl glycosides was performed on an HP 6890 GC interfaced to a 5975b MSD , using an All Tech EC-1 fused silica capillary column ( 30 m × 0 . 25 mm ID ) . The analysis was performed at the Complex Carbohydrate Center Research Center ( CCRC ) of University of Georgia , Athens , USA . DNA ( 80–150 ng μl−1 ) was extracted from 100 mg of the rhizosphere plant tissues using a DNA isolation kit ( Mo Bio Laboratories , Carlsbad , USA ) . The PCR control for microbial DNA isolation was performed on 16S rRNA genes . PCR was performed using eubacterial primers 27F ( 5′-AGAGTTTGATCCTGGCTCAG-3′ ) and 1492R ( 5′-GGTTACCTTGTTACGACTT-3′ ) , and the product was approximately a 1 , 450 bp fragment . Amplification was carried out with 1 μM of each primer in 3 mM MgCl2 , 20 μM of each dNTP , 1 . 25 units of Taq polymerase ( Promega ) in a total volume of 20 μl of 1X reaction buffer ( Promega , Madison , USA ) . PCR conditions included an initial denaturation at 95 °C for 3 minutes followed by 35 cycles of denaturation at 94 °C for 1 minute , annealing at 56 °C for 1 minute , and elongation at 72 °C for 1 . 5 minutes , with a final elongation at 72 °C for 7 minutes . DNA was resolved using an agarose gel run at 100 V for 30 minutes for analysis of total DNA and amplification of PCR products , respectively . Gels were visualized by ethidium bromide staining under UV light in a gel documentation system . The products obtained were purified with a NucleoSpin Gel extraction kit ( Clontech , Palo Alto , USA ) . 16S rRNA gene PCR and sequencing of rhizosphere and plant tissues were carried out using the Caporaso protocol [45] . We extended the Caporaso approach [46] to a dual barcode scheme for each sample and replaced the Golay barcodes with a different set of Illumina-compatible barcodes that were designed to balance base composition and tolerate up to 4 sequencing errors in barcode sequences . The forward primer used was AATGATACGGCGACCACCGAGATCTACAC[Barcode]TATGGTAATTGTGTGCCAGCMGCCGCGGTAA , and the reverse primer was CAAGCAGAAGACGGCATACGAGAT[Barcode]AGTCAGTCAGCCGGACTACHVGGGTWTCTAAT . The barcodes used were designed to allow pooling of multiple samples within a single MiSeq run . Ten cycles of PCR with barcoded primers were performed at low annealing temperature ( 55 °C ) ; samples were then pooled and cleaned using a Qiagen column to remove the unincorporated primers . At this stage , an additional 10 or 20 cycles of PCR were performed on the pool using the Illumina paired-end flowcell primers with a higher annealing temperature ( 65 °C ) . The resulting PCR product was subjected to QC with an Agilent Bioanalyzer and estimated concentration using KAPA Biosystems qPCR kit . The samples were diluted to the appropriate loading concentration for a MiSeq run , spiked with 25% phiX control library , and sequenced using an Illumina MiSeq instrument with the manufacturer’s standard 150 nucleotides paired-end dual-index sequencing protocol and the custom sequencing primers . The Illumina sequences were obtained from 2 MiSeq runs ( 2X150 bp paired-end ) and were demultiplexed using a custom script ( https://figshare . com/s/04997ae7f7d18b53174a ) . The 822 , 804 reads thus obtained were trimmed to a Phred-equivalent of 20 and filtered for adaptor contamination using BBDuk ( of the BBTool packages https://sourceforge . net/projects/bbmap/ ) . Because of the low quality of the reverse mate pair , reads of the forward mate were used in the analysis . The reads were then preprocessed and analyzed using DADA2 [47] . Prescribed standard filtering parameters were used , such as PhiX contamination check and removal of reads with more than 2 errors or ambiguous bases or with an expected error greater than 2 . Chimeras were identified and removed using the removeBimeraDenovo function of DADA2 . The clean reads were then collapsed into sequence variants and classified using RDP training set ( version 14 ) . The sequence variants that were classified as chloroplast or mitochondria were removed from further analyses . From samples with library size ranging from 84 to 20 , 597 reads ( mean of 4 , 703 ) , 995 unique sequence variants were identified . Alpha and Beta diversity metrics were generated using the Phyloseq 3 . 4 . 2 R packages [48] . Alpha diversity was calculated using Shannon and Simpson indices . Additionally , a PCoA plot based upon Bray-Curtis dissimilarity matrix was used to visualize the differences in samples . The sequence variant table was used to generate a heat map following the variance-stabilizing transformation in DESeq2 [49] . Permanova tests were run using the adonis function from the vegan 3 . 4 . 3 . R package ( https://CRAN . R-project . org/package=vegan ) performed on the NMDS ordination . Illumina sequencing libraries from the same DNA extractions as above were made using an adaptation of the Nextera transposase-based library construction method with multiplex barcoding . Samples were then sequenced on the MiSeq and HiSeq instruments . Illumina sequences thus obtained were demultiplexed and trimmed using Trimmomatic ( ver 0 . 33 ) [50] with the following parameter: Illuminaclip 2:30:10 , Headcrop:15 , Leading:20 , Trailing:20 , Sliding window:4:20 , and Minlen:100 . The reads were then screened for PhiX and maize sequences ( genomic , chloroplast , and mitochondrial ) using Bowtie2 aligner [51] against the PhiX genome ( Genbank acc# NC_001422 . 1 ) and Z . mays cultivar B73 draft genome ( RefSeq assembly acc# GCF_000005005 . 2 ) . The clean reads were assigned taxonomy using Kaiju [52] with the nr database . To calculate the beta diversity of the samples , we used Phylosift ( ver 1 . 0 . 1 ) [53] , which identifies and places reads matching 37 conserved phylogenetic marker genes on a reference tree . From these placements , an Edge-PCA analysis [54] was carried using Guppy [54] . Peptide sequences of the 6 core nif genes ( nifH , nifD , nifE , nifK , nifN , nifB ) and alternate nitrogenase ( anfG , vnfG ) from known diazotrophs as previously published [33] were retrieved from GenPept as a reference . A multiple-sequence alignment of these sequences was generated as a reference alignment using ClustalW2 [55] . A blast search ( E-value < 0 . 001 ) of 6 frame-translated metagenomic reads was conducted against these reference sequences . The hits were then aligned against the multiple sequences alignment of reference using clustal-Omega [56] followed by generation of phylogenetic trees for every individual nif gene , using Fasttree2 . 1 [57] with a WAG model of amino acid evolution and gamma20 likelihood . The reads were assigned as belonging to the nif genes if they were inside the clade of the reference sequences . Each read that had significant similarity to one of the core nif genes was further analyzed by phylogenetic analysis to confirm its assignment as one of the 6 core nif genes . The counts of nif genes thus obtained were normalized by recA counts ( determined using recA TIGRFAM HMM [58] with HMMER3 and an E-value cutoff of e−10 ) . For ARA with the mucilage , 2 ml of freshly collected mucilage from 1 or 2 aerial roots ( Sierra Mixe ) or several plants ( teosinte ) grown in the field were introduced in 14 . 5 ml vials ( Wheaton , Millville , USA ) that were tightly closed . For ARA with added bacteria , A . brasilense , H . seropedicae , and B . unamae were grown in BSE medium for 48 hours . Then , bacteria were collected by centrifugation ( 5 minutes , 4 , 000 × g ) and suspended in the Fahraeus medium . The 14 . 5 ml vials , each containing 5 ml of mucilage previously stored for several months at –20 °C to reduce endogenous nitrogen-fixing bacteria , were inoculated with the bacterial suspension at a final OD600nm = 0 . 01 . Control tubes were prepared either without bacteria or with 5 ml of Fahraeus medium instead of mucilage . Then , 850 μl of acetylene ( Airgas ) was injected into each vial . OD600nm was measured for each tube after 72 hours . For both conditions , controls without acetylene were performed in parallel . For ARA with aerial roots , 1 aerial root without mucilage was introduced in each 14 . 5 ml vial ( 10 replicates ) . One ml of acetylene ( Airgas ) was injected into each vial . For ARA with seedlings , Sierra Mixe seedlings were inoculated with A . brasilense and grown for 3 weeks . Plants were then transferred to 500 ml jars , and 50 ml of acetylene was injected in each jar . For ARA with underground roots , pieces of roots ( about 10 cm long ) were collected from plants grown in pots and introduced into 14 . 5 ml vials ( 3 replicates ) . Ethylene quantification was made by injecting 1 ml of the air phase , sampled after 72 hours , on a gas chromatography ( GC-2010 Shimadzu ) equipped with a Rt-Alumina BOND/KCL column ( Restek ) . The enrichment of mucilage in 15N atom was achieved by removing 4 ml of headspace gas and replacing it with 4 ml of either 15N2 ( Sigma-Aldrich ) or 14N2 nitrogen gas directly into a vial containing 1 . 0 mL of mucilage . Mucilage was collected from Sierra Mixe maize plants grown in Sierra Mixe and stored at 4 °C for up to 2 weeks between sampling and the determination of 15N2 assimilation . The mucilage samples were incubated at 37 °C for 0 and 70 hours in the presence of 15N2 . 15N2 assimilation was stopped by freezing the mucilage samples at −20 °C . The samples were then freeze-dried and weighed . The 15N2 analysis in the mucilage samples was performed at the UC Davis Stable Isotope Facility ( Davis , USA ) and the UW-Madison Soil Science Facility ( Madison , USA ) . Statistical analysis was performed using SYSTAT version 10 ( Chicago , USA ) . For measurement in collected mucilage , 2 ml of mucilage was introduced in a 15 ml tube . The probe ( robust oxygen mini probe , Pyroscience ) was introduced 8 mm deep in the mucilage and oxygen measurements performed until stabilization of the signal was observed . Control corresponds to free-oxygen concentration in the liquid Fahraeus medium . One-point calibration was made in aerated water , as advised by the manufacturer . Aerial roots were collected from Sierra Mixe maize grown at the Biotron greenhouse facility ( University of Wisconsin , Madison , USA ) . Mucilage was generated from each of these aerial roots by incubating them in 5 ml of water at room temperature for 48 hours . Mucilage , along with the aerial roots , was inoculated with A . brasilense Sp7 . Then , 10%–15% ( v/v ) of 15N2 gas was pumped into the vials , and the samples were incubated at 30 °C for 48 hours . After incubation of mucilage alone , or aerial roots alone , pheophytin extracted from these aerial roots was subjected to IRMS analysis . To obtain pheophytin , chlorophyll was extracted from aerial roots and converted to pheophytin by acid treatment , following as described [59] . 14N2-treated mucilage and aerial roots were used as negative controls . The proportion ( % ) of nitrogen derived from biological nitrogen fixation ( %Ndfa ) was estimated from the 15N natural abundance ( expressed in delta units , ‰ ) of the Sierra Mixe maize ( δ15Nfixing plant ) and that of the reference plant species ( δ15Nref ) . In each of the 2011 and 2012 field seasons in Sierra Mixe , 90–114 individual maize samples ( depending on the year ) and 270 reference plant samples , representing 8–10 species ( depending on the year ) of non-nitrogen-fixing plants , were analyzed . For the single time point in 2010 , 12 individual maize samples and 33 reference plant samples , representing 8 species of non-nitrogen-fixing plants , were analyzed . The reference plant species in the field were identified using universal 18S PCR analysis from DNA sampled using FTA Plant Saver card ( GE Life Sciences , Pittsburg , USA ) simultaneously with tissue samples collected for 15N analysis . PCRs were performed using Sigma’s Extract-N-Amp Plant PCR Kit according to the manufacturer for sequencing and BLASTN comparison . For 15N natural abundance , the third-youngest leaf of Sierra Mixe maize or reference plants was collected from Field 3 and 4 from the second to the sixth month postplanting and analyzed for N-isotope composition . Total organic nitrogen was determined by Kjeldahl digestion followed by steam distillation . Analysis for natural 15N abundance was carried out as described by Bremer and van Kessel [38] . The 15N analysis was performed at UCD Stable Isotope Facility ( http://stableisotopefacility . ucdavis . edu/13cand15n . html ) . The percentage of nitrogen derived from nitrogen fixation ( %Ndfa ) was calculated as follows: %Ndfa=δ15Nreference−δ15Nfixingplantδ15Nreference−BX100 , where “δ15N” is stable nitrogen isotopes , “ref” is the value from non-N-fixing reference plants , “fixing plant” is Sierra Mixe maize , and “B” is the 15N abundance in the air , assumed to be 0 . 0‰ . In 2016 , 3 locations were chosen: Field 3 , land that had not been planted to crops for over 10 years; Field 4 , land that had maize for 1 year; and Field 5 , land with continuous maize . A randomized complete block design trial was established at each site , with 5 replicates with 4 varieties . Each plot consisted of 6 matas surrounded by a common border of SM2 on all sides and a double border on outside rows . A mata is the traditional planting design in the Sierra Mixe region , similar to a hill plot in which multiple plants are seeded together . Each mata was planted with 5 seeds and thinned to 3 seeds for 18 plants per plot . Matas were planted in a grid 80 × 100 cm from each other . 15N was applied at a dose of 0 . 36 grams per plot in a liquid solution , reaching the desired enrichment of 1% for all 3 fields , with 50 ml added per mata . In 2017 , the same 3 locations were planted with same design and entries , except Field 3 consisted only of H377 and SM2 entries . In all 3 fields , 15N was applied at a dose of 0 . 95 grams per plot , reaching the desired enrichment of greater than 1% . A solution was spread evenly over each plot using a garden watering can , such that the whole experimental area received an equal amount of enriched 15N . Plants were covered with plastic bags at the time of application ( V5 ) to ensure that 15N was not directly applied to the leaves and that the 15N was uniformly available to all plants . Soil samples were taken from a 0–60 cm depth in each plot , blended , and sent for analysis at UC Davis Soil lab . Means were calculated across locations . In 2016 , at V9 and V12 , 1 mata ( 3 plants ) was sampled; and at Tassel , 4 matas ( 12 plants ) were sampled . In 2017 , a single sampling of 6 matas ( 18 plants ) was sampled at Tassel . For each sampling , plants were dug out to include all roots . Because of the high rainfall ( 2 , 100 mm concentrated in the growing season from June to October ) , roots were shallow for both reference and test varieties . Each plant was photographed , and data were recorded for the number of plants , plant height , the total fresh weight of shoots , and roots and stem diameter . Whole plants were chopped , ground , subsampled , and dried in an oven to record total dry weight for shoots and roots . Well-blended subsamples were taken and shipped to Davis to measure Total N and 15N . Total nitrogen and Atom% 15N were determined for each plot at each time point for the shoot . Total organic nitrogen was determined by Kjeldahl digestion followed by steam distillation . The 15N analysis was performed at UCD Stable Isotope Facility ( http://stableisotopefacility . ucdavis . edu/13cand15n . html ) . %Ndfa was calculated using the 15N-enrichment method [40] as %Ndfa= ( 1−Atom%15Nexcess ( N2fixer ) Atom%15Nexcess ( reference ) ) X100 . Atom% excess is calculated with sample value obtained from the UC Davis Stable Isotope Facility– 0 . 37 ( n in air ) . Nitrogen difference ( %NDiff ) method was calculated as %Ndiff=Nkg/ha ( N2fixer ) -Nkg/ha ( reference ) kgN/ha ( reference ) X100 . The average of Tornado F21 and H377 was used as the reference to calculate %Ndfa and %Ndiff . For Ndiff calculations , the area harvested was adjusted based on matas harvested at each time point . Data were analyzed using the R lme4 package . Data were checked for outliers and subjected to ANOVA and mean separation using Least Significant Difference ( p = 0 . 05 ) for each location . LSDs were calculated only when ANOVA F-tests were significant at p = 0 . 05 . For %Ndfa and %Ndiff , single-degree-of-freedom contrasts were calculated to compare test varieties to the mean of the reference varieties ( P > 0 . 05 ) . Pearson correlation coefficients were calculated between %Ndfa and %Ndiff .
|
Nitrogen is an essential nutrient for plants , and for many nonlegume crops , the requirement for nitrogen is primarily met by the use of inorganic fertilizers . These fertilizers are produced from fossil fuel by energy-intensive processes that are estimated to use 1% to 2% of the total global energy supply and produce an equivalent share of greenhouse gases . Because maize ( Zea mays L . ) is a significant recipient of nitrogen fertilization , a research goal for decades has been to identify or engineer mechanisms for biological fixation of atmospheric nitrogen in association with this crop . We hypothesized that isolated indigenous landraces of maize grown using traditional practices with little or no fertilizer might have evolved strategies to improve plant performance under low-nitrogen nutrient conditions . Here , we show that for one such maize landrace grown in nitrogen-depleted fields near Oaxaca , Mexico , 29%–82% of the plant nitrogen is derived from atmospheric nitrogen . High levels of nitrogen fixation are supported , at least in part , by the abundant production of a sugar-rich mucilage associated with aerial roots that provides a home to a complex nitrogen-fixing microbiome .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"ecology",
"and",
"environmental",
"sciences",
"microbiome",
"chemical",
"compounds",
"split-decomposition",
"method",
"microbiology",
"plant",
"physiology",
"diazo",
"compounds",
"organic",
"compounds",
"multiple",
"alignment",
"calculation",
"cereal",
"crops",
"plant",
"science",
"model",
"organisms",
"experimental",
"organism",
"systems",
"crops",
"molecular",
"biology",
"techniques",
"plant",
"ecology",
"plants",
"microbial",
"genomics",
"research",
"and",
"analysis",
"methods",
"grasses",
"crop",
"science",
"artificial",
"gene",
"amplification",
"and",
"extension",
"medical",
"microbiology",
"maize",
"chemistry",
"molecular",
"biology",
"agriculture",
"rhizosphere",
"nitrogen",
"fixation",
"eukaryota",
"plant",
"and",
"algal",
"models",
"organic",
"chemistry",
"ecology",
"computational",
"techniques",
"polymerase",
"chain",
"reaction",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"genomics",
"plant-environment",
"interactions",
"organisms"
] |
2018
|
Nitrogen fixation in a landrace of maize is supported by a mucilage-associated diazotrophic microbiota
|
Mycobacterium tuberculosis , the etiological agent of TB , possesses a cholesterol catabolic pathway implicated in pathogenesis . This pathway includes an iron-dependent extradiol dioxygenase , HsaC , that cleaves catechols . Immuno-compromised mice infected with a ΔhsaC mutant of M . tuberculosis H37Rv survived 50% longer than mice infected with the wild-type strain . In guinea pigs , the mutant disseminated more slowly to the spleen , persisted less successfully in the lung , and caused little pathology . These data establish that , while cholesterol metabolism by M . tuberculosis appears to be most important during the chronic stage of infection , it begins much earlier and may contribute to the pathogen's dissemination within the host . Purified HsaC efficiently cleaved the catecholic cholesterol metabolite , DHSA ( 3 , 4-dihydroxy-9 , 10-seconandrost-1 , 3 , 5 ( 10 ) -triene-9 , 17-dione; kcat/Km = 14 . 4±0 . 5 µM−1 s−1 ) , and was inactivated by a halogenated substrate analogue ( partition coefficient<50 ) . Remarkably , cholesterol caused loss of viability in the ΔhsaC mutant , consistent with catechol toxicity . Structures of HsaC:DHSA binary complexes at 2 . 1 Å revealed two catechol-binding modes: bidentate binding to the active site iron , as has been reported in similar enzymes , and , unexpectedly , monodentate binding . The position of the bicyclo-alkanone moiety of DHSA was very similar in the two binding modes , suggesting that this interaction is a determinant in the initial substrate-binding event . These data provide insights into the binding of catechols by extradiol dioxygenases and facilitate inhibitor design .
Mycobacterium tuberculosis , the leading cause of mortality among bacterial pathogens , infects one-third of the human population and is responsible for approximately 2 million deaths annually . The global threat of TB has risen alarmingly due to two factors: the bacterium's deadly synergy with HIV [1] and the emergence of multidrug-resistant strains , including extensively drug-resistant strains ( XDR-TB ) that are virtually untreatable with current chemotherapies [2] . An important factor that contributes to the disease's prevalence is the pathogen's unusual ability to survive for long periods of time , and even to replicate , in the macrophage [1] . The mechanisms by which M . tuberculosis persists in the macrophage remain largely unknown , but such mechanisms are good targets for novel therapeutic agents . A suite of genes critical for survival of M . tuberculosis in the macrophage [3] was recently discovered to be involved in cholesterol degradation [4] . As in the aerobic bacterial degradation of other steroids , the core 4-ringed structure is degraded via opening of ring B with concomitant aromatization of ring A . The resulting phenolic metabolite is hydroxylated , yielding a catechol , 3 , 4-dihydroxy-9 , 10-seco-nandrost-1 , 3 , 5 ( 10 ) -triene-9 , 17-dione ( DHSA ) . HsaC catalyzes the meta-cleavage of DHSA to produce 4 , 9-DSHA ( 4 , 5-9 , 10-diseco-3-hydroxy-5 , 9 , 17-trioxoandrosta-1 ( 10 ) , 2-diene-4-oic acid; Figure 1 ) . Recent work by Pandey and Sassetti [5] indicates that in vitro , the pathogen uses different parts of the cholesterol molecule for energy and the biosynthesis of phthiocerol dimycocerosate ( PDIM ) , a virulence-associated lipid , respectively . Using a mutant defective in the mce4-encoded cholesterol transporter [6] , Pandey and Sassetti further demonstrated that cholesterol uptake is essential for persistence in the lungs of chronically infected mice and for growth in IFN-γ-activated macrophages that predominate during the chronic phase of the illness . However , this deletion impaired in vitro growth on cholesterol only modestly . HsaC shares ∼40% amino acid sequence identity with BphC ( EC 1 . 13 . 11 . 39 ) , a well-characterized type I extradiol dioxygenase that cleaves 2 , 3-dihydroxybiphenyl ( DHB ) and that is potently inhibited by 2′ , 6′-diCl DHB [7] , a polychlorinated biphenyl metabolite ( Figure 1 ) . Extradiol dioxygenases typically utilize Fe ( II ) in a 2-His 1-carboxylate facial triad coordination environment to catalyze the cleavage of catechols and their analogues . In the proposed mechanism , based on biochemical , spectroscopic , kinetic and structural studies [8] , [9] , [10] , the catecholic substrate binds first to the enzyme's Fe ( II ) center in a bidentate manner , displacing two solvent ligands . Thus activated , the ferrous center binds O2 , leading to the formation of an Fe ( II ) –bound alkylperoxo intermediate . The latter undergoes heterolytic O-O bond cleavage and Criegee rearrangement involving 1 , 2-alkenyl migration to produce a lactone intermediate . Hydrolysis of the latter affords the ring-cleaved product . Several of the proposed intermediates were recently substantiated in structural studies of homoprotocatechuate 2 , 3-dioxygenase ( HPCD ) and a slow substrate , 4-nitrocatechol [11] . Nevertheless , some steps of the catalytic cycle remain unclear , including the multi-step binding of the catecholic substrate [12] . We report herein studies of HsaC from M . tuberculosis H37Rv . An hsaC-null gene deletion mutant was generated and tested in liquid culture and in animal models to assess the role of HsaC in cholesterol degradation and pathogenicity . The specificity of the enzyme was investigated , and crystal structures of HsaC were obtained in its substrate-free form and in complex with the steroid metabolite , DHSA . The results provide insights into the binding of catechols to extradiol dioxygenases and the role of cholesterol metabolism in pathogenesis .
To characterize HsaC from M . tuberculosis H37Rv , we anaerobically purified the enzyme to >99% apparent homogeneity from a recombinant E . coli strain . Purified enzyme contained 0 . 92 equivalents of iron . To stabilize HsaC for steady state kinetic assays , the enzyme was diluted in 20 mM HEPES , 80 mM NaCl , pH 7 . 5 supplemented with 5% t-butanol , 2 mM dithiothreitol , 0 . 1 mg/ml bovine serum albumin and stored on ice under an inert atmosphere . Due to the oxidative inactivation of both the enzyme and DHSA in air-saturated buffer , kinetic studies were performed using buffer equilibrated with 5% oxygen in nitrogen to obtain better quality data . Steady-state kinetic studies revealed that HsaC has 90-times greater specificity for the steroid metabolite , DHSA , over DHB , the preferred substrate of BphC ( Table 1 ) . To facilitate further kinetic characterization of HsaC , we designed a substrate analogue , DHDS ( Figure 1 ) , which incorporated potentially important features of DHSA including the methyl group on the catecholic ring and a saturated 2-carbon bridge between the two ring systems . The specificity of HsaC for DHSA was 10-times greater than for DHDS . 2′ , 6′-diCl DHB , a PCB metabolite that potently inhibits ( 7±1 nM ) and oxidatively inactivates BphC [7] , and 4-Cl DHDS , a chlorinated substrate analogue were cleaved very slowly by HsaC ( partition ratios<50 ) . While the Km values of HsaC for the PCB metabolite and the chlorinated steroid metabolite were ∼1 , 000-fold greater than those of BphC , both compounds have clear potential as competitive inhibitors of the mycobacterial dioxygenase . The steady-state utilization of O2 by HsaC was evaluated in the presence of DHDS due to the ease of preparation of this compound . We anticipate that the reactivity of the enzyme with O2 will be very similar in the presence of DHDS and DHSA as the two compounds have similarly substituted catecholic rings . The apparent KmO2 of HsaC was 90±20 µM: 13-fold less than that of BphC [13] and nearly 3-times less than the concentration of O2 in air-saturated buffer . Nevertheless , the specificity of HsaC for O2 is only 5-times less than that of BphC ( 0 . 20±0 . 01 µM−1 s−1 vs . 1 . 0±0 . 1 µM−1 s−1 ) . Extradiol dioxygenases are subject to oxidative inactivation during catalytic turnover [14] . Accordingly , we investigated the susceptibility of HsaC to inactivation during the steady-state cleavage of each of the catecholic substrates using the partition ratio , the amount of substrate consumed per mole of enzyme inactivated . As reported for BphC , HsaC was more susceptible to inactivation by poorer substrates ( Table 1 ) . Nevertheless , the observed partition ratios are more than 2 orders of magnitude less than what has been reported for other extradiol dioxygenases for their preferred substrates [14] , [15] . Finally , 2′ , 6′-diCl DHB inactivated HsaC with a partition ratio similar to that in BphC ( <50 ) . Crystal structures of HsaC were obtained in its substrate-free form and in complex with DHSA at resolutions of 2 . 0 and 2 . 2 Å , respectively . The asymmetric unit of the crystals contains two well-ordered molecules . Crystallographic four-fold symmetry of the two molecules in the asymmetric unit indicates the enzyme is octameric , like BphC . Crystallographic statistics are summarized in Table 2 . The overall fold of HsaC is that of a two-domain type I extradiol dioxygenase , with the structure most closely resembling that of BphC [16] ( rmsd of 1 . 16 Å for the 275 common Cα atoms; Figure 2 ) . The active site is located within the central cavity of the slightly larger C-terminal domain , with the catalytically essential mononuclear Fe2+ ligated by His145 , His215 and Glu266 . In the resting state enzyme , the coordination sphere is completed by two solvent molecules ( wat1 and wat2 ) such that the metal ion's coordination geometry is square pyramidal . The metal-ligand distances ( Table S1 ) and ligand-metal-ligand angles ( Table S2 ) are within experimental error of those observed in BphC [16] . The two most significant structural differences between HsaC and BphC appear to be associated with the larger substrate-binding pocket of the former ( 550 Å3 in HsaC versus 420 Å3 in BphC , as calculated by VOIDOO [17] ) . First , the loop-helix-loop segment comprising residues 172–190 in HsaC , which contributes to the external wall of the substrate-binding pocket , angles outwards and contains a 6-residue insertion with respect to BphC , increasing the opening of the substrate-binding pocket by up to 10 Å . Second , the distal portion of the substrate-binding pocket , which accommodates the non-catecholic portion of the substrate , is lined with fewer bulky residues in HsaC . For example Met175 , Phe202 , His209 in BphC ( PDB 1HAN ) are Leu174 , Met207 and Val214 in HsaC . Both the insertion and the smaller residues occur in other steroid-degrading extradiol dioxygenases [4] . In crystals of HsaC soaked anaerobically with DHSA , the active site cavity of each molecule in the asymmetric unit contained additional electron density that corresponds to the steroid metabolite , DHSA ( Figure 3 ) . The structure of the protein in the two molecules is essentially identical to that of the substrate-free enzyme ( rmsd of 0 . 39 Å for the 597 common Cα atoms ) except as noted below . However , in both molecules , the iron is hexacoordinate with a distorted octahedral geometry instead of being pentacoordinate with square pyramidal geometry as in the resting state enzyme . Remarkably , DHSA binds in different modes in each of the molecules , with the catecholic ring coordinating the Fe in bidentate and monodentate manners in molecules A and B , respectively ( Table 3 ) . The sites appear to be relatively well ordered in each molecule . This interpretation is supported by ligand-omit and Fo−Fc difference maps calculated using phases derived from the model in the absence of any ligands . These maps indicate that the electron density was compatible with the active sites of molecules A and B being fully occupied with DHSA ( Figure S1 ) , although the latter is slightly less ordered in molecule B . To further rule out alternate interpretations of the electron density in molecule B , the density was also fit using DHSA in a second orientation ( rotated 180 degrees upon the substrate's orthogonal axis ) and the product , DSHA . After refinement , these trials yielded strong positive residual density in proximity to the iron and very high temperature factors around the catechol ring , indicating that the current refined models are correct . The asymmetric , bidentate binding of the catecholic moiety in molecule A ( Figure 3 ) corresponds to that which has been reported in other extradiol dioxygenase:substrate complexes [11] , [15] . Briefly , the proximal hydroxyl ( O4 ) of the catecholic ring binds the Fe in the site trans to His145 and the distal hydroxyl ( O3 ) binds trans to His215 , displacing the two water molecules in the resting state enzyme ( Figure 3A ) . The coordination sphere is completed by a solvent species ( wat267 ) trans to Glu266 , presumed to occupy the O2-binding site . The catecholic moiety is bound asymmetrically to the iron in the sense that the O3-Fe distance is longer than that of O4-Fe , consistent with the monoanionic nature of catechol , as observed in BphC and homoprotocatechuate 2 , 3-dioxygenase ( HPCD ) . Other hydrogen bonds involving the catecholic hydroxyls reported in BphC are conserved in HsaC . In molecule B , the catecholic ring is bound to the iron in a monodentate manner via the 4-hydroxyl ( proximal ) group with an O4-Fe distance of 2 . 8 Å ( Figure 3B ) . The 3-hydroxyl group forms a long hydrogen bond with Asn249 ( 3 . 0 Å ) and a water molecule that is coordinated to the metal instead of with Asp250 as in the bidentate binding mode . With respect to its conformation in molecule A , the catecholic moiety of DHSA is rotated 60° clock-wise around the ligand's C6–C7 bond such that the O3 hydroxyl is 3 . 7 Å away from the Fe . The DHSA has greater temperature factors ( mean 51 Å2 ) in molecule B than in molecule A ( mean 35 Å2 ) , consistent with a greater degree of disorder and lesser binding affinity of the monodentate-bound catechol versus the bidentate-bound molecule . In contrast to the different binding modes of the catecholic ring in the two molecules , the bicycloalkanone moiety of the bound DHSA occupies strikingly similar conformations in the two complexes , suggesting that this moiety is a major determinant in the binding of the substrate . More precisely , the bicycloalkanone moiety occupies a largely hydrophobic portion of the substrate-binding site , contacting Leu174 , Leu190 , Leu205 , Val 214 , and Phe294 ( Figure 3C ) . These five residues are conserved in extradiol dioxygenases known or thought to preferentially cleave DHSA [4] . In both molecules , the carbonyl oxygen at C9 is orientated towards the iron ligand His215 ( O9 ) while that at C17 interacts with up to three ordered water molecules ( O17 ) . In the case of molecule A , the protein's C-terminus forms part of the substrate-binding pocket , sequestering the binding site from bulk solvent . In molecule B , the C-terminus is partially disordered . A similar partial disorder at the C-terminus was also observed in the structure of ligand-free HsaC , suggesting that crystal contacts may favor a more ordered conformation of the three C terminal residues ( residue 298–300 ) in molecule A . To assess the role of HsaC in cholesterol catabolism , we generated a precise null deletion mutant of hsaC in M . tuberculosis H37Rv by specialized transduction ( Figure 4A and 4B ) . Growth on cholesterol and other organic substrates was tested using a minimal medium . This medium supported some background growth in the absence of added substrate . However , growth of wild-type H37Rv was measurably enhanced in the presence of cholesterol ( Figure 5A ) , confirming that M . tuberculosis can utilize this steroid as a growth substrate . In contrast , the ΔhsaC mutant completely failed to grow on cholesterol while growth on glycerol was not impaired . Indeed , the ΔhsaC mutant displayed two notable phenotypes . First , the ΔhsaC mutant developed a pink color in the medium ( Figure 5B ) , indicating the accumulation of catechols and their non-enzymatic oxidation to o-benzoquinones and condensation products , as observed in the ΔhsaC mutant of R . jostii RHA1 [4] . Second , the mutant lost viability in the presence of cholesterol , displaying a ten-fold decrease in CFU over 14-day growth experiment ( Figure 5A ) . To evaluate the role of cholesterol metabolism in pathogenesis , we tested the ΔhsaC mutant in two animal models: immuno-compromised SCID mice and guinea pigs . Mice intravenously infected with 105 CFU of the ΔhsaC mutant ( median survival time 33 . 5 days±0 . 5 SD ) survived substantially longer ( p<0 . 0001 , log-rank test ) than those infected with wild-type H37Rv ( median survival time 22 . 4 days±0 . 9 SD ) or the complemented mutant ΔhsaC attBL5::pMV361::hsaC ( median survival time 26 . 9 days±1 . 4 SD ) ( Figure 5C ) . These data corroborate the predicted importance of cholesterol catabolism for virulence of M . tuberculosis and emphasize the critical role of HsaC within this pathway . They further suggest that M . tuberculosis utilizes cholesterol early during infection , prior to the onset of adaptive immunity . Guinea pigs infected via aerosol with ∼102 CFU of the ΔhsaC mutant had similar bacillary loads in the lungs at 4 weeks post-infection as compared to both H37Rv and the complemented mutant strain . However at week 8 , there were significantly fewer ( p<0 . 01 , two-way ANOVA ) ΔhsaC organisms in the lung compared to either wild-type or complemented strains ( Figure 6A ) . The spleens from the ΔhsaC-infected animals showed significantly ( p<0 . 05 ) lower bacterial loads 4 weeks post-infection , suggesting an impaired dissemination to the organ . While implantation of the ΔhsaC mutant was slightly lower than the groups challenged with the wild-type and complemented strain ( day 1 ) , the differences were not statistically significant . In accordance with the CFU data , there were more grossly visible tubercles in lungs of animals infected with the wild-type or complemented strain compared to the mutant ( Figure 6B ) . Microscopically , there were fewer lung granulomas ( Table 3 ) at both week 4 ( p<0 . 001 ) and week 8 ( p<0 . 001 ) in mutant-infected guinea pigs ( Figure 6C ) . Moreover , those in the ΔhsaC-infected guinea pig lungs were smaller and had less necrosis .
The phenotype of the ΔhsaC H37Rv mutant in cholesterol-containing medium , SCID mice and guinea pigs provides clear evidence that cholesterol metabolism contributes to the survival of M . tuberculosis in the host . The high specificity ( kcat/Km ) of HsaC for DHSA and the occupation of the enzyme's large , hydrophobic substrate-binding pocket with the bicyclo-alkanone moiety of the cholesterol metabolite are consistent with the enzyme's role in cholesterol metabolism , corroborating our previous demonstration that deletion of hsaC blocked growth on cholesterol in the related actinomycete , R . jostii RHA1 [4] . The first evidence for the role of cholesterol metabolism during pathogenesis was derived from genome-wide insertional mutagenesis studies [3] and the up-regulation of cholesterol catabolic genes during infection of macrophages [18] . Most recently , co-infection studies of mice using a mutant defective in cholesterol uptake indicated that cholesterol catabolism plays an important role in the chronic phases of infection [5] . The in vitro growth of M . tuberculosis on cholesterol and the loss of viability of the ΔhsaC mutant in the presence of cholesterol indicate that the attenuation of this mutant in the animal models is due to two factors: blockage of a catabolic pathway and the toxicity of catechols and/or quinones . The cytotoxicity of the latter compounds can arise from at least two mechanisms: ( a ) redox cycling between quinones and catechols to generate reactive oxygen species and ( b ) covalent modification of cellular components by the electrophilic o-benzoquinone [19] . This toxicity might be mitigated in the animal models by the fact that M . tuberculosis utilizes multiple growth substrates in vivo . Regardless of the precise mechanism of attenuation in the ΔhsaC mutant , the current data unambiguously establish that M . tuberculosis metabolizes cholesterol during infection . Moreover , the ΔhsaC mutant effectively provides a sensitive probe of the conditions under which cholesterol catabolism occurs , even when the latter is not essential . The most striking result of the animal studies was the reduction in granulomas in guinea pigs infected with the ΔhsaC mutant . This is consistent with the conclusion of Pandey and Sassetti [5] , and correlates with the recent finding of tubercule bacilli in close association with lipid droplets and crystalline cholesterol in a mouse model of caseating granulomas [20] . Indeed , histopathology studies have reported the progressive accumulation of cholesterol-rich lipid in alveolar macrophages leading to caseating granulomas in humans [21] , [22] . Nevertheless , the current studies further indicate that cholesterol metabolism by M . tuberculosis contributes to bacillary multiplication during earlier stages of infection and to the dissemination of the pathogen in the host . The ΔhsaC mutant likely enabled detection of this effect due to the accumulation of a toxic metabolite . However , another difference between the studies is that the mce4 permease mutant did not completely block growth on , nor metabolism of , cholesterol . Curiously , a different mce4 mutant was much less attenuated [23] . Finally , the phenomenon of comparable bacillary counts accompanied by reduced lung pathology has been described for some sigma factors mutants and a whiB3 mutant [24] . Nevertheless , it is unclear whether HsaC or an HsaC-dependent product is required for an inflammatory response in animal lungs while being dispensable for growth . HsaC appears to be significantly more susceptible to oxidative inactivation during catalytic turnover than other characterized extradiol dioxygenases . For example , the partition ratios of HsaC for each of DHSA and DHDS are 50-fold less than that of BphC for its preferred substrate , DHB [13] . Some meta-cleavage pathways , such as the xylene catabolic pathway of Pseudomonas putida mt-2 , have recruited a ferredoxin that reduces the catalytically essential iron of extradiol dioxygenases that is adventitiously oxidized during catalytic turnover , enabling the growth of the organism on a broader range of compounds [15] . BLAST searches indicate that the M . tuberculosis genome does not encode such a ferredoxin . This does not preclude the possibility that another reductase or electron-transfer protein plays this physiological role . The susceptibility of HsaC to oxidative inactivation could reflect relatively low levels of O2 in M . tuberculosis-infected lungs . Tuberculous granulomas in lungs of guinea pigs , rabbits and non-human primates were found to be positive for the hypoxia marker pimonidazole hydrochloride ( PIMO ) [25] and the oxygen tension in small pulmonary lesions in infected rabbits were about 3% that of uninfected lungs and below the KmO2 of HsaC . Interestingly , hypoxic conditions have been shown to upregulate a number of genes in M . tuberculosis , including fadD19 , an acyl CoA synthetase in the cholesterol metabolism pathway [26] . Although transposon mutagenesis studies have identified many cholesterol metabolic genes as essential for survival in activated macrophages [3] , there is almost no correlation between the up-regulation of genes in response to low O2 and macrophage activation [18] . An intriguing possibility is that M . tuberculosis sequesters O2 for HsaC and other oxygenases of the pathway to improve the degradation of steroids in certain cellular environments . Indeed , trHbO , one of two truncated hemoglobins harbored by the pathogen and encoded by glbO , has been proposed to increase the availability of O2 for respiration [27] . Moreover , the heterologous expression of related hemoglobins increased the rate of the microbial degradation of aromatic compounds by dioxygenase-dependent pathways [28] . In a recent study , glbO was found to be most strongly up-regulated by hypoxia , and was also up-regulated late during infection of macrophages [29] . Nevertheless , the KmO2 of HsaC is almost two orders of magnitude greater than that of some extradiol dioxygenases isolated from hypoxic soil environments [30] , suggesting that this enzyme , and by extension the cholesterol catabolic pathway of M . tuberculosis , has not evolved to function in extremely hypoxic environments . The structure of the monodentate-bound HsaC:DHSA complex was unexpected , but potentially provides insights into the initial substrate-binding steps of extradiol dioxygenases . Observation of this species is reminiscent of the trapping of three catalytic intermediates in different protein molecules of a single crystal of HPCD [11] , another extradiol dioxygenase . In that case , stabilization of different intermediates at different active sites was ascribed in part to crystal packing forces . In the current studies , the different packing forces affecting molecules A and B , reflected in the greater disorder of the C-terminal residues of molecule B , may contribute to stabilization of the different ligand binding mode . Irrespective of how the monodentate-bound catechol was stabilized , this species was proposed to occur by Groce et al . , who described the binding reaction of 4-nitrocatechol ( 4-NC ) to HPCD as proceeding via three observable steps in addition to an unobserved initial rapid association step [12] . 4NC is a poor substrate for HPCD , binding to the active site as a dianion instead of as a monoanion observed for physiological substrates . Nevertheless , the initial binding of the catechol to the iron was proposed to be monodentate via the hydroxyl group attached to the ring carbon that is eventually subject to nucleophilic attack by the activated oxygen intermediate , consistent with the current structural data . Finally , the observation of the bicycloalkanone moiety of the DHSA in essentially identical positions ( rmsd = 0 . 25 Å ) in the bidentate and monodentate complexes suggests that this moiety is a determinant in the initial complex that is proposed to form reversibly between extradiol dioxygenases and their substrates . Although 2′ , 6′-diCl DHB and 4-Cl DHDS efficiently inactivated HsaC during catalytic turnover , their respective modes of action likely differ , reflecting steric and electronic considerations , respectively . 2′ , 6′-DiCl DHB strongly inhibits BphC ( Kic = 7±1 nM ) due to partial occlusion of the likely O2-binding site by one of the chloro substituents [7] . The O2-binding site , defined by Val148 , Phe187 and Ala198 is conserved in HsaC ( Val147 , Phe192 and Ala203 ) . 2′ , 6′-diCl DHB does not inhibit HsaC as effectively as BphC , likely due to the poorer fit of the non-hydroxylated phenyl ring into the active site ( Figure S2 ) . By contrast , 4-Cl DHDS likely inactivates HsaC due to the electron-withdrawing group on the catecholic ring . This basis of inactivation has been reported in a range of extradiol enzymes , including BphC [14] and human 3-hydroxy-anthranilate-3 , 4-dioxygenase [31] , an enzyme essential to the biosynthesis of quinolinate from tryptophan . While the precise role of cholesterol metabolism by Mtb in human patients remains to be determined , particularly considering the limitations of the various animal models , the presented structural and kinetic data should facilitate the design of more potent inhibitors of HsaC .
DHB and 2′ , 6′-diCl DHB were synthesized according to established procedures [32] . DHDS ( 6; Figure 7 ) was prepared starting from commercially available 2-methoxy-5-methylphenol ( 1 ) which was converted into an intermediate MOM derivative allowing a directed ortho metalation ( DoM ) and iodination reaction sequence to form the corresponding aryl iodide 2 . Compound 2 was subjected to Heck coupling conditions , as precedented [33] , to afford the stilbene 3 which was reduced ( 4 ) and deprotected ( 5 ) to give the requisite catechol 6 . The latter was purified by silica gel chromatography and its identity was confirmed by 1H and 13C NMR . 1H NMR of DHDS ( 400 MHz , CDCl3 ) δ/ppm: 7 . 37 ( d , J = 7 . 1 Hz , 1H ) , 7 . 24–7 . 13 ( m , 3H ) , 6 . 65 ( d , J = 8 . 1 Hz , 1H ) , 6 . 60 ( d , J = 8 . 1 Hz , 1H ) , 5 . 20 ( s , 1H ) , 4 . 86 ( br s , 1H ) , 2 . 93 ( s , 4H ) , 2 . 23 ( s , 1H ) . 13C NMR of DHDS ( 100 MHz , CDCl3 ) δ/ppm: 142 . 2 , 141 . 0 , 139 . 4 , 133 . 8 , 130 . 6 , 129 . 5 , 129 . 4 , 127 . 5 , 126 . 9 , 126 . 4 , 121 . 4 , 112 . 6 , 32 . 9 , 27 . 1 , 18 . 8 . 1H NMR of 4-Cl DHDS ( 400 MHz , CDCl3 ) δ/ppm: 7 . 38–7 . 36 ( m , 1 H ) , 7 . 20–7 . 16 ( m , 3 H ) , 6 . 70 ( s , 1 H ) , 5 . 52 ( br s , 1 H ) , 5 . 35 ( br s , 1 H ) , 2 . 96–2 . 94 ( m , 4 H ) , 2 . 18 ( s , 3 H ) . Full experimental details of the synthesis and characterization data of DHDS and related compounds will be presented in a future publication ( J-X . Wang , L . D . Eltis , and V . Snieckus , unpublished results ) . DHSA was generated by incubating a culture of the ΔhsaC mutant of Rhodococcus jostii RHA1 [4] with cholesterol . Briefly , several colonies were used to inoculate 100 ml W minimal salt medium [34] containing 20 mM pyruvate . At mid-log phase ( OD600 of 1 . 0 ) , 50 ml of preculture was used to inoculate 5 litres W media containing 20 mM pyruvate and 0 . 5 mM cholesterol . Cells are harvested at OD600 of 1 . 5 and pellet was resuspended in 0 . 5 litres W media containing 0 . 5 mM cholesterol in a 2-litre baffled flask . Production of metabolites in culture supernatant was monitored using HPLC . At highest DHSA production , the culture supernatant was collected by centrifugation , acidified using 0 . 5% orthophosphoric acid to ∼pH 6 , and then extracted twice with 0 . 5 volumes of ethyl acetate . The ethyl acetate fractions were pooled , dried with anhydrous magnesium sulfate , and evaporated to dryness with a rotary evaporator . The residue was dissolved in a 44:56 mixture of methanol/water containing 0 . 5% phosphoric acid and purified using a Waters model 2695 HPLC ( Milford , MA ) equipped with a Prodigy 10-µm ODS-Prep column ( 4 . 6×250 mm; Phenomenex , Torrance , CA ) . Metabolites were eluted using the same methanol/water solvent at a flow rate of 1 ml/min . The eluate was monitored at 280 nm . Fractions containing DHSA ( tR∼35 min ) were pooled , added to 10 volumes of water , and extracted as described above . All other chemicals were of analytical grade or higher . HsaC from M . tuberculosis H37Rv was produced in Escherichia coli , as previously described [4] . M . tuberculosis H37Rv strains were grown in a minimal medium ( KH2PO4 1 g/l , Na2HPO4 2 . 5 g/l , asparagine 0 . 5 g/l , ferric ammonium citrate 50 mg/l , MgSO4×7 H2O 0 . 5 g/l , CaCl2 0 . 5 mg/l , ZnSO4 0 . 1 mg/l , Tyloxapol 0 . 05% , v/v ) containing 0 . 1% ( v/v ) glycerol or 0 . 02% ( w/v ) cholesterol . Cholesterol was added from a 25 mg/ml stock solution dissolved in isopropanol . Minimal medium containing 0 . 8% ( v/v ) isopropanol was used as a control . Growth was monitored by measuring colony forming units ( CFU ) by plating serial dilutions of cultures onto Middlebrook 7H10 agar supplemented with 10% ( v/v ) OADC enrichment ( Becton Dickinson Microbiology Systems , Spark , MD ) and 0 . 5% ( v/v ) glycerol . For generating an allelic exchange construct designed to replace the hsaC gene ( Rv3568c ) with a γδres-sacB-hyg-γδres cassette comprising the sacB and hygromycin resistance genes flanked by res-sites of the γδ-resolvase , upstream and downstream flanking DNA regions of hsaC were amplified by PCR employing the oligonucleotide pair Rv3568c-LL ( 5′-TTTTTTTTCCATAAATTGGTCCGCTGGTGGGCAAC TCGTT-3′ ) and Rv3568c-LR ( 5′-TTTTTTTTCCATTTCTTGGCCTTCGGCATTCGCGCATC-3′ ) introducing Van91I restriction sites ( underlined ) for amplification of the upstream flanking region Rv3568c-L , and the oligonucleotide pair Rv3568c-RL ( 5′-TTTTTTTTGCATAGATTGC AGCCGAGTGGTCAGCCCGTAT-3′ ) and Rv3568c-RR ( 5′-TTTTTTTTGCATCTTTTGCTAA CGGCGGTTCCAACGACA-3′ ) introducing BstAPI restriction sites ( underlined ) for amplification of the downstream flanking region Rv3568c-R . Subsequently , Rv3568c-L and Rv3568c-R were digested with Van91I or BstAPI , respectively , and ligated with Van91I-digested p0004S vector arms ( T . Hsu and W . R . Jacobs Jr . , unpublished results ) , resulting in the knock-out construct pRv3568cS which was then linearized with PacI and cloned and packaged into the temperature-sensitive phage phAE159 ( J . Kriakov and W . R . Jacobs Jr . , unpublished results ) as described [35] , yielding the knock-out phage phRv3568cS . Allelic exchange in M . tuberculosis H37Rv using the phage phRv3568cS was achieved by specialized transduction as reported previously [35] , resulting in deletion of nucleotides 220–375 of the hsaC gene ( 903 bp ) and replacement by the γδres-sacB-hyg-γδres cassette ( Figure 4A ) . The obtained mutants were verified by Southern analysis of XhoI-digested genomic DNA isolated from independent mutant clones as well as the wild-type using radiolabeled Rv3568c-R as probe ( Figure 4B ) . For complementation of the ΔhsaC mutant , the hsaC gene was amplified by PCR using the oligonucleotides 5′-TTTTTTCAGCTGCAATGAGCATCCGGTCGCTGGGC-3′ ( 5′ primer ) and 5′-TTTTTTAAGCTTCTAGCCGCGAGCGCCTACGGTG-3′ ( 3′ primer ) and cloned via the primer-introduced restriction sites ( underlined ) as a PvuII-HindIII fragment downstream of the constitutive hsp60 promoter into plasmid pMV361KanR , which allows single copy integration into the genome of M . tuberculosis [36] and complementation in trans , resulting in complemented mutant strain ΔhsaC attBL5::pMV361::hsaC . SCID/NCr ( BALB/c background ) mice ( 4- to 6-week-old females ) were infected intravenously through the lateral tail vein with 105 CFU of various M . tuberculosis H37Rv strains suspended in 200 µl PBS containing 0 . 05% Tween 80 . Ten mice per group were infected and survival of mice was monitored . Specific pathogen-free outbred Hartley strain guinea pigs ( 250–300 g; Charles River Breeding Laboratories , Inc . ( Wilmington , MA ) ) were infected via the respiratory route in an aerosol chamber ( University of Wisconsin Engineering Shops ( Madison , WI ) ) with a nebulizer concentration of 2×107 CFU/ml of the three strains of M . tuberculosis H37Rv [37] ( n = 15 guinea pigs per group ) . Animals were euthanized on day 1 , 4 weeks and 8 weeks post-infection . The right lower lung lobe and half of the spleen was homogenized in sterile saline and appropriate 10-fold dilutions were inoculated on M7H10 agar plates [38] . The lower left lung lobe and half of the spleen were taken for histopathology . Following 3 weeks of incubation at 37°C , the colonies were counted and the data were transformed in log10 values for statistical analysis . Mouse and guinea pig infection protocols were approved by the Animal Care and Use Committee at Albert Einstein College of Medicine and at Texas A&M University , respectively . The number of low power ( 20× ) fields was counted for each specimen . Within each field , the number of granulomas was also tabulated permitting the calculation of the number of granulomas per low-power microscopic field . Because the size and extent of necrosis of each granuloma varies , a subjective determination on a scale of 1–4 of disease severity was also assessed so that both quantitative and qualitative measures could be used to describe the extent of tissue damage in a manner similar to a recently described method [39] . HsaC was purified anaerobically using a two-column protocol derived from that used to purify BphC [13] . Briefly , cells from 3 litres of culture were resuspended in 30 ml of 10 mM TRIS , pH 7 . 5 containing 1 mM MgCl2 , 1 mM CaCl2 and 0 . 1 mg/ml Dnase I and disrupted using a French Press operated at 20 , 000 psi . The cell debris was removed by ultracentrifugation ( 120 , 000g×45 min ) . The clear supernatant fluid ( ∼40 ml ) was decanted , referred to as the raw extract , and divided into two equal portions . Each portion was loaded onto a column packed with Source15 Phenyl resin ( 2×9 cm ) and equilibrated with 10 mM TRIS , pH 7 . 5 containing 1 M ammonium sulphate . The column was operated at a flow rate of 5 ml/min . The enzyme activity was eluted with a linear gradient of 1 to 0 M ammonium sulphate over 8 column volumes . Fractions ( 10 ml ) containing activity from the two runs were concentrated to 10 ml with a stirred cell concentrator equipped with a YM10 membrane ( Amicon , Oakville , Ontario ) and loaded onto a Mono Q anion exchange column ( 1×8 cm ) equilibrated with 10 mM TRIS , pH 7 . 5 containing 5% t-butanol , 2 mM dithiothreitol ( DTT ) and 0 . 25 mM ferrous ammonium sulphate . The column was operated at a flow rate of 2 ml/min . The enzyme activity was eluted with a linear gradient of 0 . 2 to 0 . 4 M NaCl over 20 column volumes . Fractions exhibiting activity were combined , exchanged into the column equilibration buffer , concentrated to 20–25 mg/ml protein , and flash frozen as beads in liquid N2 . Purified HsaC was stored at −80°C for several months without significant loss of activity . Aliquots of HsaC were thawed immediately before use and exchanged into 20 mM HEPES , 80 mM NaCl ( I = 0 . 1 ) , pH 7 . 0 containing 5% t-butanol using a desalting column . Samples of HsaC were further diluted for enzyme kinetics using the same buffer containing 0 . 1 mg/ml BSA and 2 mM DTT , except in the inactivation experiments . For the latter , enzyme was diluted in the same buffer without DTT and were used within two hours . HsaC activity was verified at the beginning and end of each set of experiments . Protein and iron concentrations were evaluated colorimetrically using the Bradford method [40] and Ferene S [41] , respectively . Enzyme activity was routinely measured by following the consumption of O2 using a Clark-type polarographic electrode as described previously [13] unless otherwise stated . Experiments were performed in a total volume of 1 . 3 ml 20 mM HEPES , 80 mM NaCl , pH 7 . 0 , 25 . 0±0 . 1°C equilibrated with 5% O2 in N2 ( 103±3 µM dissolved O2 ) . Reaction buffers containing different concentrations of dissolved O2 were prepared by bubbling them with mixtures of O2 and N2 gases and transferring them to the reaction chamber as described previously [13] . The amount of active HsaC was defined by the iron content of the sample . Steady-state kinetic parameters were calculated using LEONORA [42] . Cleavage of 2′6′-diCl DHB and was measured by following the rate of appearance of the ring-cleaved product using a Cary 5000 spectrophotometer equipped with a thermojacketed cuvette holder ( Varian , Walnut Creek , CA ) . Initial velocities were determined from a least-squares analysis of the linear portion of the progress curves . Partition ratios expressing the number of substrate molecules consumed per enzyme molecule inactivated were determined spectrophotometrically for DHB , 2′ , 6′-diCl DHB , DHDS , and DHSA by following the appearance of the ring-cleaved products at 434 nm ( ε = 23 . 4 mM−1 cm−1 ) , 391 nm ( ε = 36 . 5 mM−1 cm−1 ) , 396 nm ( ε = 6 . 3 mM−1 cm−1 ) , and 392 nm ( ε = 7 . 6 mM−1 cm−1 ) , respectively . The partition ratio for 4-Cl DHDS was determined by oxygraph electrode due to a very low extinction coefficient . Partition ratios were determined under saturating substrate conditions ( [S]≫Km ) . Crystals of substrate-free HsaC were grown anaerobically at room temperature using the hanging drop method ( protein 15 mg/ml , crystallization solution: 12–15% PEG 3350 , 0 . 2 M ammonium tartrate , 25% ethylene glycol ) . Single crystals appeared in 2–5 days and grew to their full size ( 200 µm×200 µm ) in two weeks . The crystals were frozen anaerobically in liquid N2 prior to diffraction experiments . The complex with DHSA was formed by adding 0 . 2–0 . 5 µl of crude extract containing DHSA dissolved in t-butanol directly to the drop containing the crystal and incubating for up to 2 hr ( anaerobically ) at room temperature . The HsaC:DHSA crystals were flash frozen in liquid N2 . X-ray data collections were performed under cryogenic conditions using an in-house rotating anode X-ray generator ( CuKα radiation , λ = 1 . 542 Å ) and at the Advanced Light Source ( ALS , Beamline 8 . 2 . 2 ) . Data were processed using HKL2000 [43] . Molecular replacement was performed using PHASER [44] and the structure of substrate-free BphC ( PDB accession code 1HAN ) with Fe and waters deleted as a search model . The highest scoring solution placed a dimer in the asymmetric unit , which was used as a starting model for re-building and structure refinement , which was performed using CNS [45] ( simulated annealing ) and REFMAC [46] in alternation with manual rebuilding using COOT [47] . For the HsaC:DHSA complexes , difference Fourier electron density maps revealed additional density within the active site consistent with a bound DHSA . The diffraction data and properties of the refined model are characterized in Table 2 . A model for the substrate was established using the PRODRG server . Electron density maps were calculated with the CCP4 suite ( FFT function ) . Structural figures and graphical rendering were made by using PYMOL [48] . The final model of HsaC:DHSA contains a dimer of HsaC covering 299/298 residues of each chain , two DHSA molecules , and 485 water molecules . The final model of substrate-free HsaC contains a dimer of HsaC covering 295 residues ( 2–296 ) of each chain , one tartrate and 712 water molecules . The coordinates for HsaC:DHSA and HsaC alone were deposited in the Protein Data Bank ( www . pdb . org ) with accession codes 2ZI8 and 2ZYQ , respectively .
|
Mycobacterium tuberculosis , the etiological agent of TB , is the most devastating infectious agent of mortality worldwide: it is carried by one-third of all humans and kills nearly two million people annually . Recent work has established that the pathogen metabolizes cholesterol , although the role of this metabolism in pathogenesis remains unclear . In the current study , we demonstrate that HsaC is a key enzyme in the cholesterol catabolic pathway and that it can be inactivated by compounds that resemble its substrate . Using molecular genetic approaches , we demonstrated that the enzyme is essential for the growth of M . tuberculosis on cholesterol and that a lack of this enzyme impairs the survival of the pathogen in each of two animal models . These studies provide definitive evidence that M . tuberculosis metabolizes cholesterol during infection and that this metabolism occurs during the early stages of infection . The oxygen-utilizing enzymes of the cholesterol catabolic pathway , of which HsaC is but one example , are intriguing potential chemotherapeutic targets , as their inhibition can lead to toxic metabolites , including reactive oxygen species . Overall , our study combines a variety of approaches to provide novel insights into a disease of global importance and into the mechanism of an interesting class of enzymes .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biochemistry/biocatalysis",
"biochemistry/biomacromolecule-ligand",
"interactions",
"infectious",
"diseases/bacterial",
"infections",
"microbiology/microbial",
"physiology",
"and",
"metabolism",
"microbiology/cellular",
"microbiology",
"and",
"pathogenesis"
] |
2009
|
Studies of a Ring-Cleaving Dioxygenase Illuminate the Role of Cholesterol Metabolism in the Pathogenesis of Mycobacterium tuberculosis
|
Cutaneous leishmaniasis ( CL ) is a parasitic disease causing chronic , ulcerating skin lesions . Most humans infected with the causative Leishmania protozoa are asymptomatic . Leishmania spp . are usually introduced by sand flies into the dermis of mammalian hosts in the presence of bacteria from either the host skin , sand fly gut or both . We hypothesized that bacteria at the dermal inoculation site of Leishmania major will influence the severity of infection that ensues . A C57BL/6 mouse ear model of single or coinfection with Leishmania major , Staphylococcus aureus , or both showed that single pathogen infections caused localized lesions that peaked after 2–3 days for S . aureus and 3 weeks for L . major infection , but that coinfection produced lesions that were two-fold larger than single infection throughout 4 weeks after coinfection . Coinfection increased S . aureus burdens over 7 days , whereas L . major burdens ( 3 , 7 , 28 days ) were the same in singly and coinfected ears . Inflammatory lesions throughout the first 4 weeks of coinfection had more neutrophils than did singly infected lesions , and the recruited neutrophils from early ( day 1 ) lesions had similar phagocytic and NADPH oxidase capacities . However , most neutrophils were apoptotic , and transcription of immunomodulatory genes that promote efferocytosis was not upregulated , suggesting that the increased numbers of neutrophils may , in part , reflect defective clearance and resolution of the inflammatory response . In addition , the presence of more IL-17A-producing γδ and non-γδ T cells in early lesions ( 1–7 days ) , and L . major antigen-responsive Th17 cells after 28 days of coinfection , with a corresponding increase in IL-1β , may recruit more naïve neutrophils into the inflammatory site . Neutralization studies suggest that IL-17A contributed to an enhanced inflammatory response , whereas IL-1β has an important role in controlling bacterial replication . Taken together , these data suggest that coinfection of L . major infection with S . aureus exacerbates disease , both by promoting more inflammation and neutrophil recruitment and by increasing neutrophil apoptosis and delaying resolution of the inflammatory response . These data illustrate the profound impact that coinfecting microorganisms can exert on inflammatory lesion pathology and host adaptive immune responses .
Leishmaniasis constitutes a spectrum of diseases with distinct clinical forms usually caused by different species of Leishmania protozoa [1] . Each of the Leishmania species can also lead to highly variable clinical outcomes in different individuals , ranging from asymptomatic to severe infection . The most common disease form is cutaneous leishmaniasis ( CL ) , which presents as lesions that often ulcerate and usually spontaneously resolve within weeks to months [2] . L . major is a species responsible for a substantial portion of the CL burden in the Eastern Hemisphere [3] . An underappreciated variable influencing CL disease outcome is the local microbial flora at the site of mammalian infection [4–8] . However , specific interactions between Leishmania parasites , local bacteria , and the host immune response are underexplored . Leishmania spp . are usually introduced into a susceptible mammalian host through the bite of an infected female phlebotomine sand fly . The insect bites by repeated probing activity , a behavior that amplifies in flies harboring Leishmania spp . parasites , forming a pool of blood in the dermis into which the parasite is inoculated [9–11] . In addition to commensal bacteria on the host’s skin , recent data show the infection site is also exposed to microbes from the sand fly gut [12] . Ulcerated lesions provide a portal for bacterial invasion , occasionally leading to superinfection [13–15] . Staphylococcus and Streptococcus species are the two most common bacterial genera that have been detected in surveys of CL lesion microbiota [13–15] . Sand fly midgut microbiota include bacterial species belonging to the families Staphylococcaceae and Streptococcaceae within the phylum Firmicutes [16 , 17] . Bacteria egested by sand flies into the skin during a blood meal can also incite an inflammatory response [12] . Staphylococcus aureus infection is known to activate the NLRP3 inflammasome , a multi-protein complex that activates caspase-1 which in turn cleaves and prompts release of IL-1β and IL-18 [18–20] . Inflammasomes are activated during human CL caused by L . braziliensis , and local IL-1β exacerbates lesion pathology in murine models of CL [21–25] . IL-1β also plays a role during infections with visceralizing species of Leishmania [12 , 26–28] . Based on the above observations , we hypothesized that bacteria coinfecting the site of L . major infection will activate proinflammatory mediators and thereby modify the host response to L . major infection . Our data showed that S . aureus coinfection indeed had a profound influence on the outcome of L . major lesions , leading to lesion exacerbation within the first four weeks of coinfection .
All experiments with vertebrate animals were performed in accordance with recommendations in the Guide for the Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committees ( IACUC ) of the University of Iowa ( protocol 7071099 ) and the Iowa City Veterans’ Affairs Medical Center ( ACORP protocol 1690501 ) . All procedures , including anesthesia and experimental endpoints , were performed in accordance with American Veterinary Medicine Association ( AVMA ) guidelines , and were approved by review committees at the University of Iowa and the Iowa City VA Medical Center . Four- to six-week-old C57BL/6N female mice purchased from Charles River were used in the experiments in this study . Mice were housed under specific pathogen-free conditions at the Iowa City Veteran’s Affairs Medical Center Animal Research Facility . Procedures for L . major and S . aureus preparation and co-inoculation by intradermal injection into mice ears are described in the protocol available online: dx . doi . org/10 . 17504/protocols . io . vdse26e . L . major IA-2 strain was recently isolated from a patient who acquired CL in Iran . The promastigote forms of either wild-type L . major IA-2 or IA-2 expressing genes encoding luciferase and mCherry were grown at 26°C in Schneider’s Drosophila medium + L-glutamine ( Gibco by Life Technologies ) supplemented with 10% heat-inactivated fetal bovine serum ( FBS ) ( SAFC Industries ) , 2 mM L-glutamine ( Gibco by Life Technologies ) , 50 μg/mL gentamicin sulfate ( IBI Scientific ) , and 1 . 2 μg/mL biopterin ( Cayman Chemical Company ) . Metacyclic promastigotes were isolated by Ficoll-Paque PLUS ( GE Healthcare ) density gradient as previously described [29] and suspended to a concentration of 106 parasites in 10 μL of phosphate buffered saline ( PBS; Gibco by Life Technologies ) . For experimental conditions , parasite suspensions were mixed with liquid media containing buffer along or S . aureus immediately before inoculation into mice . Newman is a methicillin-sensitive S . aureus [MSSA; α-toxin ( hla ) positive] strain [30] . S . aureus LAC::lux and S . aureus LAC pCM29 ( chloramphenicol resistant ) are methicillin-resistant USA300 strains that express bacterial luciferase or green fluorescent protein ( GFP ) , respectively [MRSA , hla+] [31–33] . S . aureus MNPE [hla+ , toxic shock syndrome superantigen ( TSST-1 ) positive] is a USA200 strain , which was kindly provided by Dr . Patrick Schlievert of the University of Iowa [34] . Bacteria were grown at 37°C on a semisolid tryptic soy agar ( TSA ) plate overnight . Single colonies were cultured in tryptic soy broth ( TSB ) at 37°C with shaking overnight . Overnight cultures were diluted 1:100 in TSB and grown to an optical density ( OD600 ) of 0 . 5 . Ten mL of bacterial cultures were washed by centrifugation in PBS . Based on the estimate that an OD600 of 1 . 0 corresponds to 3x105 colony-forming units ( CFU ) per μL , 104 CFUs were suspended in 10 μL PBS either with or without L . major . Bacterial CFUs in the final injection doses were confirmed by serial dilutions on TSA plates incubated overnight at 37°C . Mice were anesthetized by intraperitoneal ( i . p . ) injection of ketamine ( 80 mg/kg , Ketalar , Par Pharmaceutical Cos . , Inc . ) and xylazine ( 10 mg/kg , AnaSed , LLOYD Laboratories , LLOYD Inc . ) , and then intradermally injected in one ear pinna with 106 L . major parasites , 104 CFUs of S . aureus , or a mixture of both in 10 μL volume . Lesion dimensions were measured daily for one week followed by weekly measurements for three additional weeks . Lesion measurements were made using a Mitutoyo Flat Anvil Dial Thickness Gage ( 0–22 mm ) in 0 . 01 mm increments for thickness , and a ruler for length and width . Lesion volume was calculated using the formula for volume of an ellipsoid: lesionvolume=43×π×length2×width2×thickness2 . For the five-day S . aureus preliminary dosage experiment ( S1 Fig ) , lesion area was calculated using the formula for the area of an ellipse: lesionarea=π×length2×width2 . At experiment endpoints , mice were euthanized in accordance with AVMA guidelines as approved by the University of Iowa IACUC . After intradermal inoculation of 104 CFUs of S . aureus LAC::lux , expressing the Photorhabdus luminescence lux operon . Mice were imaged under anesthesia by inhalation of 2% isoflurane ( Piramal Enterprises Limited ) . Photons , which are emitted only from live , luminescent bacteria , were quantified during a 1-minute exposure using the IVIS-200 ( in vivo imaging system ) and Living Image software from Xenogen . Total light emissions ( flux ) in a uniformly defined circular region of interest over the ear infection site were quantified at different time points of infection in each mouse . Ears were harvested at four weeks post-injection . Active lesions were bisected , and a portion was paraffin-embedded . Three μm sections were cut and stained with hematoxylin and eosin and analyzed using a semi-quantitative histological scoring method [35] . The remaining ear tissue was homogenized . Portions were either suspended in Cell Lysis Solution ( QIAGEN Sciences ) and DNA was extracted using the Puregene DNA isolation Kit ( Gentra ) protocol for mouse tail snips or suspended in 1 mL of TRIzol Reagent ( Invitrogen ) and RNA was extracted using R . Z . N . A Total RNA Kit I ( Omega Bio-Tek R6834-02 ) . Lysostaphin ( 0 . 01 U/mL , AMBI Products , LLC ) was added to DNA extraction buffer for experiments requiring S . aureus quantification by qPCR . Quantitative polymerase chain reaction ( qPCR ) was performed on DNA extracted from tissues of infected or uninfected ( control ) mice using kinetoplastid DNA ( kDNA5 ) primers as previously described [36] . Parasite genome equivalents were calculated based on a standard curve with varied amounts of L . major promastigote DNA in a constant amount of mouse ear DNA . S . aureus relative genome equivalents were determined using primers and probe specific to the thermonuclease nuc gene , and comparison to a standard curve of S . aureus DNA diluted in mouse ear DNA [37] . Quantitative PCR reactions were performed in 96-well fast plates in an Applied Biosystems QuantStudio 7 Flex Real-Time PCR System ( Life Technologies ) . Preliminary experiments documented that the qPCR method correlates with bacterial counts . To verify live S . aureus counts , ears were homogenized in 0 . 5 mL of PBS and homogenized using a Tissue Master 125 ( OMNI International ) . 10 μL of ear homogenate was spread onto TSA plates , incubated at 37°C overnight , colony-forming units ( CFUs ) were counted , and the total number of live S . aureus bacteria was calculated . Reverse transcriptase-quantitative polymerase chain reaction ( RT-qPCR ) was performed on RNA obtained from harvested ear skin . Complementary DNA ( cDNA ) was generated using the Superscript III Reverse Transcriptase First Strand Synthesis System ( Invitrogen ) using the manufacturer’s protocol with random hexamers . Samples were pre-amplified using pooled Taqman assays ( ABI ) and PreAmp Master Mix ( Fluidigm ) according to the Gene Expression Preamplification with Fluidigm PreAmp Master Mix and TaqMan Assays protocol ( Fluidigm ) . Expression of 48 control and inflammatory genes was assessed using a 48x48 qPCR dynamic array ( Fluidigm ) , with a panel of Taqman assays ( ABI ) , according to manufacturer’s instructions . Data were calculated by the ΔΔCT method , using either GAPDH or GUSB in each experiment to normalize transcripts within samples . Each normalized ΔCT value was compared to the average ΔCT for the same transcript in sham injected mice to get the -ΔΔCT , yielding the log2 ( fold change ) . The submandibular lymph nodes draining ear lesions were homogenized in 200 μl of RP10 media , which consists of 450 mL RPMI 1640 ( Gibco by Life Technologies ) , 50 mL FBS , 1 mM L-glutamine ( Gibco by Life Technologies ) , 100 units/mL penicillin/streptomycin ( Gibco by Life Technologies ) in 1 . 5 μL microfuge tubes mini-pestles . One times 105 viable lymph node cells in 100 μL were transferred to each well of a 96-well round-bottom plate ( COSTAR ) . Cells were incubated for 72 hours at 37°C , 5% CO2 with 3x105 L . major IA-2 live promastigotes as a source of live parasite antigen . After 48 hours , supernatants were collected and stored at -80°C . Cytokines were quantified using multiplex fluorescent bead arrays for murine IL-2 , IL-4 , IL-5 , IL-10 , IL-12 ( p70 ) , IL-17A , IFNγ , and TNFα from ( Biorad ) on a Luminex 200 detection instrument ( Luminex Corporation ) . Single cell suspensions of dermal cells were obtained at the indicated times post-infection ( p . i . ) from mouse ears by separating dermal sheets and incubating dermis-side down in 0 . 5 mL of 0 . 2 mg/mL of Liberase DL ( Roche ) in RP10 medium in a 24-well plate for 1 hour at 37°C , 5% CO2 , agitating every 15 minutes . One μL of Benzonase Nuclease , Purity >99% ( EMD Millipore ) was added during the last 15 minutes . Skin cells were homogenized by passage through a 70 μm nylon cell strainer , centrifuged at 3000 rpm for 10 minutes , and resuspended in RP10 . Single cell suspensions from draining submandibular lymph nodes were obtained as described above for lymph node cell stimulation and multiplex assays . Total cell number was determined by counting cells in 10 μL of each sample on a hemocytometer . Cells were washed three times in FACS buffer [PBS , 2 μM EDTA ( Fisher Scientific ) , 1% v/v FBS ( SAFC Industries ) and 0 . 1% w/v sodium azide ( Sigma-Aldrich ) ] and suspended in antibodies to surface markers diluted 1:400 in FACS buffer . Cells stained for surface markers were washed by centrifugation , fixed in PBS supplemented with 2% paraformaldehyde ( Sigma-Aldrich ) ( fixation buffer ) , and stored protected from light at 4°C overnight . Cells harvested for intracellular cytokine staining ( ICS ) were cultured for 4–5 hours in RP10 with 2 μg/mL of brefeldin A ( eBiosciences ) , 0 . 1 μg/mL of phorbol 12-myristate 13-acetate ( PMA; Sigma-Aldrich ) , and 1 μg/mL of ionomycin ( Sigma-Aldrich ) at 37°C , 5% CO2 . Cells were suspended in Permeabilization Buffer ( 0 . 1% saponin , 0 . 09% sodium azide , eBioscience ) with 1:200 dilutions of intracellular anti-cytokine antibodies at 4°C for 20–30 minutes . Cells were resuspended in fixation buffer at 4°C until analysis . For IL-17A ICS of CD11b+ cells , stimulated cells were resuspended in TruStain FcX anti-mouse CD16/32 antibody ( Biolegend ) at 1 . 0 μg/well in 50 μL for 10 minutes on ice prior to immunostaining . Cells were surface stained and permeabilized as described above , and half of each sample was stained for intracellular IL-17A with 1:200 dilutions of PE anti-IL17A clone TC11-18H10 . 1 ( Biolegend ) . The other half of each sample was stained with isotype control PE rat IgG1 kappa ( Biolegend ) . Cells were resuspended in fixation buffer at 4°C until analysis . Fluorescent anti-mouse antibodies used for surface staining: PE-Cy7 anti-CD4 clone GK1 . 5 , and FITC anti-CD8a clone 53–6 . 7 from eBioscience; Brilliant Violet 421 anti-CD45 clone 30-F11 , APC-Cy7 anti-CD11b clone M1/70 , Alexa Fluorophore 700 anti-CD90 . 2 ( Thy1 . 2 ) clone 30-H12 , FITC anti-TCR γ/δ clone GL3 , APC anti-Ly6G clone 1A8 , and FITC anti-Ly6C clone HK1 . 4 from BioLegend; PerCP-Cy5 . 5 anti-CD11b M1/70 from BD Bioscience . Intracellular cytokine stains were PE-Cy7 anti-IFNγ clone XMG1 . 2 , APC anti-pro-IL-1β clone NJTEN3 from eBioscience , and PE anti-IL17A clone TC11-18H10 . 1 from Biolegend . Flow cytometry was performed using the Becton Dickinson LSR II with 405 nm , 488 nm , 561 nm , and 639 nm lasers using BD FACSDiva ( BD Biosciences ) software . ICS gates were determined using fluorescence minus one ( FMO ) controls . Percentage of IL-17A+ CD11b+ cells was determined by subtracting the percent of PE+ cells in the IgG1 isotype controls from the percent of PE IL-17A+ cells , and the adjusted percentage was then multiplied by the total number of cells in that sample . Data were analyzed using FlowJo Software . Ears from 3 days p . i . were snap frozen in 1 . 5 mL microcentrifuge tubes in liquid nitrogen and stored at -80°C . Frozen ears were transferred to 5 mL round bottom polystyrene tubes in 300 μL of Cell/Tissue Extraction Buffer [100 mM Tris , pH 7 . 4 ( RPI Corp . ) , 150 mM NaCl ( RPI Corp . ) , 1 mM EGTA ( Fisher Scientific ) , 1 mM EDTA ( Fisher Scientific ) , 1% Triton X-100 ( Sigma ) , 0 . 5% sodium deoxycholate ( Sigma-Aldrich ) ] with 1 mM protease inhibitor cocktail ( Roche ) . Ears were homogenized using a Tissue Master 125 ( OMNI International ) and agitated on a rotating platform for 2 hours at 4°C . Tissue homogenates were microcentrifuged for 20 minutes at 13 , 000 rpm at 4°C and the supernatants were transferred into clean 1 . 5 mL microcentrifuge tubes and stored at -80°C . A murine IL-1β ELISA ( R&D Systems ) was performed on collected supernatants from homogenized ears and read using the FLUOstar Omega plate reader ( BMG Labtech ) . Single cells suspensions were obtained one day p . i . from mouse ears as described above . Cells were surface stained with Brilliant Violet 711 anti-Ly6G , Brilliant Violet 421 anti-CD45 clone 30-F11 , and APC-Cy7 anti-CD11b clone M1/70 from Biolegend . To assess the ability of neutrophils to activate the phagocyte NADPH oxidase , cells were incubated in PBS with 100 ng/mL of PMA ( Sigma Aldrich ) and 10 μM dihydrorhodamine 123 ( DHR123 , Sigma Aldrich ) for 15 minutes at 37°C , 5% CO2 , and then washed in PBS and analyzed by flow cytometry . To assay for apoptosis , cell suspensions were resuspended in 100 μL Annexin V binding buffer ( Biolegend ) and incubated with 5 μL of APC Annexin V ( Biolegend ) for 10–15 min at room temperature in the dark . Within 10 minutes prior to analysis by flow cytometry , 10 μL of Propidium Iodide Staining Solution ( Biolegend ) was added to each sample . Female C57BL/6 mice were injected intraperitoneally with 0 . 5 mg anti-mouse IL-17A clone 17F3 antibody ( Bio X Cell InVivoMab ) or isotype control ( InVivoMAb mouse IgG1 clone MOPC-21 ) . Other infected mice were injected with 0 . 5 mg anti-IL-1β antibody ( Bio X Cell InVivoMAb anti-mouse/rat IL-1β clone B122 ) or isotype control ( InVivoMAb polyclonal Armenian hamster IgG ) in PBS every 3 days starting one day prior to intradermal ear infection , for a total of 9 days . Lesion size was monitored as described above . Statistical analyses were performed using GraphPad Prism software . Data were analyzed by one-way or two-way ANOVA with Tukey’s post-test for multiple comparisons , or by student’s t-test .
Bacteria can be introduced into the skin at the time of Leishmania inoculation by the bite of a sand fly or as CL lesions develop and ulcerate . This led us to use a murine model of CL to examine the phenotypic effects of bacteria introduced at different times or body site locations relative to the L . major parasitic infection . We chose to use S . aureus , a bacterium found in the sand fly gut [16] and commonly present at the site of human CL lesions [13–15] , at a subclinical 104 bacterial colony-forming units ( CFUs ) dose , which results in detectable but minimal swelling , and no ulceration in murine skin ( S1 Fig ) . The simultaneous administration of L . major and S . aureus ( L+S in figures ) at a single site led to a significant and pronounced exacerbation of pathology compared to mice infected with either L . major or S . aureus alone ( Lm or Sa in figures ) ( Fig 1 ) . Administration of these organisms simultaneously at different body sites , sequentially at different body sites , or sequentially at the same body site in either order did not exacerbate pathologic changes compared to single infections . These results are specifically delineated in the Fig 1 legend . Infections of C57BL/6 mice with S . aureus alone led to small lesions that peaked 2–3 days after infection and without ulceration . In contrast , L . major infections led to local swelling as the lesion developed gradually over 2–3 weeks and culminated in an ulcerated lesion by 3–4 weeks after infection ( Fig 1A , 1E and 1G ) . Simultaneous inoculation of both L . major and S . aureus at the same site produced significantly exacerbated lesions , both during the first 7 days when S . aureus lesions formed and subsided , and throughout 2–4 weeks of coinfection when L . major lesions developed ( Fig 1A ) . Parasite burdens after single or coinfection did not differ at days 3 , 7 and 28 p . i . , despite the widely divergent lesion sizes ( Fig 1B , S2 Fig ) . A 10-fold lower dose ( 103 CFUs ) of S . aureus produced similarly exacerbated lesions during coinfection with L . major , and parasite burdens were also similar between these single and coinfection groups ( S3 Fig ) . However , the early load of viable S . aureus was significantly higher in the presence of L . major and S . aureus coinfection compared to S . aureus alone , illustrated with light emissions from luminescent bacteria over the first 4 days of infection ( Fig 1C ) . S . aureus loads , determined by qPCR and by CFU , were largely undetectable at day 28 when L . major lesions were large and often ulcerated ( Fig 1D ) . Other combinations of the timing or intradermal sites of microbial challenge did not lead to exacerbated lesion pathology or L . major parasite load ( Fig 1E , 1F , 1G and 1H ) . As a model of late bacterial secondary infection , mice were intradermally infected with 106 metacyclic L . major promastigotes . Ulcerating lesions formed over 21 days , after which late secondary infection was modeled by intralesional injection of 104 CFUs of the MSSA S . aureus Newman strain . Secondary S . aureus infection did not alter the L . major lesion size throughout , or the parasite burden at 28 days post-L . major infection ( Fig 1E and 1F ) . Based on the hypothesis that systemic changes in dermal immunity resulted from prior cutaneous S . aureus infection , we tested the effect of augmenting bacterial burdens at different skin sites or prior to L . major challenge . L . major and S . aureus were inoculated simultaneously but in opposite ears , or L . major was inoculated into an ear that had been infected with S . aureus 10 days previously ( Fig 1G and 1H ) . The data show that bacterial inoculation prior to or at a different site from that of L . major inoculation had no effect on the size of parasite-induced skin lesions . Consistently , L . major burdens were not significantly different among any of the infection groups at 28 days post-L . major infection ( Fig 1H ) . Since neither pathogen burden differed between infection groups at 28 days post-coinfection , we hypothesized that changes in immune cell infiltration into coinfected ears exacerbated L . major lesions . Lesion pathology at low magnification revealed that ears that had been coinfected demonstrated increased ear thickness and greater inflammatory cell infiltrates than did tissue histology in all other infection groups ( Fig 2A ) . Semi-quantitative histologic scores showed that lymphocytes and histiocytes predominated in both L . major singly infected and L . major-S . aureus coinfected ears at day 28 p . i . ( Fig 2B and 2C ) . Unexpectedly in light of the chronicity of infection ( day 28 ) , there were significantly more neutrophils at day 28 p . i . in L . major-S . aureus coinfected ears than in single L . major infections , which correlated with enhanced lesion size ( Fig 2D ) . To confirm the identity of inflammatory cells in ears of coinfected or singly infected mice , we stained cells recovered from lesions for inflammatory cell surface markers at different times of infection ( Fig 2E and 2F ) . Significantly more neutrophils ( CD45+ CD11b+ Ly6Ghi Ly6Cint ) , inflammatory monocytes ( CD45+ CD11b+ Ly6G- Ly6Chi ) , myeloid dendritic cells ( CD45+ CD11b+ CD11c+ ) , and γδ T cells ( CD45+ Thy1 . 2+ γδ TCR+ ) were recovered from coinfected ears at day 1 p . i . compared to PBS , S . aureus , or L . major singly infected , or PBS inoculated mice ( Fig 2E and 2F ) . There were also more neutrophils , myeloid dendritic cells , γδ and non-γδ T cells in coinfected ears on day 7 p . i . , compared to PBS or L . major infection groups ( Fig 2E and 2F ) . Early L . major-S . aureus coinfected lesions exhibited enhanced S . aureus replication despite the presence of more neutrophils ( Fig 1C , Fig 2E ) . Because neutrophils are critical for the clearance of cutaneous S . aureus infections and wound resolution [38–40] it seemed paradoxical that there were more neutrophils yet a greater burden of S . aureus during the first three days of coinfection with L . major ( Fig 1C , Figs 2E and 3A ) . To understand the mechanisms underlying this apparent paradox , we first explored the hypothesis that neutrophil phagocytic or microbicidal capacities were defective during coinfection . To assess the phagocytic capacity of recruited neutrophils , we used S . aureus LAC expressing green fluorescent protein ( GFP ) and L . major IA-2 expressing luciferase and mCherry to detect selectively each microbe after intradermal injection into mouse ears . Neutrophils from coinfected lesions one day p . i . phagocytosed either pathogen ( Fig 3B ) . The phagocytosis of S . aureus or L . major was the same in the absence or presence of coinfection . Thus , coinfection did not compromise the capacity of neutrophils to ingest the pathogens . Because neutrophil killing of S . aureus predominantly depends on a functional phagocyte NADPH oxidase , we used flow cytometric analysis of dihydrorhodamine ( DHR ) oxidation to assess the ability of recruited neutrophils to generate reactive oxygen species in response to the soluble agonist , phorbol myristate acetate ( PMA ) [41] . Based on the average DHR GMFI of PMA-stimulated neutrophils from the infection site , neutrophils containing either microbe had functional NADPH oxidases , with activities highest in neutrophils recovered from lesions containing S . aureus ( Fig 3C ) . Taken together , these data suggest that the presence of more neutrophils in the lesions from coinfected tissue did not reflect defective antimicrobial capacity of the recruited neutrophils . We next tested the hypothesis that the excess of neutrophils in coinfected lesions might reflect a defect in the resolution of the inflammatory response . That is , a failure of the recruited neutrophils to undergo apoptosis and be efferocytosed by tissue macrophages . To test this hypothesis , we quantified apoptotic neutrophils by annexin V ( AnnV ) staining and measured the expression of efferocytosis-related genes by RT-qPCR . AnnV and propidium iodide ( PI ) staining of ear cells from all groups of mice coinfected one day p . i . revealed a greater percentage of neutrophils undergoing apoptosis ( AnnV+PI- ) or necrotic cells ( AnnV+PI+ ) , compared to cells from mice injected with PBS , L . major , or S . aureus alone ( Fig 3D and 3E , S4 Fig ) . Given that coinfected ears had over 9-fold greater absolute numbers of neutrophils than ears infected with either pathogen alone ( Fig 2E ) , this AnnV staining indicates the majority of those neutrophils are apoptotic ( Fig 3D and 3E ) . Thus , significantly more neutrophils recovered from coinfected lesions were apoptotic compared to all other infection groups . Resolution of inflammation depends in part on efferocytosis , a process that limits inflammatory and promotes anti-inflammatory responses through the uptake of apoptotic cells by tissue macrophages [42 , 43] . Many ligand-receptor pairs initiate uptake of apoptotic cells , but a common ligand is phosphatidylserine exposed on the apoptotic or “to-be-eaten” cell . The potential biological importance of greater numbers of cutaneous AnnV+ neutrophils led us to examine the local expression of genes often associated with efferocytosis , reasoning that the transcripts will parallel the immune “tone” or abundant immune functions active locally . Total RNA content in ears of S . aureus and/or L . major singly- or coinfected ears or uninfected control ears were extracted 3 days p . i . RT-qPCR assays revealed lower total mRNA present in S . aureus singly and coinfected ears compared to L . major-infected ears for genes such as immunoregulatory cytokines IL-10 and IL-13 , and matrix metalloproteinase-9 ( MMP9 ) , an enzyme important for cutaneous wound healing ( S5A & S5B Fig ) . Additionally , S . aureus-infected ears had lower mRNA levels for LXRα and PPARδ , two nuclear receptors associated with efferocytosis ( S5C Fig ) . Taken together , these data suggest that many of the neutrophils recovered from skin coinfected with S . aureus were apoptotic . In a setting where factors important for efferocytosis and resolution of inflammation were downregulated , the recruited neutrophils would accumulate but be ineffective at clearing infection , all of which may promote S . aureus survival and replication . In fact , in vitro studies demonstrate that human neutrophils harboring viable S . aureus display increased AnnV binding but are not efferocytosed by human monocyte-derived macrophages [44] . The presence of more immune cells in the early stages , and the persistent neutrophil infiltrate in lesions at the late stages of coinfection led us to examine potential differences in cytokines expressed . After 4 weeks of infection , most transcripts encoding inflammatory or modulatory cytokines were similar in abundance in ear tissues extracted from L . major-infected and coinfected groups . These included transcripts encoding innate , Type 1 , and Type 2 cytokines or chemokines ( S6 Fig and S1 Table ) . To examine antigen-responsive adaptive immune responses , we restimulated draining lymph node ( LN ) cells from infected mice with total Leishmania antigen and measured cytokines released in a fluorescent bead-based multiplex cytokine assay . Although L . major infection led to increased IFNγ and IL-4 compared to uninfected , in PBS-injected mice , the only Type 1 ( Th1-type ) or Type 2 ( Th2-type ) cytokines elevated above single infections were IL-2 and IL-5 ( Fig 4 ) . These results fail to implicate Type 1 and Type 2 T cells in the exacerbated pathology of coinfection . There were also no significant differences in IL-10 , IL-12 ( p70 ) , or TNFα released by cells from any of the infection groups . However , IL-17A was found at significantly higher concentrations in supernatants of cells from antigen-stimulated coinfected draining lymph nodes compared to L . major-infected alone draining lymph nodes ( Fig 4 ) . This result raises the possibility that Leishmania-specific Type 17 helper T cells ( Th17 ) develop during coinfection but not during single infections in the C57BL/6 model . In order to identify IL-17A producing cells during L . major-S . aureus coinfection , we performed intracellular cytokine and surface marker staining on LN or ear-derived inflammatory cells extracted one day p . i . Inflammatory cells were incubated in PMA and ionomycin in the presence of Brefeldin A , and then stained and analyzed by flow cytometry using the gating strategy shown in S7 , S8 , and S9 Figs . Intracellular and surface staining revealed greater numbers of IL-17A-producing γδ T cells and non-γδ T cells observed in coinfected ears at both 1 and 7 days p . i . ( Fig 5A ) . IL-17A+ CD11b+ cells were also detected in coinfected ears at 1 day p . i . ( Fig 5B and 5C ) . We also observed more IFNγ-producing non-γδ T cells in coinfected compared to singly-infected ears at days 1 and 7 p . i . ( Fig 5E ) , which may implicate innate lymphoid cells at the infection site early during coinfection . After 28 days of infection , there were few IL-17A+ cells in the ears of all infection groups ( Fig 5A ) . However , there were significantly more IL-17A+ CD4+ ( non-γδ ) T cells in the draining LNs of coinfected mice compared to mice infected with either L . major or S . aureus alone at 28 days p . i . ( Fig 5D ) . These data suggest the source of IL-17A differs in the acute versus chronic L . major-S . aureus coinfected lesions . The correlation between L . major-S . aureus coinfection exacerbation and elevated IL-17A from γδ T cells and Th17 cells led us to examine potential upstream factors promoting an IL-17A response in different cell types . Inflammasome-derived IL-1β can promote the expression of IL-17A by γδ T cells [45 , 46] and Th17 cells [47 , 48] , as well as the differentiation of naïve T cells into Th17 cells [49] . Pro-IL-1β is upregulated at a transcriptional level by priming conditions , and secreted IL-1β must undergo proteolytic cleavage by caspase-1 in order to be biologically active [50] . Additionally , IL-6 and particularly IL-23 are important for the differentiation and maintenance of Th17 cells [51 , 52] . We used flow cytometry to measure pro-IL-1β abundance in inflammatory cells extracted from the ears of singly or coinfected mice . After 1 or 7 days p . i . , we observed significantly more CD45+ CD11b+ myeloid cells with intracellular pro-IL-1β in coinfected ears compared to S . aureus-infected ears after 1 or 28 days , and compared to L . major-infected ears after 1 , 7 , or 28 days p . i . ( Fig 6A ) . Most of the pro-IL-1β+ CD11b+ cells during early infection showed surface marker staining consistent with neutrophils ( CD45+ CD11b+ Ly6Ghi Ly6Cint ) ( Fig 6B ) . The abundance of active IL-1β by ELISA in coinfected ears at day 1 and day 3 p . i . compared to all other infection groups ( Fig 6C ) . These increased levels of IL-1β early during coinfection may be due to the elevated S . aureus burdens observed during the early stages of coinfection with L . major . Expression of IL-6 at day 3 , and IL-23 at days 3 and 7 was assessed by RT-qPCR . These assays revealed significantly greater abundance of IL-6 mRNA in the ears of the L . major single infection group , but lower IL-23 mRNA in L . major single infected mice , compared to myeloid cells from the ears of mice infected either with S . aureus alone or coinfected with L . major and S . aureus ( Fig 6D and 6E ) . These observations suggest that IL-23 and IL-1β responding to S . aureus in the skin might contribute to enhanced production of IL-17A from innate T cells at early times of infection . As a corollary , augmented IL-17A early during infection might drive the differentiation of naïve T cells into Th17 cells in the later stages of coinfection . The coinfection phenotype was not limited to the S . aureus Newman strain . L . major IA-2 coinfections with S . aureus MNPE , a methicillin-sensitive USA200 strain that produces the superantigen toxic shock syndrome toxin ( TSST-1 ) , recapitulated the lesion exacerbation seen with S . aureus Newman MSSA ( Fig 7A ) . These lesions revealed more highly elevated IL-17A production by γδ T cells in coinfection ( Fig 7B ) . Thus , two clinically relevant S . aureus isolates that express different secreted toxins exacerbate lesions in a murine model of cutaneous leishmaniasis in a similar manner . We hypothesized that the increased production of IL-17A during L . major-S . aureus coinfection may contribute to the observed enhanced pathology , and that IL-17A expression might be stimulated by IL-1β . We tested this hypothesis by treating L . major singly infected or L . major-S . aureus coinfected mice with neutralizing anti-IL-17A or anti-IL-1β antibodies through the first two weeks infection ( Fig 8A and 8B ) . Coinfected groups treated with anti-IL-17A antibody developed two-fold reduction in early lesion volumes compared to mice treated with isotype control antibody at day 2 p . i . ( Fig 8A ) , consistent with a lesion-exacerbating effect of IL-17A . However , coinfected groups treated with anti-IL-1β antibody developed two-fold larger lesion volumes compared to isotype control treated mice at day 4 p . i . ( Fig 8B ) . This was not due to an inability of anti-IL-1β antibodies to reach the ear skin site , although antibodies were only able to decrease the level of IL-1β in the ears by 64% ( S10 Fig ) . Neutralizing antibodies to IL-17A or IL-1β had no effect on ear lesions of mice infected with L . major alone , suggesting they influence immune responses elicited in the presence of bacteria . The anti-IL-17A data suggest that IL-17A is partially responsible for the lesion exacerbation that occurs during coinfection , but this does not exclude a contribution of other immune mediators and bacterial or parasitic factors to the development of disease . We hypothesized that the results of neutralizing IL-1β may reflect competing actions of IL-1β in this model . IL-1β promotes clearance of S . aureus from skin lesions , an observation illustrated by our control mice singly infected with S . aureus alone , in which anti-IL-1β caused a significant exacerbation of bacterial loads ( Fig 8C ) . A second effect of IL-1β may be to induce IL-17A , which in turn would recruit neutrophils to lesion sites . We surmise that at the early time points shown in our study , the defective bacterial killing in the face of diminished IL-1β in the vicinity might dominate over the effects of lowering IL-17A responses in terms of lesion size . The data suggest that the several proinflammatory functions of IL-1β may participate in controlling bacterial replication and lesion size during the early phases of cutaneous L . major-S . aureus coinfection .
Determinants of the outcome of host defense against infection include input from both host and microbe and interactions that are bidirectional . The immune environment into which a pathogen is introduced plays a critical role in the fate of microbial infection [4] . Recent studies on microbiomes have demonstrated that commensal microbes present at a non-sterile body site , such as the skin , modulate local immune environments [4–6 , 15 , 53] . Adding to these complex and reciprocal interactions , studies on polymicrobial infections have revealed a pattern of synergy between coinfecting organisms with overlapping biogeography [54–57] . In the case of vector-borne infectious diseases such as cutaneous leishmaniasis , microbiota from the non-sterile gut environment of the insect vector can also be introduced in concert with the parasite into mammalian skin [12 , 16] . The current study tested the hypothesis that the simultaneous presence of bacteria on mammalian skin at the site of inoculation with Leishmania spp . modifies the local immune response , either enhancing parasite killing or exacerbating leishmanial disease . Simultaneous intradermal coinfection of C57BL/6 mice with L . major and S . aureus resulted in exacerbation of lesions characterized by the presence of more neutrophils . The timing and site of coinfection were critical in our studies , because coincident infection with S . aureus inoculated prior to L . major , or concurrently at a different body site , did not alter the course of severity of L . major lesion development or burdens . These findings mirror results of a study of concurrent infections of hamsters with Leishmania braziliensis panamensis and either S . aureus or Pasteurella multocida , which demonstrated enhanced early lesion size and bacterial burdens , although late lesions were not changed [58] . Other models of cutaneous leishmaniasis , including studies of gene knockout mice ( NLRP10 and TNFRp55 ) [59–61] describe enhanced lesion severity without an associated increase in L . major burden [59 , 62] . In contrast to the absence of an increase in parasite load , augmented lesion size during early coinfection correlated with elevated S . aureus burdens in the first three days of coinfection , despite the enhanced numbers of neutrophils , cells that typically ingest , kill , and eliminate S . aureus from tissues . Compared to single infections , L . major-S . aureus coinfection resulted in the presence of more neutrophils both in the early and the later stages of infection , times corresponding to the peaks of pathology due to S . aureus or L . major , respectively . Because of the paradoxical increase in the numbers of S . aureus in the face of elevated neutrophil numbers , we assessed neutrophil functions critical for antimicrobial action . At day 1 p . i . , neutrophils recovered from L . major-S . aureus coinfected lesions phagocytosed either pathogen and had a functional NADPH oxidase , evidence suggesting that the antimicrobial machinery of recruited neutrophils was intact . Although both S . aureus and L . major can extend neutrophil lifespan [40 , 63 , 64] , we found that a high percentage of neutrophils from S . aureus-containing lesions , either alone or with L . major , were apoptotic . In contrast , similar percentages of viable ( i . e . AnnV-PI- ) were recovered in all experimental conditions . Considering the markedly elevated number of total neutrophils in coinfected compared to singly infected groups , the net result was a large number of apoptotic neutrophils in L . major-S . aureus coinfected groups . Typically , apoptotic neutrophils are cleared by efferocytosis . However , expression of several transcripts associated with efferocytosis and wound healing , such as IL-10 , IL-13 , and MMP9 , were downregulated at 3 days p . i . in S . aureus and coinfected ears . Taken together , these data demonstrate that neutrophils recruited to the sites of coinfection with L . major-S . aureus underwent apoptosis but were not cleared , thus contributing in part , to the large numbers of neutrophils in lesions but with failure to kill S . aureus or resolve the inflammation . Although the mRNA transcript profiles for our selected chemokines/cytokines were similar between L . major singly or coinfected mouse ears at 28 days p . i . , we found more IL-17A production from draining LN cells of coinfected mice in response to L . major antigen during coinfection . It is well established that acute S . aureus infection upregulates IL-17A , which contributes to the recruitment of neutrophils and consequently clearance of the pathogen [65] . Cutaneous bacteria can have a major impact on host immune status and the development of a Type 17 response to pathogens . Germ-free mice have fewer IL-17A-producing T cells at baseline and display less pathology in response to L . major infection despite higher parasite loads . In contrast , germ-free mice recolonized with S . epidermidis had restored numbers of IL-17A+ T cells in the skin and L . major lesions size similar to that of specific pathogen-free mice [4] . Th17 cells and IL-17A are also implicated in the immunopathology of murine models of CL [66–70] . Susceptible BALB/c mice have higher IL-17A levels in L . major lesions compared to resistant C57BL/6 mice [67] . Importantly , the levels of IL-17A produced by draining LN cells from our C57BL/6 four weeks post-single L . major infection group in response to whole promastigote L . major antigen were comparable to the levels they observed from the draining LN cells of L . major-infected C57BL/6 mice in response to soluble Leishmania antigen [67] . Although IL-17A has previously been implicated in CL , development of an elevated bystander Type 17 response to L . major in specific pathogen-free mice when coinfected with S . aureus , or the phenotypic consequences of such response , has not been described prior to the current study . Bacterial-leishmanial coinfection is likely to occur in nature when an infected sand fly bites a mammalian host . When a sand fly takes a blood meal , it generates a pool of blood in the skin due to repeated probing activity . Bacteria from host skin , the sand fly midgut , and/or the environment may be deposited along with Leishmania spp . parasites in these dermal blood pools . Dey et al . recently demonstrated that culturable bacteria are deposited by sand flies transmitting Leishmania donovani , a cause of visceral leishmaniasis . The presence of bacteria contributes to the local priming and activation of the NLRP3 inflammasome , production of IL-1β , sustained recruitment of neutrophils , and enhanced dissemination of parasites to visceral organs [12] . NLRP3 , ASC , or caspase-1/caspase-11 deficient BALB/c mice produced less IL-17A in L . major-infected footpads compared with wild-type mice [24] . Furthermore , elevated levels of IL-1β were detected in supernatants from cultured human L . braziliensis cutaneous leishmaniasis lesion biopsies compared to healthy skin controls , and a greater percentage of myeloid cells , particularly granulocytes , from lesion biopsies were pro-IL-1β+ compared to peripheral blood mononuclear cells from those same patients [21] . Thus , our observations may model the events occurring in human disease . Our results suggest that L . major coinfection with S . aureus increases expression of pro-IL-1β and activation of IL-1β , which increases neutrophil infiltration , and stimulates γδ and non-γδ T cells in the skin to produce IL-17A . Interestingly , neutrophils , which have been implicated in the immunopathology of L . major lesions [23 , 71] , were a major cellular source of pro-IL-1β during coinfection . Both single infections with S . aureus , and double infection with both pathogens led to enhanced expression of IL-23 at 7 days p . i . IL-23 is important for the maintenance of Th17 cells [51 , 52] , and may contribute to the elevated IL-17A production in mice coinfected with L . major and S . aureus for four weeks . A functional consequence of local IL-17A was suggested by antibody neutralization of IL-17A , a manipulation that partially ameliorated the exacerbated lesions observed during early coinfection . Interestingly , antibody neutralization of IL-1β did not mirror the results of anti-IL-17A treatment , but instead caused even further lesion exacerbation of coinfected lesions . This suggests that in the L . major-S . aureus coinfection context , IL-1β is important not only for promoting IL-17A production , but also the control of bacterial load . These data support a model in which L . major and S . aureus synergize to generate a microenvironment that promotes IL-1β and IL-23 production from myeloid cells , which contributes to IL-17A production by γδ and non-γδ T cells , the differentiation of naïve T cells into Th17 cells , and the continued recruitment of neutrophils . These neutrophils become apoptotic but are in an environment that delays their clearance and resolution of inflammation , which together , lead to increased lesion severity during coinfection . In summary , we found that co-inoculation of Staphylococcus aureus with Leishmania major at the same site in a murine skin infection model increased the S . aureus bacterial load but did not alter the local burden of the protozoan . Nonetheless , there was a local activation of IL-17A-mediated inflammatory responses and corresponding neutrophil recruitment to the site of coinfection , with exacerbation of both early and late phases of the inflammatory lesion . The large numbers of neutrophils recovered from the coinfections reflected increased recruitment due to IL-17A and decreased efferocytosis and clearance , though the molecular basis for this is not known . The exacerbated lesion size corresponded to an IL-17A response , which developed from an innate response of skin γδ and non-γδ T cells during the first week of infection , into an adaptive response of L . major antigen-responsive Th17 cells at later time points ( Fig 9 ) . These findings provide insight into the interactions that occur between L . major and microbiota from the sand fly midgut or mammalian host skin , and may contribute to the development of novel therapies that reduce immunopathology during cutaneous leishmaniasis .
|
Cutaneous leishmaniasis ( CL ) is a vector-borne ulcerating skin disease affecting several million people worldwide . The causative Leishmania spp . protozoa are transmitted by infected phlebotomine sand flies . During a sand fly bite , bacteria can be coincidentally inoculated into the dermis with the parasite . Staphylococcus aureus is the most common bacterium in CL skin lesions . Symptomatic CL is characterized by papulonodular skin lesions that ulcerate and resolve with scarring , although most cutaneous Leishmania infections are asymptomatic . We sought to explore factors that determine whether infection with a cutaneous Leishmania species would result in symptomatic CL rather than asymptomatic infection . We hypothesized that local bacteria promote the development of symptomatic CL lesions during infection with Leishmania major . We discovered that cutaneous lesions were significantly larger in mice inoculated simultaneously with S . aureus and L . major than in mice infected with either organism alone . Coinfection led to increased S . aureus growth in skin lesions , whereas L . major parasite numbers were unchanged by coinfection . The size of the exacerbated lesion correlated with early increased numbers of neutrophils and elevated levels of proinflammatory cytokines IL-1β and IL-17A during the first 7 days , and with sustained increases in IL-17A through 28 days of coinfection . Neutralizing antibody experiments suggested IL-17A was partially responsible for lesion exacerbation during coinfection , whereas IL-1β was important for both control of early lesion exacerbation and promotion of IL-17A production . These data suggest that treatment of symptomatic CL targeting the parasite , local commensal bacteria , and host proinflammatory IL-17A immune responses might improve the outcome of CL .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"blood",
"cells",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"pathology",
"and",
"laboratory",
"medicine",
"ears",
"pathogens",
"immunology",
"microbiology",
"parasitic",
"diseases",
"staphylococcus",
"aureus",
"animal",
"models",
"otology",
"model",
"organisms",
"ear",
"infections",
"experimental",
"organism",
"systems",
"bacteria",
"neutrophils",
"bacterial",
"pathogens",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"white",
"blood",
"cells",
"animal",
"cells",
"staphylococcus",
"medical",
"microbiology",
"animal",
"studies",
"microbial",
"pathogens",
"t",
"cells",
"mouse",
"models",
"head",
"otorhinolaryngology",
"cell",
"biology",
"anatomy",
"co-infections",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"organisms"
] |
2019
|
Coinfection with Leishmania major and Staphylococcus aureus enhances the pathologic responses to both microbes through a pathway involving IL-17A
|
Assembly of kinetochore complexes , involving greater than one hundred proteins , is essential for chromosome segregation and genome stability . Neocentromeres , or new centromeres , occur when kinetochores assemble de novo , at DNA loci not previously associated with kinetochore proteins , and they restore chromosome segregation to chromosomes lacking a functional centromere . Neocentromeres have been observed in a number of diseases and may play an evolutionary role in adaptation or speciation . However , the consequences of neocentromere formation on chromosome missegregation rates , gene expression , and three-dimensional ( 3D ) nuclear structure are not well understood . Here , we used Candida albicans , an organism with small , epigenetically-inherited centromeres , as a model system to study the functions of twenty different neocentromere loci along a single chromosome , chromosome 5 . Comparison of neocentromere properties relative to native centromere functions revealed that all twenty neocentromeres mediated chromosome segregation , albeit to different degrees . Some neocentromeres also caused reduced levels of transcription from genes found within the neocentromere region . Furthermore , like native centromeres , neocentromeres clustered in 3D with active/functional centromeres , indicating that formation of a new centromere mediates the reorganization of 3D nuclear architecture . This demonstrates that centromere clustering depends on epigenetically defined function and not on the primary DNA sequence , and that neocentromere function is independent of its distance from the native centromere position . Together , the results show that a neocentromere can form at many loci along a chromosome and can support the assembly of a functional kinetochore that exhibits native centromere functions including chromosome segregation accuracy and centromere clustering within the nucleus .
Genome stability requires accurate chromosome segregation . Faithful chromosome segregation requires the assembly of a kinetochore complex on the centromere DNA region of each chromosome . The kinetochore is a large complex of more than 100 proteins and is essential for the attachment of the spindle microtubules to each chromosome during cell division [1] . Defects in chromosome segregation accuracy can cause DNA damage and chromosome rearrangements as well as aneuploidy , an imbalance in the numbers of individual chromosomes [2 , 3 , 4] . In most eukaryotes , the mechanisms that specify centromeres and that direct kinetochore assembly to a particular chromosomal region are epigenetic , rather than strictly sequence-dependent . CENP-A , a variant histone H3 protein , is an essential component of centromeric chromatin ( reviewed in [5] ) . In addition to the presence of CENP-A at centromeres , centromeric chromatin is marked by other histone modifications . For example , histone H3K9 methylation and other indicators of heterochromatin mark pericentromere regions in humans and many other species [5] . Hypoacetylation of histone H4 is associated with gene silencing and is observed at centromere chromatin in budding yeast [6] . In Schizosaccharomyces pombe , the kinetochore mediates silencing of marker genes within the central core of the centromere [7] . However , recent data suggest that despite the association of many repressive marks at centromeres , transcription of non-coding RNA within the central core of centromere sequences is required for normal centromere function ( reviewed in [8] ) . Transcription at centromeres must be carefully regulated because transcriptional levels that are either too low or too high are detrimental to kinetochore assembly [5 , 9] . However , we do not know how these optimal transcriptional levels are maintained nor whether kinetochore assembly has a direct role in regulating transcription . Recent work has highlighted a consistent feature of functional centromeres in many organisms including S . cerevisiae , Drosophila and humans: they cluster together within a specific region within the 3D organization of the nucleus [10 , 11] . Centromere clustering provides a defining feature of yeast centromeres that has been used to identify centromeres in fungi with uncharacterized centromeres [12 , 13] . In S . cerevisiae and C . albicans , centromere clustering to a single focus is dependent on kinetochore-microtubule interactions , as strains lacking kinetochore components such as the Dam1 complex have clustering defects [14 , 15 , 16 , 17] . In other organisms including Drosophila [10] , mouse , and human , centromeres cluster to multiple nuclear locations [11] . In Drosophila , clustering requires nucleoplasmin-like protein ( NLP ) and the insulator protein CTCF [10] . Interestingly , in Drosophila , tethering of kinetochore proteins to a plasmid causes association with the clusters [10] . Also , interfering with clustering disrupts pericentric heterochromatin causing increased expression of pericentric repeats [10] . This suggests that centromere clustering also may be important in transcriptional regulation at centromeres . The position of the centromere on a given chromosome is inherited , such that syntenic centromeric loci are detected in related species [18 , 19 , 20] . Although the position of a centromere is generally stable through many generations , chromosome rearrangements , deletions , or amplifications sometimes form acentric chromosome fragments . Neocentromeres that assemble de novo at DNA loci not previously associated with kinetochore proteins can restore the ability of an acentric chromosome fragment to segregate efficiently [21] . In rare cases , neocentromeres form in otherwise normal chromosomes , without physical deletion of the native centromere , presumably following inactivation of the native centromere through unknown mechanisms [22 , 23] . Evidence of centromere repositioning is observed rarely in human patients , but has been detected as “evolutionary new centromeres” in the genomes of humans , macaques , and donkeys [20 , 24] . Evolutionary new centromeres are repositioning events that become fixed in the population and are thought to be important steps in speciation [19 , 25] . More than 100 human neocentromere locations have been identified [26] , with the majority found in patients with developmental disabilities and others found in cancer tissues [26] . For example , neocentromeres are characteristic cytogenetic features of well-differentiated liposarcomas [27] . Recent work has identified neochromosomes , many of which are predicted to have neocentromeres , in approximately 3% of cancers [28] . Several model systems have been developed to study neocentromere formation and function including Drosophila [29] , S . pombe [30 , 31] , C , albicans [32 , 33] , and chicken cells [34] . Neocentromere locations in Drosophila and S . pombe are limited to specific chromosomal domains . For example , neocentromeres in Drosophila have been identified at pericentric regions [29] , and mature neocentromeres in S . pombe form most frequently at subtelomeric regions and require adjacent heterochromatin for functionality [31] . Neocentromeres in humans and in chicken DT40 cells localize to diverse positions , many of which lack adjacent heterochromatin [34] . Thus , the range of possible neocentromere positions changes in different systems . C . albicans has been established as a model for neocentromere formation . The small , regional centromeres of C . albicans all have unique DNA sequences of approximately 3-5kb bound by CENP-A [35] . Several centromeres , most predominately centromere 5 ( CEN5 ) , are flanked by inverted repeat sequences unique to that centromere [32] . Following deletion of native centromere DNA sequences , functional kinetochores assemble , evidenced by the appearance of CENP-A and other kinetochore proteins at new loci [32 , 33] . Neocentromeres also specify early replication timing , similar to native centromeres [36] . Neocentromeres can form either proximal or distal to the native centromere [32 , 33] . Neocentromere positions are inherited from one generation to the next , but neocentromere positions are less stable than native centromere positions . At low frequency , neocentromere positions shifted locally as detected by reversible silencing of a URA3 gene in proximal neocentromere strains . Additionally , in one transformant from Ketel et al . , the isolate was saved prior to neocentromere position stabilization and multiple neocentromere positions were isolated from a single transformant [32] . In most systems it has been difficult to compare the function of neocentromeres to native centromeres . In humans , some neocentromeres appear to be more prone to chromosome segregation errors than native centromeres . One characterized human neocentromere also has defects in the localization of Aurora B kinase , an essential regulator of kinetochore-microtubule attachments , and in error correction [37] . Neocentromere mosaicism , defined as the presence of the neocentromere in a subset of somatic cells , suggests that the chromosome carrying the neocentromere was lost in a subpopulation of the cells . Based on the available data in humans , it is not clear if the mosaic neocentromeres are due to processes related to the formation of the neocentromere , selective disadvantages of maintaining the neocentric chromosome , and/or defects in segregation accuracy of the neocentromere [26] . Importantly , other neocentromeres are found consistently in all patient tissues and appear to segregate accurately [26] . The possibility that different neocentromere loci have different levels of chromosome segregation accuracy is intriguing , but technical issues , such as differences in genetic background between individuals and difficulty in quantitating chromosome segregation , complicate rigorous comparisons of human neocentromeres . Using C . albicans as a model system allows us to eliminate both of these obstacles . First , all C . albicans neocentromeres can be isolated from the same parental strain , which reduces the effect of genetic diversity . Second , a sensitive method to quantify small to moderate increases in chromosome loss is readily available , based upon selection for loss of the URA3 marker gene by growth of cells on 5-fluorourotic acid ( 5-FOA ) [38] . In this work , we characterized twenty neocentromere loci on C . albicans chromosome 5 ( Chr5 ) . These neocentromeres were assembled at intergenic regions as well as at loci containing ORFs , where the neocentromere repressed ORF transcription . Some , but not all neocentromere strains had higher chromosome loss rates than strains with native centromeres . Thus , as in humans , neocentromeres in C . albicans can have variable degrees of functionality at different loci . Finally , neocentromere formation drives reorganization of interchromosomal interactions , such that the functional neocentromere , like native centromeres on unperturbed chromosomes , clusters with active native centromeres on other chromosomes . This indicates that the three-dimensional ( 3D ) organization of centromere clustering is a dynamic process and is dependent upon epigenetic kinetochore function rather than upon DNA sequence in C . albicans .
Strains that survive deletion of the 7 . 6kb centromere region on Chr5 , which includes the central core and both flanking inverted repeat sequences , form neocentromeres at locations proximal to ( within 4kb of the deleted region ) or more distal to the native centromere locus [32 , 33] . In Ketel et al . , proximal neocentromere strains all were centered at 464 . 5kb and two independent transformants resulted in four distal neocentromeres at loci along the length of Chr5 [32] . Thakur and Sanyal ( 2013 ) also deleted a 7 . 2kb region of CEN5 and all 6 neocentromeres characterized were centered nearby the deleted sequence at ~459kb and ~478kb [33] . To ask if neocentromere loci are limited to specific chromosome arm regions , we isolated additional transformants in which CEN5 sequences were replaced with URA3 ( S1 Fig ) . Combined with the transformants described in Ketel et al . [32] , ~ 50% of transformation events resulted in distal neocentromeres ( S1 Fig ) . Previously , three of the four distal neocentromere positions were isolates obtained from a single transformant stock from neocentromere movement or displacement following sorbose treatment , a nutrient stress condition resulting in high rates of homozygosis of Chr5 [32 , 39] . In addition to neocentromere movement following stress , colony purification of a single transformant revealed sub-clones with different neocentromere positions . Subsequently , some of these neocentromeres were detectable in bulk analysis of the original stored stock . Thus , we hypothesized that , immediately following transformation , neocentromeres may be unstable and that subpopulations of a transformant colony might contain different neocentromere loci . Therefore , in addition to testing multiple colonies from the new transformants , we also isolated and characterized additional single colonies from previously published transformants with distal neocentromeres . We then identified the positions of the neocentromeres by chromatin immunoprecipitation ( ChIP ) with anti-CENP-A antibodies on the newly isolated strains as well as on additional isolates of the previous transformants ( S2 Fig ) . In addition , the neocentromere position originally identified at ~170kb in strains YJB10779 and YJB10780 [32] was mapped at higher resolution to two adjacent , non-overlapping neocentromere positions centered at 173 . 5kb and 166kb , respectively ( S3 Fig ) . One distal neocentromere position was identified in isolates from two independent transformants in our lab , and two neocentromere positions overlapped with the neocentromere positions identified in Thakur and Sanyal [33] . All other neocentromere positions were observed only from a single transformant , indicating that the screen for neocentromere positions has not yet saturated all possible loci , but that positions capable of supporting neocentromere function are likely not infinite . Together , this brings the total number of neocentromere positions to twenty including the proximal neocentromere position at 464 . 5kb ( Fig 1A , S1 Table ) . Using these twenty neocentromeres , we searched for DNA sequence features that could significantly distinguish neocentromere loci from native centromeres and/or from all other regions of Chr5 . While neocentromere regions are more variable in size than native centromere regions , no significant differences were found between the size of the CENP-A bound DNA sequence in neocentromeres and native centromeres ( unpaired t-test , p>0 . 05 ) ( Fig 1B ) . The GC% values ± SD for neocentromeres ( 33 . 7 ± 3 . 3% ) , native centromeres ( 35 . 0 ± 1 . 0% ) , or size-matched random DNA regions on chromosome 5 ( 32 . 6 ± 2 . 2% ) also were not significantly different ( one-way ANOVA , p>0 . 05 ) . Skew inversions for G/C distribution were previously identified at C . albicans centromeres as fossils of the long-term presence of early origins of DNA replication ( S4A Fig ) [36] . This is thought to occur because leading strands and lagging strands cause biased rates of C→G transversions and this bias would ‘flip’ at a constitutive origin where leading strands emerge in opposite directions [40] . Importantly , neocentromere formation promotes early/efficient replication initiation [36] , yet neocentromeres have not been constitutive early origins over the long time scales necessary to accumulate skew inversion patterns ( where the skew level crosses the X-axis 0 line , S4 Fig ) ; consistent with this , we did not identify consistent G/C skew inversion patterns at the neocentromere loci ( S4B Fig ) or at random loci ( S4C Fig ) . Native centromeres in C . albicans are associated with three different types of repeat elements: inverted repeats , tandem repeats , and transposon-associated repeats ( S5A Fig ) and repeat elements are important in de novo kinetochore assembly on a plasmid in the closely related species C . tropicalis [41] . Yet the distance between the center of each neocentromere and the closest repeat element ( mean distance ± SEM , 3622 ± 745bp ) was not significantly different from the corresponding distances of random loci and repeat elements on Chr5 ( 4165 ± 1013bp ) ( unpaired t-test , p>0 . 05 ) ( S5B Fig ) . Early observations with small numbers of distal neocentromeres suggested that neocentromeres formed in intergenic regions [32] . However , more detailed mapping and the increased number of neocentromere strains revealed neocentromeres that mapped within ORFs as well . Indeed , neocentromere regions and random sequences were similarly likely to be intergenic versus genic ( unpaired t-test , p>0 . 05 ) ( Fig 1C ) . ORFs overlapped with the neocentromere position by at least 100bp for 19 of the 20 neocentromere positions ( S1 Table ) . Using more conservative criteria , 17 of 20 neocentromeres have >500bp 5’ ORF overlap , >1000bp 3’ ORF overlap , or overlap of the entire ORF . The relationship between kinetochore assembly and transcription is complex , as low transcription levels benefit centromere function and high levels of transcription are incompatible with the presence of a functional kinetochore [8 , 9] . We next asked if ORFs with the potential to become neocentromeres are transcribed under standard laboratory conditions , by analyzing previously published RNA-seq data for C . albicans grown in YPD medium at 30°C [42] . C . albicans were grown in YPD immediately preceding the centromere deletion event that promoted neocentromere formation , so these transcription levels likely represent the transcription state of the cell prior to the induction of neocentromere formation . In the Bruno et al . data set , transcripts in YPD grown cells ranged from -4 . 6 to 13 . 6 ( on a log2 scale of reads per kilobase per million mapped reads ( RPKM ) ) . Interestingly , 13 of 20 neocentromere positions have transcripts within the neocentromere DNA region at levels equal to or greater than the median expression level of all ORFs in YPD in the RNA-seq data set ( S1 Table ) . This suggests that neocentromeres assembled at ORFs that are normally transcribed in the context of native chromosomes . To ask if neocentromeric chromatin suppresses the expression of genes within the CENP-A binding region as has been observed with marker genes at native centromeres in S . pombe [32 , 43] , we measured transcription levels of native genes within defined neocentromere regions . We conducted qRT-PCR in strains with an active neocentromere at the given locus , and at least two other strains with neocentromeres formed at other Chr5 loci . Importantly , for a given region , transcript levels at an active neocentromere were lower than those at the same loci in strains without the neocentromere at that locus ( Fig 2A–2C and S6A–S6F Fig ) . By contrast , a neighboring transcript just outside the CENP-A binding region ( Fig 2D ) , a transcript with the promoter >500bp from the CENP-A binding region ( S6G Fig ) , and a transcript of a gene on a different chromosome ( S6H Fig ) showed no detectable difference in expression among three neocentromere strains strains . Thus , active neocentromeres suppress transcription at loci where they are assembled . In humans , different neocentromeres appear to have different degrees of chromosome segregation accuracy ( reviewed in [26] ) . To directly test whether different neocentromeres have different chromosome segregation accuracy , we compared chromosome segregation by following the loss of URA3 in 12 heterozygous neocentromere strains where one copy of Chr5 maintains the native centromere and the other copy of Chr5 contains the neocentromere and the URA3 marker . The URA3 loss rate for the native centromere strain was approximately 1 . 0e-05 ( Fig 3A ) . URA3 loss rates for neocentromere strains ranged from approximately 4 . 6e-06 to 6 . 2e-04 ( Fig 3A ) . Thus , some neocentromere strains had URA3 loss rates very similar to the native centromere strain , while others had increased URA3 loss rates of up to 60-fold higher than the native centromere strain . ANOVA analysis indicated that different neocentromere positions have statistically significant differences in URA3 loss rate ( p<0 . 01 ) . URA3 loss is observed as the combined consequence of chromosome loss , shorter range recombination events and loss-of-function mutations in the URA3 gene , which can be distinguished by SNP-RFLP analysis of markers on both arms of Chr5 . In the seven neocentromere isolates with the highest URA3 loss rates , loss of heterozygosity across all markers tested on Chr5 was elevated compared to the homozygosis of markers observed in strains with segregation driven by native centromere loci [44] ( S2 Table ) . Thus , URA3 was primarily lost via increased whole chromosome loss ( homozygosis of all tested Chr5 SNP-RFLP markers ) , rather than an increase in recombination events ( homozygosis of only some of the markers along Chr5 ) in neocentromere strains . We next asked whether neocentromere function , measured as chromosome segregation accuracy , correlated with any of the other characteristics of neocentromere loci . No significant correlations were found for chromosome loss rates ( Fig 3 ) relative to neocentromere length or the distance between the neocentromere and the nearest repeat element ( S7A and S7B Fig ) . There was a slight positive correlation between chromosome loss rates and the fraction of the CENP-A binding region containing ORF sequences ( S7C Fig ) . The correlation was much stronger when the fraction of ORF overlap and the RNA-seq transcription data were combined to estimate total transcriptional activity in the region prior to neocentromere formation ( R2 = 0 . 71 ) . Higher transcriptional activity correlates positively with higher chromosome loss ( Fig 4A ) . This indicates that better neocentromere function is associated with DNA positions that normally have lower levels of transcription . Certain types of stress , such as exposure to high temperature , elevate whole chromosome loss rates in C . albicans [44 , 45] . As expected , growth at 39°C significantly increased URA3 loss rates ( p<0 . 01 ) compared to growth at 30°C for neocentromeres as well as for native centromeres ( Fig 3B ) . Again , whole chromosome loss was the likely mechanism based upon SNP-RFLP analysis of several markers ( S2 Table ) . Interestingly , the fold-change in URA3 loss between 30°C and 39°C was lower for the neocentromere strains , perhaps because chromosome loss rates were already elevated relative to the native centromeres at 30°C . Specific chromatin modification patterns are required for accurate chromosome segregation . In Saccharomyces cerevisiae , nicotinamide treatment inhibits NAD+ histone deacetylation by Sir2-family proteins , resulting in centromere dysfunction [6] . Pericentromeric regions in C . albicans have elevated levels of H4K16 acetylation relative to the central core region [46] . Treatment with 2mM nicotinamide increased the relative level of H4K16 acetylation within the centromeric central core ( S8A Fig ) . Thus , we asked if 2mM nicotinamide affected Chr5 loss rates in C . albicans with normal CEN5 or different Chr5 neocentromeres . Overall , loss rates for native centromere strains and for 6 of 7 neocentromere strains increased with nicotinamide exposure ( ANOVA , p<0 . 01 ) , measured as URA3 loss rates ( S8B Fig ) . Treatment with 100μM nocodazole , an inhibitor of microtubule polymerization , also increased URA3 loss rates for all tested neocentromeres ( S9 Fig ) . In both drug treatments , the majority of the Ura- isolates had SNP-RFLP markers indicating whole chromosome loss ( S2 Table ) . Similar to heat stress , the fold-change between no drug and nicotinamide treatment or nocodazole treatment was less for the neocentromere strains , perhaps due to the higher initial chromosome loss rates . Together these results indicate that neocentromere strains are not hypersensitive to factors that disrupt centromere function . Nonetheless , the same types of stresses and drugs that affect native centromeres affect most neocentromere strains . Neocentromere formation can occur along the entire length of the chromosome in C . albicans ( Fig 1A ) including , but not restricted to positions proximal to the native centromere [32 , 33 , 34] . Importantly , this data clearly refutes the suggestion that only neocentromeres close to the native centromere position are truly functional in C . albicans [33] . Furthermore , a comparison of URA3 loss rates at different neocentromeres found no correlation between chromosome loss and distance of the neocentromere from the native centromere position ( Fig 4B ) . Therefore , neocentromeres both close and far from the native centromere position can be highly functional in C . albicans . Yeast centromeres cluster in a single nuclear location , providing a driving force for the nuclear organization in fungi [47 , 48] . Mapping of chromosomal interactions with the chromatin conformation capture assay Hi-C is an effective way to identify functional centromere regions based on their 3D colocalization with one another [12 , 13 , 49 , 50] . Therefore , we tested the hypothesis that centromere clustering is an epigenetic feature of centromere function and is independent of physical or sequence-based features of the native centromere position . This hypothesis predicts that neocentromere formation would reorganize interchromosomal interactions , bringing the newly formed neocentromere on one chromosome with the native centromeres on the remaining chromosomes . We used Hi-C to identify all chromatin interactions for a strain with all centromeres at native locations and for two strains with an active neocentromere at different positions: one with homozygous neocentromeres centered near the left telomere at 4 . 5kb ( YJB10777 ) and one with homozygous neocentromeres centered at 166kb ( YJB10780 ) . We mapped Hi-C data from these three strains to the C . albicans reference genome and further processed the mapped read pairs to produce raw and normalized Hi-C contact maps ( Methods ) . In the wild-type C . albicans strains , native centromeres on all chromosomes clustered with one another in 3D ( Fig 5 ) . This clustering was evident from the strong enrichment of centromere interactions , apparent both in the raw ( data available via the Short Read Archive ) and in the normalized interchromosomal contact maps ( Fig 5A ) , as well as from the peaks of interchromosomal interactions near all centromere pairs when the native CEN5 location was used as the interaction probe to create virtual 4C plots ( Fig 5B ) . The peaks between all centromere pairs were conserved when any other native centromere was used as the probe ( S10 Fig ) . These interchromosomal interaction patterns are similar to what has been seen in other yeast species [49] . Next , centromere positions were predicted solely from the Hi-C data using the Centurion algorithm [49] . For data from the wild-type strain , six of eight centromere midpoint predictions fell within the boundaries and the remaining two were within 2kb of native centromere positions estimated by CENP-A ChIP mapping ( S3 Table ) . This result suggests that the generated Hi-C data provides sufficient information to accurately locate the centromeres . In the wild-type strain data ( prior to neocentromere formation ) , regions where neocentromeres could form did not exhibit strong interactions with other centromeres . The DNA region near 4 . 5kb showed local Hi-C interactions with neighboring regions on Chr5 and some interactions with other telomeric regions , but not with centromeres in the wild-type strain ( Fig 6A ) . Similarly , the DNA region near 166kb showed local interactions with neighboring regions on Chr5 , but not with native centromeres in the wild type ( Fig 6B ) . In contrast , in the two strains with homozygous active neocentromeres centered at 4 . 5kb and 166kb ( in which native CEN5 had been deleted ) , the neocentromere regions clustered with centromeres on other chromosomes . In the strain with the neocentromere at the 4 . 5kb locus , the DNA region ~ 4 . 5kb from the left telomere interacted with all other centromeres ( Fig 6C ) . Similarly , the homozygous active neocentromere near 166kb interacted with the native centromere loci of all other chromosomes ( Fig 6D ) . Furthermore , genome-wide analysis of interactions demonstrated that the native centromere region on Chr5 did not interact with other centromeres and that centromeres on all chromosomes established reciprocal interactions with the neocentromere positions ( Fig 7 and S11 Fig ) . In addition , the Centurion algorithm accurately predicted the neocentromere positions ( as previously determined by chromatin immunoprecipitation with the centromere-specific histone CENP-A ) further demonstrating that neocentromere formation was accompanied by an overall change in chromosome organization ( S3 Table ) . Thus , neocentromeres acquire an important phenotype that is characteristic of centromeres—the ability to cluster with native centromeres on all other chromosomes .
Many loci on Chr5 are capable of supporting neocentromere formation . The most frequent position is immediately adjacent to the native centromere [32] . This proximal neocentromere position exhibits reversible silencing of the URA3 marker used to delete the centromere and inverted repeat sequences [32] . Indeed , ~50% of transformants tested exhibited URA3 silencing ( 11/22 ) , indicative of proximal neocentromere formation ( S1 Fig ) [32] . Silencing was only observed when the center of the CENP-A region was within 2kb of the deleted sequence . In addition , three neocentromere positions did not exhibit URA3 silencing , but were located within 30kb of the native centromere position ( S1 Table ) . The increased likelihood of neocentromere formation near the native centromere , compared to any other single location on Chr5 , is similar to neocentromere formation tendencies on chromosome Z in chicken DT40 cells where 76 . 2% of neocentromeres formed near the native centromere position [34] . The regions near the native centromere are potentially enriched for non-centromeric CENP-A in a “CENP-A cloud” [34] . Furthermore , these centromere proximal regions are in closer 3D proximity to one another in the wild-type nucleus ( Fig 5 ) , which may increase the opportunity for closer sequences to capture kinetochore proteins upon their release when the native centromere is deleted . Importantly , we characterized sixteen neocentromere loci across the chromosome with kinetochore assembly occurring more than 30kb from the native centromere ( S1 Table ) [32] . These neocentromeres are functional , not only in chromosome segregation but also in centromere clustering , and thus are clearly able to serve as active neocentromeres . This is in contrast to the suggestions of Thakur and Sanyal ( 2013 ) , who analyzed 6 neocentromeres on Chr5 , and argued that C . albicans only forms neocentromeres near the native centromere [33] . Clearly , if sufficient numbers of neocentromeres are collected and analyzed , a significant proportion of them form active neocentromeres at distal positions on Chr5 greater than 30kb from the native centromere . Once formed , neocentromeres on all regions of the chromosome promote chromosome segregation . Six of twelve tested neocentromeres had loss rates within 5-fold of native centromeres , further demonstrating that many distal neocentromere loci are active and functional ( Fig 3 ) . The functional similarity between distal neocentromeres and native centromeres is reinforced by their similar increase in chromosome loss rates in response to stressors such as high temperature , NAD-dependent histone deacetylase inhibitors , and the microtubule-destabilizing agent nocodazole ( Fig 3 , S8 Fig and S9 Fig ) . Strikingly , some neocentromeres conferred more accurate chromosome segregation than others ( Fig 3 ) and this correlated with the estimated total transcriptional activity in the region prior to neocentromere formation ( Fig 4 ) . Specifically , higher transcriptional activity correlates positively with higher chromosome loss . This indicates that chromosome segregation is more efficient in regions with lower native transcription levels . Consistent with this idea , in S . cerevisiae , low levels of transcription are compatible with centromere function , high levels of transcription are disruptive [9 , 51] . We posit that kinetochore assembly on DNA competes with transcription initiation and thus , the stronger the affinity of the transcription machinery for the DNA , the weaker , or less functional is the assembled kinetochore on the neocentromere . Neocentromere formation results in transcriptional repression of ORFs within the centromeric chromatin , further supporting the model that kinetochore assembly competes with transcription initiation . Transcriptional repression is evident at proximal neocentromeres , where the presence of URA3 facilitates the detection of reversible silencing [32] , but is also clear from analysis of transcript levels by qRT-PCR ( Fig 2 ) . In Shang et al . , repression of a gene located within the CENP-A boundaries of a single neocentromere was found to be repressed following neocentromere formation [34] . Our data showing repression of nine different genes at five C . albicans neocentromeres ( Fig 2 and S6 Fig ) supports the idea that transcriptional repression is a conserved feature of neocentromeres . Because the inhibition of transcription was limited to the CENP-A bound region and did not extend to neighboring genes ( Fig 2 and S6 Fig ) , we propose that CENP-A recruitment to neocentromeres , and the resulting chromatin structure and kinetochore complex assembly , hinders the transit of transcription complexes through the region . Most neocentromere loci were initiated by deleting one of the two CEN5 copies , such that in the original isolates only one allele would be associated with CENP-A and repressed , while the other would be expressed . Interestingly , four ORFs within neocentromere CENP-A binding regions have homologs in S . cerevisiae that are essential for growth under normal laboratory conditions: orf19 . 3166 , orf19 . 3161 , orf19 . 4221 , and orf19 . 4230 . For three of these neocentromere loci , we were unable to isolate homozygous centromere deletion strains with CENP-A assembled on these genes suggesting that neocentromere formation in regions with putative essential genes can only occur on one allele . The fourth putative essential gene ( orf19 . 3161 ) is found within the proximal neocentromere position that is positionally unstable as seen by reversible silencing of URA3 , perhaps to allow access of RNA polymerase within the region [32] . We suggest that it may not be possible to form functional neocentromeres on both copies of an essential gene , as it would reduce transcription to levels that would be detrimental to growth and survival . Importantly , the data clearly reveals that distal neocentromeres direct 3D centromere clustering like native centromeres ( Fig 5 , Fig 6 and Fig 7 ) and independent of their distance to the native centromere location . Indeed , not only does the distance of the neocentromere from the deleted native centromere not correlate with neocentromere chromosome segregation function ( Fig 4 ) , it also does not appear to affect centromere clustering . Thus , our study of neocentromeres demonstrates , for the first time , that centromere clustering , which has been observed in many fungi and can be used to identify functional centromeres , is an epigenetic feature of an active centromere and is independent of DNA sequence or chromosomal context . Importantly , the neocentromere centered at 4 . 5kb clearly clustered with other centromeres despite having the second-highest chromosome loss rate , albeit less than 0 . 05% ( Fig 3 ) . Thus , even neocentromeres at the lower end of the chromosome segregation function scale still recapitulate a remarkable number of centromere features , including kinetochore assembly , chromosome segregation , and centromere clustering . The variability in chromosome segregation accuracy of different neocentromeres has implications for our understanding of evolution and cancer . Some instances of speciation involve the formation of evolutionary new centromeres through centromere repositioning events that become fixed in the population [19 , 25] . A significant fraction of randomly isolated neocentromeres likely have the properties necessary to become evolutionary new centromeres . Other neocentromeres have elevated levels of chromosome segregation errors that could produce aneuploid progeny , which frequently are unfit [52] and yet sometimes promote survival under specific stress conditions , particularly in mitotic or somatic cells [53 , 54 , 55] . Approximately 3% of cancer cells have neochromosomes , many of which must have assembled neocentromeres [28] . Decreased chromosome segregation accuracy in neocentromere-containing cancer cells may promote the development of chemotherapy resistance . For example , aneuploidy gives rise to gene copy number variations that can confer resistance to chemotreatment in ovarian cancer [56] . On the other hand , very high levels of chromosome loss may decrease the survival of the cancer cells as extreme genome instability is associated with better prognosis in breast cancer patients [57] . Thus , in cancer cells , neocentromeres with lower levels of chromosome segregation accuracy might synergize with chemotherapy treatments to promote very high levels of genome instability and in turn , improve patient prognosis . It is not yet clear what mechanisms determine the relative chromosome loss rates at different neocentromere positions in any organism or cell type . Native centromeres recruit error correction proteins [37 , 58] , spindle assembly checkpoint proteins [59] , and structural complexes such as condensin and monopolin [58 , 60] that are required for optimal chromosome segregation efficiency . We propose that the differences in chromosome segregation accuracy at different active neocentromeres may be due to the differential ability of neocentromeres to recruit these proteins or complexes . Whether this is due to the underlying DNA sequence and its transcriptional state , or to the epigenetic recruitment of other factors remains to be explored .
CEN5 was deleted as previously described [32] . Lithium acetate transformation of PCR products with at least 70 bp of homology to the targeted gene was used for strain construction . Briefly , strains to be transformed were inoculated in liquid YPAD ( 10g/L yeast extract , 20g/L bactopeptone , 0 . 04g/L adenine , 0 . 08g/L uridine , 20g/L dextrose ) and grown at 30°C for 16–18 h . Cultures were then diluted 1:166 in YPAD and grown at 30°C for 3–4 h . Cells were washed with water , then TELiAc ( 10mM Tris pH 7 . 5 , 1mM EDTA , 100mM LiAc ) and incubated in TELiAc with transformation DNA and 50μg sheared salmon sperm DNA ( Ambion ) for 30 min . 4 volumes PLATE mix ( 40% PEG , 10mM Tris pH 7 . 5 , 1mM EDTA , 100mM LiAc ) was then added and the transformation mix was incubated for 16–18 h at 20–24°C . Transformations were heat shocked at 42°C for 1 h , then plated on selective media with the exception of NAT1 marker transformations , which were recovered on non-selective media for 6 h prior to replica plating to selective media containing 400 μg/ml nourseothricin ( Werner BioAgents ) . Strains were checked by PCR of genomic DNA . ChIP was performed essentially as described in [32] . ChIP was performed using rabbit anti-Cse4 ( CaCENP-A ) antibodies [32] , rabbit anti-histone H4 antibodies [61] , and rabbit anti-histone H4K16Ac antibodies ( Abcam ) . DNA pull-down efficiency was measured by qPCR using the Universal Probe Library ( Roche Applied Science ) with a LightCycler 480 PCR machine ( Roche Applied Science ) according to the manufacturer’s instructions . Enrichment was calculated as relative quantification of ( +Ab/Input ) - ( -Ab/Input ) using the second-derivative maximum to determine CT values and corrections for primer efficiency values with the LightCycler 480 software ( Roche Applied Science ) . For H4K16 ChIP , the H4K16 ChIP was normalized to ChIP of total H4 . Custom microarrays ( Agilent SurePrint 8x60k ) were designed with 60bp probes targeted towards all centromere sequences and the complete chromosome sequences of Chr4 , Chr5 and Chr7 . Labeling of ChIP DNA ( input and anti-Cse4 IP ) and hybridization of the arrays were performed according to the manufacturer’s instructions . Arrays were scanned with an Agilent SureScan scanner . Images were processed with the Agilent Feature Extraction software . The Log2 IP/WCE ratio data was normalized and plotted by chromosome position . Neocentromere positions were identified by areas of enrichment of Cse4 and were confirmed by qPCR . Strains were inoculated into YPAD and grown at 30°C for 16–18 hr . Cultures were then diluted 1:100 into YPAD and grown at 30°C for 4 hr . RNA was prepared using the MasterPure yeast RNA purification kit ( Epicentre ) according to the manufacturer’s instructions . RNA was treated with DNase ( Epicentre ) to remove contaminating genomic DNA . cDNA was prepared using the ProtoScript M-MuLV First Strand cDNA Synthesis Kit ( New England Biolabs ) according to the manufacturer’s instructions with oligo dT primers . cDNA was measured by qPCR using the Universal Probe Library ( Roche Applied Science ) with a LightCycler 480 PCR machine ( Roche Applied Science ) or Rotor-Gene SYBR Green master mix ( Qiagen ) with a Rotor-Gene cycler ( Qiagen ) according to the manufacturer’s instructions . Expression was calculated as the amount of cDNA from the gene of interest relative to the amount of TEF1 cDNA in the same sample using the second-derivative maximum to determine CT values and corrections for primer efficiency values . Fluctuation analysis of loss rates was performed as described elsewhere [62] using the method of the median [63] . Briefly , strains were streaked for single colonies and grown on SDC-Uri for 2 days at 30°C . Per strain , 8 independent colonies were inoculated into 1ml liquid non-selective medium ( YPAD ) and grown overnight at 30°C with shaking . For heat stress assays , cultures were incubated at 39°C . For nicotinamide assays , colonies were inoculated in YPAD + 2mM nicotinamide and incubated at 30°C . For nocodazole assays , cells were inoculated in YPAD + 100μM nocodazole . Cultures were harvested by centrifugation and washed once in 1ml of sterile water . Dilutions were plated onto nonselective YPAD for total cell counts and selective media ( SD+FOA for URA3 loss ) ( Gold Biotechnology ) . Plates were incubated at 30°C for 2–3 days , and colony counts were used to calculate the rate of FOAR/cell division [62] . At least 8 individual colonies that lost URA3 were isolated from 5-FOA plates after incubation in the fluctuation analysis . Colonies were streaked on YPAD plates and incubated at 30°C for 24 hr . Following genomic DNA extraction , PCR was performed on the right ( 5R ) and left ( 5L ) ends of Chr5 using primers as previously described [44] . Restriction digests were performed on resulting PCR products with Alu1 ( 5R ) at 37°C and Taq1 ( 5L ) at 60°C for approximately 16 hr . Restriction digests were run on 3% agarose gels to check for SNP homozygosity or heterozygosity ( digested or non-digested alleles based on SNP present within PCR ) [44] . Hi-C experiments were performed as described previously using the Sau3AI restriction enzyme to digest the chromatin [64] . Sequencing was performed using 80bp paired-end reads . Reads were trimmed by 10bp from each end and remaining 60bps were mapped to C . albicans reference genome using BWA [65] with no mismatches allowed . Uniquely mapped read pairs were further binned into non-overlapping 10kb windows to create raw contact maps which were subsequently normalized using an iterative correction method [66] . The resulting normalized contact maps were used for heatmaps , virtual 4C plots and for prediction of centromere coordinates using Centurion algorithm [49] . Sequencing data for Hi-C libraries are available from the Short Read Archive accession number PRJNA308106 . C . albicans genome information was obtained from the Candida Genome Database at www . candidagenome . org . Assembly 21 was used for mapping neocentromere coordinates . Inverted repeats were identified with the Inverted Repeats Database ( https://tandem . bu . edu/cgi-bin/irdb/irdb . exe ) , and tandem repeats were identified with the Tandem Repeats Database ( https://tandem . bu . edu/cgi-bin/trdb/trdb . exe ) . GC% and GC skew ( ( G-C ) / ( G+C ) ) was calculated with FastPCR . Gene essentiality information for homologs of C . albicans genes was obtained from the Saccharomyces Genome Database at www . yeastgenome . org .
|
The accurate segregation of chromosomes during cell division is essential for maintaining genome integrity . The centromere is the DNA region on each chromosome where assembly of a large protein complex , the kinetochore , is required to maintain proper chromosome segregation . In addition , active centromeres exhibit a specific three-dimensional organization within the nucleus: the centromeres associate with one another in a clustered manner . Neocentromeres , or new centromeres , appear at new places along the chromosome when a native centromere becomes non-functional . We used a yeast model , Candida albicans , and isolated twenty instances in which neocentromeres had formed at different positions . All of these neocentromeres were able to direct chromosome segregation , but some had increased error rates . Like native centromeres , these neocentromeres cluster in the nucleus with the other active centromeres . This implies that formation of a neocentromere leads to reorganization of the three-dimensional structure of the nucleus so that different regions of the chromosome are in closer contact to regions of other chromosomes . Recent work suggests that approximately 3% of cancers may contain chromosomes with neocentromeres . Our observations that many neocentromeres have increased error rates provides insight into genome instability in cancer cells . Changes in chromosome copy number may benefit the cancer cells by increasing numbers of oncogenes and/or drug resistance genes , but may also sensitize the cells to chemotherapy approaches that target chromosome segregation mechanisms .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"chromosome",
"structure",
"and",
"function",
"centromeres",
"pathogens",
"microbiology",
"dna",
"transcription",
"fungi",
"model",
"organisms",
"genome",
"analysis",
"fungal",
"pathogens",
"research",
"and",
"analysis",
"methods",
"mycology",
"genomics",
"chromosome",
"biology",
"chromosomal",
"disorders",
"medical",
"microbiology",
"gene",
"expression",
"microbial",
"pathogens",
"biological",
"databases",
"genetic",
"loci",
"yeast",
"clinical",
"genetics",
"candida",
"chromosomes",
"cell",
"biology",
"database",
"and",
"informatics",
"methods",
"genetics",
"biology",
"and",
"life",
"sciences",
"yeast",
"and",
"fungal",
"models",
"computational",
"biology",
"genomic",
"databases",
"organisms",
"candida",
"albicans"
] |
2016
|
Neocentromeres Provide Chromosome Segregation Accuracy and Centromere Clustering to Multiple Loci along a Candida albicans Chromosome
|
Advances in high-throughput genotyping and the International HapMap Project have enabled association studies at the whole-genome level . We have constructed whole-genome genotyping panels of over 550 , 000 ( HumanHap550 ) and 650 , 000 ( HumanHap650Y ) SNP loci by choosing tag SNPs from all populations genotyped by the International HapMap Project . These panels also contain additional SNP content in regions that have historically been overrepresented in diseases , such as nonsynonymous sites , the MHC region , copy number variant regions and mitochondrial DNA . We estimate that the tag SNP loci in these panels cover the majority of all common variation in the genome as measured by coverage of both all common HapMap SNPs and an independent set of SNPs derived from complete resequencing of genes obtained from SeattleSNPs . We also estimate that , given a sample size of 1 , 000 cases and 1 , 000 controls , these panels have the power to detect single disease loci of moderate risk ( λ ∼ 1 . 8–2 . 0 ) . Relative risks as low as λ ∼ 1 . 1–1 . 3 can be detected using 10 , 000 cases and 10 , 000 controls depending on the sample population and disease model . If multiple loci are involved , the power increases significantly to detect at least one locus such that relative risks 20%–35% lower can be detected with 80% power if between two and four independent loci are involved . Although our SNP selection was based on HapMap data , which is a subset of all common SNPs , these panels effectively capture the majority of all common variation and provide high power to detect risk alleles that are not represented in the HapMap data .
Researchers have used several approaches to identify genetic variation that predispose individuals to common diseases . For Mendelian disease traits , the method of choice has been linkage analysis in families [1] . For many complex disease traits where many alleles contribute to the trait , the effect of any single gene or locus may be relatively small and difficult to detect by linkage analysis . A more sensitive approach to identifying risk alleles with smaller gene effects is to employ a case-control association study in which allelic markers , such as SNPs , are used to find regions of the genome enriched ( or depleted ) in a particular risk allele or haplotype between the cases and controls . This study design can provide more power to detect relatively small gene effects [1] . Historically , association studies have employed markers in candidate genes; unfortunately this approach requires a priori knowledge about which genes to choose . A more comprehensive and agnostic approach is to employ markers encompassing the entire genome . Whole-genome association studies survey common genetic variation by probing a dense set of SNPs across the genome . Because whole-genome studies allow researchers to identify genes not previously known to be involved with disease etiology and may be more sensitive for detecting the multiple small gene effects often found in complex disease traits , they may be the most appropriate method to identify variants that predispose individuals to common diseases . Whole-genome association studies require a dense map of hundreds of thousands of SNPs across the genome and sufficiently large sample sizes to provide power to detect relatively small gene effects [2 , 3] , but knowledge of the underlying LD ( LD ) structure can be used to minimize the number of genotyped SNPs [4 , 5] . The International HapMap Project [6] has provided a densely mapped and validated set of SNP loci genotyped in four populations that can be used to select these SNPs . The complete HapMap dataset ( release 21 ) contains over 2 million common SNPs ( minor allele frequency [MAF] ≥0 . 05 ) in each population studied ( Utah residents with ancestry from Northern and Western Europe [CEU] , Han Chinese/Japanese in Tokyo , Japan [CHB + JPT] , and Yoruba in Ibadan , Nigeria [YRI] ) . One of the many results shown in the HapMap data is that LD is often discontinuous , appearing as “block-like” structures [7 , 8] , and a typical SNP is highly correlated with many of its neighbors in any given population . This correlation can be used to build highly efficient whole-genome genotyping panels by choosing tag SNPs that serve as proxies for many other highly correlated neighboring SNPs [9] . A tag SNP approach drastically reduces the number of genotyped loci while producing more information content than a large set of randomly-chosen loci [8 , 10] . We have recently developed a whole-genome genotyping assay [11 , 12] that has the capability of genotyping hundreds of thousands of markers , enabling whole-genome association studies on a single microarray . Using this assay , we have constructed whole-genome genotyping panels of over 550 , 000 ( HumanHap550 ) and 650 , 000 ( HumanHap650Y ) SNP loci by choosing tag SNPs from all populations in the International HapMap Project . We estimate that these tag SNP loci cover the majority of all common variation in the genome and show that most association studies using these panels should detect genetic risk factors even for complex traits where each risk allele only confers a moderate risk .
We constructed the HumanHap550 panel , a whole-genome genotyping panel of 555 , 352 SNPs , to effectively tag CEU ( European ) and CHB + JPT ( Asian ) sample populations ( Table 1 ) . A majority of the SNPs were selected by tagging the more than 2 million common HapMap SNPs , but the panel also includes variation types that have been found to be overrepresented in diseases such as nonsynonymous SNPs [1] , SNPs in the MHC region [13] , SNPs in commonly reported CNV regions [14] , and mitochondrial SNPs [15] . Because individuals with African ancestry have distinct and lower levels of LD compared to those with European or Asian ancestry [16] , we added another 100 , 000 common YRI ( African ) tag SNPs to increase coverage of the YRI samples for the HumanHap650Y panel . In this publication , genome coverage and power are calculated using the HumanHap550 panel for CEU and CHB + JPT samples and the HumanHap650Y panel for YRI samples . Coverage calculation has been previously described for a subset of ∼314 , 000 SNPs from these panels ( the HumanHap300 panel ) [17 , 18] . We assessed how well the whole-genome genotyping panels capture all variation in the human genome . To do this , we measured how well all variation is captured by at least one SNP on the array at various levels of LD as measured by r2 [19 , 20] . Since all variation in the genome is currently not known , one can use proxies for all variation to estimate genomic coverage . These include SNPs genotyped in the International HapMap Project [6] , the HapMap ENCODE resequencing and genotyping project [21] , and other complete resequencing data such as the SeattleSNPs Program for Genomic Applications ( PGA ) data ( http://pga . gs . washington . edu/ ) . Genomic coverage was estimated from HapMap release 21 and SeattleSNPs PGA genotype data . Because the PGA data contains 68 genes whose variants were publicly released after the HapMap SNPs had been selected , and HumanHap550 and HumanHap650Y tag SNPs were derived from the HapMap , this dataset provides a relatively unbiased estimate of genomic coverage beyond HapMap . Since most of the SNPs genotyped in the HapMap ENCODE regions were part of the HapMap release 16c data , estimates of genomic coverage in ENCODE regions is not an independent assessment of genomic coverage . Coverage of ENCODE regions was very similar to coverage of the entire HapMap data and is not shown . Coverage of common variation ( MAF ≥ 0 . 05 ) was estimated from the HapMap data ( release 21 ) for each population ( CEU , CHB + JPT , YRI ) . To calculate coverage , the maximum pairwise r2 value [19 , 20] between each common HapMap release 21 SNP and a SNP in either HumanHap550 or HumanHap650Y was determined . The cumulative proportion of HapMap release 21 loci captured by HumanHap550 and HumanHap650Y is shown in Figure 1 . The mean maximum r2 was 0 . 93 and 0 . 91 between HumanHap550 SNPs and common HapMap SNPs in the CEU and CHB + JPT samples , respectively , and 0 . 81 for HumanHap650Y SNPs in the YRI samples . Using a strict r2 threshold of 0 . 8 , HumanHap550 captures 90% and 87% of the common HapMap SNPs in the CEU and CHB + JPT samples , respectively , and the HumanHap650Y captures 67% of common HapMap SNPs in YRI samples . Using a less-stringent r2 threshold ( 0 . 5 ) , the HumanHap550 panel captures 96% and 95% of the common HapMap variants in the CEU and CHB + JPT samples , respectively , and the HumanHap650Y captures 85% of common variants in YRI samples . Because both SNP selection and genomic coverage of the HumanHap550 panel was based on the same HapMap data , these estimates are likely to be an overestimate of true genomic coverage . Although the International HapMap Project genotyped over 5 . 8 million SNPs in four populations , these SNPs do not necessarily represent all genomic variation due to biases in SNP discovery and genotyping . Furthermore , as mentioned previously , there is a slight upward bias for our estimate of genomic coverage because the SNP selection used for these panels was based on the samples and markers in the HapMap project . An unbiased estimate of the coverage of these panels can be obtained from regions that are fully sequenced independently from the HapMap project . Regions sequenced before the HapMap project may be overrepresented in the HapMap data because more complete knowledge existed for these regions when SNPs were selected for the HapMap project . The SeattleSNPs Program for Genomic Applications ( PGA; http://pga . gs . washington . edu/ ) has resequenced 68 genes ( as of 29 September 2006 ) related to inflammatory response in 23 and 24 individuals in common with the HapMap CEU and YRI samples , respectively . The PGA's earliest SNP entries into dbSNP are build 125 for these 68 genes , after when SNPs from the HapMap Project were selected ( http://www . hapmap . org/ ) [6] . The HumanHap550 panel has a mean maximum r2 of 0 . 84 and captures 72% of common variation in the CEU population ( r2 = 0 . 8 ) , and HumanHap650Y has a mean maxiumum r2 of 0 . 68 and captures 45% of the common variation in the YRI population ( r2 = 0 . 8 ) , respectively ( Figure 1B ) . These values are 18% and 22% less than what is seen in the HapMap data for the CEU and YRI samples , respectively . It is possible that the 68 genes in the PGA dataset may not accurately reflect all variation in the genome since these are candidate genes for inflammatory response and may be undergoing selection or have atypical patterns of natural variation . To test if these genes are undergoing selection , we calculated the average Tajima's D statistic for the 68 genes and found that they are close to neutrality and not significantly different from all genes for which data is available ( Table S1 ) . Quantifying coverage according to r2 does not fully address the probability that a study will detect risk alleles in genotype–phenotype association studies . Because the sample size required to maintain the same power is inversely proportional to r2 [3 , 22 , 23] , LD coverage allows us to estimate the reduction in power that will occur by not genotyping every SNP directly . Other factors are more important than r2 for the power of an association study ( i . e . , the probability that a risk allele will be observed at significantly higher frequencies in the cases versus controls ) , such as the risk conferred by the allele and the frequency of the disease . In addition , the frequency of the risk allele is a very important factor in how much power an association study will have to detect the risk allele because low frequency risk alleles are much more difficult to detect than high frequency risk alleles . As long as the genotyped SNPs have some correlation with the risk alleles , increasing the sample size of a study can overcome all these factors so that a study can be sufficiently powered with a large enough sample size . Selecting SNPs in a way that maximizes the coverage of all SNPs through LD allows us to maximize our power for a given sample size . For a model where there exists a single risk allele , we estimated the total power to detect a single risk allele represented by the HapMap data where each SNP is equally likely to be the risk allele . For each SNP in the HapMap data , we combined the frequency and maximum r2 with a nearby SNP on the HumanHap550 or HumanHap650Y to estimate the power to detect that SNP if it is a risk allele for various sample sizes and disease models . For each risk allele frequency , sample size , and disease model , the power was estimated from 10 , 000 simulated case-control datasets ( see Methods ) . Then , under the assumption that each SNP in the HapMap dataset is equally likely to be the risk allele , we calculated the total power , PT , to detect a risk allele represented in the HapMap data for a 5% false-positive rate after applying a Bonferroni correction to account for multiple testing ( see Methods ) . Under the multiplicative disease model , an individual's baseline risk of having the disease or phenotype is multiplied by λ if they carry one copy of the risk allele and λ2 if they carry two copies of the risk allele . Figure 2 ( red line ) shows the total power to detect an association using HumanHap550 ( CEU and CHB + JPT ) and HumanHap650Y ( YRI ) for a study design with 1 , 000 cases and 1 , 000 controls for different relative risks , λ . For the CEU and CHB + JPT samples , the power drops below 80% when the relative risk is less than ∼1 . 69 under the multiplicative model , whereas for the YRI samples the power drops below 80% when the relative risk is less than ∼1 . 85 ( Figure 2 ) . We have also performed the power analysis for an additive model where an individual's baseline risk is multiplied by λ if they carry one copy of the risk allele and 2λ if they carry two copies of the risk allele ( see Figure S1 ) . Under the additive model the power drops below 80% when the relative risk is less than ∼1 . 61 in the CEU and CHB + JPT samples and ∼1 . 80 in the YRI samples ( Figure S2 ) . These simulations were based on sample sizes of 1 , 000 cases and 1 , 000 controls but a lower relative risk can be detected by increasing the study size . To illustrate this , we also estimated power in a variety of study sample sizes to examine the minimum relative risk detectable at 80% power under the multiplicative and additive models for larger sample sizes ( Figure 3 [CEU] and Figure S2 [CHB + JPT and YRI] ) . Under the multiplicative model , disease variants with relative risks as low as ∼1 . 20 for the CEU and CHB + JPT samples and ∼1 . 24 for the YRI samples can be detected with 80% power with 10 , 000 cases and 10 , 000 controls ( Figure S2; Tables S2 and S4 ) . Under the additive model , diseases variants with relative risks as low as ∼1 . 06 for the CEU and CHB + JPT samples and ∼1 . 12 for the YRI samples can be detected with 80% power with 10 , 000 cases and 10 , 000 controls ( Figure S2; Tables S3 and S5 ) . These power estimates are strongly influenced by the frequency spectra of the SNPs in the HapMap data . While the power is relatively low to detect a single risk allele if the minor allele frequency is less than 0 . 10 , the power is significantly greater for risk alleles with minor allele frequencies greater than 0 . 40 ( dashed lines in Figure 2 ) . For example , under the multiplicative model , relative risks as low as 1 . 50–1 . 57 can be detected in 1 , 000 cases and 1 , 000 controls if the minor allele frequency is greater than 0 . 40 , but when the minor allele frequency is less than 0 . 10 , the risk allele can only be detected at this sample size if the relative risk is greater than 2 . Unknown factors such as the risk allele frequency may significantly increase the power to detect a particular disease even though the above estimates indicate that a study is underpowered . Conversely , even a supposedly well-powered study may have little ability to detect associations if the risk allele frequencies are low , unless they confer a very large risk . Since there are ∼7 . 1 million common ( MAF ≥ 0 . 05 ) SNPs in the genome [24] and 2 . 0 million , 2 . 2 million , and 2 . 5 million common SNPs have been genotyped by the HapMap Project in the CHB + JPT , CEU , and YRI populations , respectively , approximately 70% of the common risk alleles are not contained in the HapMap data . For SNPs that are not contained in the HapMap data , the power will be lower than shown in Figures 2 and 3 because the tag SNPs selected for the HumanHap550 and HumanHap650Y were chosen specifically to maximize coverage of the HapMap SNPs . To explore how the HumanHap550 and HumanHap650Y will perform for the non-HapMap SNPs , we estimated the coverage of non-HapMap SNPs using the SNPs found in 68 genes that were completely resequenced by SeattleSNPs after the release of the HapMap data . It is important to note that for this analysis we excluded all SNPs that are in the HapMap dataset to provide an unbiased estimate of the power for SNPs outside of HapMap . Following the same procedure as outlined above , we estimated the power to detect the common SNPs ( MAF ≥ 0 . 05 ) that are not in the HapMap data ( Figure 4 multiplicative model; Figure S3 additive model ) . In 1 , 000 cases and 1 , 000 controls , the minimum relative risks that we can detect at 80% power is ∼14%–22% higher for non-HapMap SNPs compared with our estimates for just the HapMap data ( Tables S2–S5 ) . In 10 , 000 cases and 10 , 000 controls , the minimum relative risks that we can detect at 80% power is ∼6%–11% higher for non-HapMap SNPs compared with our estimates for just the HapMap data ( Tables S2–S5 ) . The SeattleSNPs dataset represents a small number of SNPs ( 668 for non-HapMap SNPs for CEU and 1 , 429 for YRI ) , so even though these genes represent over ∼2 . 4 Mb of the human genome , these are rough estimates and more extensive data will help refine these numbers . Overall , these estimates of power are very promising for studies that wish to detect risk alleles both within and outside of the HapMap SNPs . The total power to detect a common risk allele in an association study is a weighted average of the power to detect a HapMap risk allele and the power to detect a non-HapMap risk allele where the weights are the relative fraction of SNPs within and outside of the HapMap data ( see Methods ) . We have combined our power calculations to estimate the genome-wide power for a variety of disease models and study designs ( Tables S2–S5 ) . In 1 , 000 cases and 1 , 000 controls , the minimum relative risks that we can detect at 80% power is ∼10%–18% higher genome wide compared with our estimates for just the HapMap data . In 10 , 000 cases and 10 , 000 controls the minimum relative risks that we can detect at 80% power is ∼4%–8% higher genome-wide compared with our estimates for just the HapMap data ( Tables S2–S5 ) . The power calculations shown above were done assuming that a single , common risk allele is represented in the genome . For most complex traits with a significant genetic component , multiple risk alleles and loci are likely to be involved [25]; therefore , the power estimates calculated under the assumption that there is only one risk allele may be overly conservative . Risk alleles have been detected in multiple genes for many diseases [26–33] . While the most desirable case power-wise is to have a single high-risk allele that explains the entire genetic component of a disease , the presence of multiple risk loci will improve the power for association studies—compared with a single loci of the same frequency and risk—if the loci are segregating independently and noninteractively in the population . For example , if there are two risk alleles segregating independently , and one confers a multiplicative relative risk of λ = 1 . 5 and the other λ = 1 . 6 , the genome-wide power to detect each SNP in 1 , 000 cases and 1 , 000 controls is 0 . 42 and 0 . 59 , respectively ( Table S2 ) . However , the power to detect at least one of the two risk loci is one minus the probability that we do not detect either locus ( 0 . 76 ) , greatly improving the power to detect any particular single locus . To further illustrate how the power improves when multiple loci are involved , we started from the whole-genome estimates and examined the case where a phenotype has genetic risk factors at two or four unlinked loci ( see Methods ) . If two unlinked loci are involved , studies with the same sample sizes and power may be able to detect smaller gene effects ( Figure 5 ) . As shown in Tables 2 and 3 , at 80% power , the minimum relative risk to detect at least one of the two risk alleles can be as much as 17% smaller for the multiplicative model and 23% for the additive model . If four unlinked loci are involved , even smaller gene effects may be detected at 80% power given the same sample size . The minimum relative risk detectable with 80% power can be as much as 23% smaller for the multiplicative model and 36% for the additive model when four unlinked loci are involved ( Tables 2 and 3 ) . These values require that each risk allele confers at least the same relative risk or greater and thus can only be considered rough estimates . Even so , the presence of multiple unlinked risk alleles will greatly improve the power of association studies .
Just as high-throughput sequencing and the human genome project have greatly advanced the study of genomics [34] , so have advances in high-throughput genotyping and the HapMap project enabled association studies at a whole-genome level . Using data from the International HapMap Project [6] , we have constructed genotyping panels for whole-genome association studies . These panels are invaluable tools for discovering etiologic variants of complex diseases . We have shown that , given a sample size of 1 , 000 cases and 1 , 000 controls , these panels have the power to detect single disease loci of moderate risk ( λ ∼1 . 8–2 . 0 ) . Relative risks as low as λ ∼1 . 2–1 . 3 can be detected using 10 , 000 cases and 10 , 000 controls , depending on the sample population and disease model . If multiple loci are involved , the power increases significantly such that relative risks 20%–35% lower can be detected with 80% power if between two and four independent loci are involved . Although our SNP selection was based on HapMap data , which is a subset of all common SNPs , these panels effectively capture the majority of all common variation and provide high power to detect risk alleles that are not represented in the HapMap . Additionally , we have focused on common ( MAF ≥ 0 . 05 ) variants in this study , and it should be stressed that because there are many more rare variants in the population , there will likely be many more rare risk alleles . The power for these rare risk alleles will be much lower than the average power that we have used ( e . g . , dashed lines in Figure 2 ) , and discovering these low-frequency risk alleles will likely require resequencing efforts . Using the known relationship between sample size and r2 [3 , 22] , we estimated the probability of having a positive outcome under various scenarios of study size and different disease models ( i . e . , risks ) . Recently , this relationship between r2 and sample size has been questioned [23] . It is true that if tag SNPs are selected using small sample sizes , the LD between SNPs will be upwardly biased compared with larger datasets but this bias will only be large if the training sample is small and/or the LD criteria used to select the tag SNPs is low . When tag SNPs are selected using high-LD criteria ( e . g . , r2 > 0 . 5 ) and sample sizes are greater than ∼50 samples , this bias is minimal ( e . g . , Figure 1 of [23] ) . Because our tag SNPs are selected using a stringent r2 criteria ( r2 > 0 . 8 ) and more than 50 individuals were genotyped in each of the CEU , CHB + JPT , and YRI populations , the bias in LD will be small for these tag SNPs [35] . One method to increase power in a whole-genome association study is to use imputation methods to infer genotypes for unobserved markers [30 , 36 , 37] . This method relies on phased genotype data from a known set of genotyped markers , i . e . , the HapMap . The reliability of the imputed genotyped data is dependent on how well correlated the observed ( genotyped ) SNPs are to the unobserved ( imputed ) genotyped SNPs . Therefore , whole-genome genotyping panels selected using tag SNPs would likely provide more reliably imputed genotypes as the same number of non-tag SNPs . Reliability of the inferred genotypes is also dependent on the study population being similar to the HapMap population being used to impute genotypes . The use of multi-marker haplotypes would also increase power in a whole-genome association study by providing greater specificity for detecting the actual risk allele [38–41] . In the extreme example where allelic heterogeneity exists and the disease variant occurs on two different haplotypes , where each haplotype does not contain any alleles in common ( i . e . , AB versus ab ) , the association would be missed using single-marker analysis , but could be detected if a haplotype analysis was used . The power estimated in this paper used single-marker analysis only , and therefore power could be improved beyond what is presented here if multi-marker haplotypes are used . Since we made the assumption that risk alleles exist as either a single noninteractive SNP or a couple of unlinked loci , it is important to understand how our results would be affected if some of the risk alleles are in LD with one another . If multiple risk alleles are in LD with one another then they could mask the signal of each other and be much more difficult to detect in simple association studies . An extreme example of how one risk allele may mask another would occur is if two risk alleles ( both of which confer the same risk ) were in perfect LD such that every chromosome contained one of the risk alleles but never contained both ( i . e . , D′ = −1 ) . In this example , the genetic risk factors would be impossible to detect using simple association studies . While this scenario may occur , it seems unlikely that this would be any more prevalent than a scenario where the risk alleles are correlated in such a way that they increase the probability of detection; e . g . , both risk alleles always occur on the same chromosome ( i . e . , D′ = 1 ) . Thus , our results probably represent the average scenario and some association studies will have more power and some will have less power , due to LD between risk alleles . A more difficult example would occur when there are epistatic interactions between loci that mask the overall gene effect signal that we would expect to get from each marker independently . Searching for these types of interactions will require at the very least multi-marker analyses . The power calculations presented here do not cover epistatic disease models and our results would substantially overestimate the power under these disease models . Complex diseases as a group vary across of range of genetic risks , environmental factors , and other complexities . At one extreme are the simpler diseases that contain relatively strong genetic components , such as Crohn Disease ( sibling relative risk , λs = 17–35 [42 , 43] ) and Type I Diabetes ( λs = 15 [44] ) . Isolating the high-risk genetic components for diseases similar to these may not require as many markers or samples because even SNPs that are loosely correlated with markers of these risks will be expected to show association . Conversely , diseases towards the other end of the spectrum of complexity , such as Type 2 Diabetes Mellitus ( λs = 3 [45] ) , where many risk alleles , each with relatively small gene effects , are expected to occur , will require denser marker sets and larger samples sizes to detect an association [46] . To detect risk alleles , a key component is the design of genome-wide association studies . For example , in choosing patient samples , case subjects with specific disease subphenotypes are collected to decrease phenotypic heterogeneity and appropriate controls are used to eliminate false positives from population stratification [31 , 47] . A second key component to genome-wide association studies is the marker set . To address this , we have designed the HumanHap550 and HumanHap650Y whole-genome genotyping panels to maximize the likelihood of detecting etiologic variants . By selecting tag SNPs , we reduced the amount of required genotyping of more than 2 million common HapMap SNPs to ∼550 , 000–650 , 000 while retaining most of the power ( e . g . , Figure 2 ) . Furthermore , to increase the chance of detecting etiologic variants , we have selected SNPs in gene regions and regions that have historically been overrepresented in disease , such as copy-number variation regions and mitochondrial DNA . Already , the success of genome-wide association studies are coming to bear and candidate risk alleles have been discovered for several disorders [26–33 , 47–56] . In summary , whole-genome tag SNP panels , such as the HumanHap550 and HumanHap650Y panels , should greatly aid in our understanding of how genetic variation affects both human health and disease .
HapMap release 16c , 20 , and 21 data were downloaded from http://www . hapmap . org/ . The filtered , nonredundant genotypes were used for all analyses . SeattleSNPs data were downloaded on 29 September 2006 from http://pga . gs . washington . edu/ . SeattleSNPs sequenced 23 CEU and 24 YRI unrelated individuals from the HapMap samples . Tajima's D values for the SeattleSNPs genes were taken directly from the SeattleSNPs website . The 18 , 411 RefSeq genes and their coordinates were downloaded from the University of California Santa Cruz ( UCSC ) genome browser ( hg17; http://genome . ucsc . edu/ ) . Genomic regions demonstrating conservation across species and the respective PHAST scores were downloaded from the UCSC genome browser ( phastConsElements . txt . gz ) . Copy number variant ( CNV ) regions of the genome were obtained from the Database of Genomic Variants ( http://projects . tcag . ca/variation/ ) , consisting of 2 , 714 CNV loci ( February 2007 ) . All SNP coordinates are shown with respect to build 35 and dbSNP124 . To validate the whole-genome genotyping panels , accuracy and completeness of the genotypes generated on both the HumanHap550 and HumanHap650Y panels were measured with respect to call rates , Mendelian inconsistencies , reproducibility , and concordance to HapMap genotype data . These data quality parameters are shown in Table 4 . Tag SNPs were by chosen from HapMap data using an algorithm incorporating the LD statistic r2 [19] . The genome was divided into 1-Mb nonoverlapping segments , and pairwise r2 values were calculated for loci within 200 kb . Approximately 314 , 000 SNP loci were first selected from the CEU population from the Phase I HapMap data ( release 16c ) . HapMap release 16c had approximately 775 , 000 SNP loci with MAF ≥ 0 . 05 in the CEU population . First , tag SNPs were chosen using a strict r2 threshold of 0 . 8 . If any SNP in a bin of correlated SNPs was within 10 kb of a RefSeq gene or in an evolutionarily conserved region ( ECR ) , the tag SNP was retained as a “must-keep” SNP . A SNP was defined as being in an ECR if the SNP mapped to one of the phastCons elements with a PHAST score ≥ 50 . A second analysis was done using a less-stringent r2 threshold of 0 . 7 , choosing additional tag SNPs genome-wide in addition to the “must-keep” tag SNPs selected from the previous analysis . This strategy provided a higher density of tag SNPs within 10 kb of genes or in ECRs . To construct the HumanHap550 , an additional ∼240 , 000 tag SNPs were selected from the Phase II HapMap data ( release 20 ) and combined with 313 , 505 HumanHap300 loci . Using the HumanHap300 tag SNP list as “must-haves , ” an analysis was conducted using the full release 20 data in the CEU population ( >2 , 100 , 000 SNPs with MAF ≥ 0 . 05 ) , prioritizing tag SNP selection for those loci that were polymorphic in all HapMap populations . Again , SNP selection in the CEU population was done choosing a higher density of tag SNPs within 10 kb of RefSeq genes and in ECRs ( r2 = 0 . 8 in gene regions/ECRs; r2 = 0 . 7 in rest of the genome ) . All tag SNPs were retained with the exception of singleton bins ( those SNPs not tagging any additional SNPs ) not within 10 kb of a gene or in an ECR . An additional tag SNP was selected for those bins with 10 or more loci . After the core set of tag SNPs were determined in the CEU population , additional tag SNPs were included from the Han Chinese/Japanese ( CHB + JPT; all bins >2 SNPs at r2 = 0 . 8 ) and Yoruba populations ( YRI; all bins >4 SNPs at r2 = 0 . 7 ) , respectively . Additional content was added to the panel including 7 , 779 nsSNPs , 177 mitochondrial SNPs ( selected from http://www . broad . harvard . edu/mpg/tagger/mito . html [57] ) , 4 , 284 SNPs in 495 reported copy number regions of the genome [58–60] , and a higher density of tag SNPs in the MHC region . After this final list was selected , any gaps ≥ 100 kb between common SNPs for each population were filled with common SNPs for that particular population . The mean spacing between consecutive common SNPs on autosomal chromosomes is 5 . 5 kb , 6 . 5 kb , and 6 . 3 kb for CEU , CHB + JPT , and YRI , respectively . To construct the HumanHap650Y , 100 , 000 additional YRI-specific tag SNPs were added to the 555 , 532 previously selected SNPs . Using the 555 , 532 tag SNPs list as “must-haves , ” an analysis was conducted using the release 20 data in theYRI population and tag SNPs from the largest bins were selected ( bins >2 SNPs , r2=0 . 7 ) . The mean spacing between consecutive common SNPs on autosomal chromosomes is 5 . 3 kb , 6 . 2 kb , and 5 . 4 kb across the genome in the CEU , CHB + JPT , and YRI populations , respectively . To calculate coverage of HapMap or SeattleSNPs , pairwise r2 values were calculated using the expectation algorithm [20] based on the genotypes from HapMap release 21 and the 68 genes resequenced in the PGA samples . Maximum r2 values were calculated for each SNP list ( HapMap release 20 or 68 PGA genes ) with a SNP on either HumanHap550 or HumanHap650Y . All pairwise combinations were considered within 200 kb . For chrX , only female individuals were used; otherwise , all unrelated individuals were used . To calculate the power to detect a risk allele with a given risk , risk allele frequency and sample size we generated a series of simulated datasets . For a single sample , we assigned the genotype homozygous for the risk allele and heterozygous or homozygous for the nonrisk allele with Hardy-Weinberg probabilities p2 , 2pq , q2 where p is the frequency of the risk allele and q = 1 − p is the frequency of the nonrisk allele . An individual sample was assigned a disease status with probabilities defined by the disease model and assigned as a case or control accordingly . For example , for an associated risk , λ , under a multiplicative model and a baseline risk , π , an individual has the probability being a case of λ2π , λπ and π depending on whether that individual is homozygous for the risk allele , heterozygous , or homozygous for the nonrisk allele , respectively [61] . Using a program written in C , we generated the allele frequencies for 10 , 000 case-control simulations following the above procedure for a wide range of risk allele frequencies , disease models , and sample sizes . We then calculated the χ2 value for the corresponding 2 × 2 table of allele frequencies and estimated the power for each case as the fraction of times that the χ2 value exceeded the p-value 0 . 05 after a Bonferroni correction for multiple testing ( χ2 ≥ 28 . 5687 for HumanHap550 and χ2 ≥ 28 . 8976 for HumanHap650 ) . This amounted to 188 , 600 , 000 simulated case control calculations for each disease model examined ( multiplicative and additive ) . Using the above power estimates , for each genotyped SNP or SNP tagged at r2 = 1 , we assigned it a power value according to its minor allele frequency for each disease model and study design ( i . e . , number of cases and controls ) . This allows us to do a single power estimation for each of the possible minor allele frequencies ( 5%–50% in increments of 1% ) rather than doing an individual calculation for all possible ∼2 . 2 million common SNPs . This method works when the SNP is directly genotyped or perfectly correlated with one of the genotyped SNPs . Alternatively , if the risk allele is not directly genotyped but instead , one or several nearby markers are genotyped , then the power is expected to be reduced if none of these genotyped markers are perfectly correlated with the risk allele . The amount that the power is reduced is equivalent to reducing the study sample size by r2 , where r2 is the maximum LD between the genotyped markers and the risk allele [22] . For example , genotyping a SNP that is in LD with a risk allele at r2 = 0 . 5 in 1 , 000 cases and controls has the same power as directly genotyping the risk allele in 500 cases and controls . In this example , the effective sample size is 500 and is equal to the actual sample size multiplied by r2 . When the power for the effective sample size was not previously calculated , we linearly interpolated between the power values calculated for sample sizes above and below to estimate the power . The probability of detecting a SNP in a case control study is the probability that the SNP is a risk allele multiplied by the power to detect it if it is a risk allele [62] . Assuming that a single risk allele is involved in a disease and each SNP is equally likely to be a risk allele , then the total power , PT , for all SNPs is just the sum of the powers for each SNP and given by the equation: where Pi is the power to detect an association at SNP i and N is the number of SNPs . Writing the probability that a SNP is in the HapMap data as PH , the total power for the entire genome , PA , can be written as: where PH is the probability that the risk allele is in the HapMap data and ( 1 − PH ) is the probability that that risk allele is outside of the HapMap data , is the power to detect one of the HapMap SNPs , and is the power to detect one of the non-HapMap SNPs . It is estimated that PH , common SNPs represented in the HapMap data , represents 30% of all common SNPs [24] , so PH = 0 . 30 . The power values , and , are taken from the simulated estimates according to the disease model and study design ( e . g . , Tables S1–S4 ) . For a given set of N risk alleles of unknown frequency and total powers P0 , P1 , … , PN , the probability of detecting at least one of these risk alleles—if they are independent of each other—is equal to 1 − ( 1 − P0 ) ( 1 − P1 ) , … , ( 1 − PN ) , assuming the risk alleles are independent of one another . While knowing the risks associated with all disease markers is not likely , the lower bound of this estimate can be calculated as 1 − ( 1 − Pi ) N where Pi is the power for the lowest-powered risk allele and N is the number of independent loci . Thus , if the power to detect a single risk allele is 0 . 5 and there are two risk alleles , then the power to detect one of these risk alleles is 0 . 75 , and if there are three risk alleles , then the power to detect at least one risk allele is 0 . 875 .
|
Advances in high-throughput genotyping technology and the International HapMap Project have enabled genetic association studies at the whole-genome level . Our paper describes two genome-wide SNP panels that contain tag SNPs derived from the International HapMap Project . Tag SNPs are proxies for groups of highly correlated SNPs . Information can be captured for the entire group of correlated SNPs by genotyping only one representative SNP , the tag SNP . These whole-genome SNP panels also contain additional content thought to be overrepresented in disease , such as amino acid–changing nonsynonymous SNPs and mitochondrial SNPs . We show that these panels cover the genome with very high efficiency as measured by coverage of all HapMap SNPs and a set of SNPs derived from completely resequenced genes from the Seattle SNPs database . We also show that these panels have high power to detect disease risk alleles for both HapMap and non-HapMap SNPs . In complex disease where multiple risk alleles are believed to be involved , we show that the ability to detect at least one risk allele with the tag SNP panels is also high .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"homo",
"(human)",
"genetics",
"and",
"genomics"
] |
2007
|
Power to Detect Risk Alleles Using Genome-Wide Tag SNP Panels
|
Pregnant women with sporotrichosis should not receive systemic antifungal therapy except in severe cases when amphotericin B is recommended . Thermotherapy is the most reported treatment described in this group of patients . It entails weeks of daily self-application of heat to the lesions , requires that the patient faithfully apply it , and it could cause skin burns . Cryosurgery is a useful therapeutic tool for many cutaneous infectious diseases , safe for pregnant women , but not well evaluated for sporotrichosis treatment in this group . The authors conducted a retrospective study describing epidemiological , clinical , and therapeutic data related to four pregnant patients with sporotrichosis treated with cryosurgery . The authors reviewed the clinical records of four pregnant patients diagnosed with cutaneous sporotrichosis and treated with cryosurgery . The sessions were carried out monthly up to clinical cure . Molecular identification of the Sporothrix species was performed in two cases using T3B PCR fingerprinting assays . All patients were in the second trimester of pregnancy and their age ranged from 18 to 34 years . With regard to clinical presentation , two patients had lymphocutaneous and two had the fixed form . S . brasiliensis was identified in two cases as the causative agent . Cryosurgery was well tolerated and the number of sessions ranged from 1 to 3 . All the patients reached a complete clinical cure . Cryosurgery was a safe , easy to perform and well tolerated method , and therefore it is suggested to be a suitable option for the treatment of cutaneous sporotrichosis in pregnant women .
Sporotrichosis is caused by dimorphic fungi of the genus Sporothrix , found in its filamentous form as saprophytes on decaying and living vegetation , and soil [1] . However , since the late 1990s , sporotrichosis in the state of Rio de Janeiro , Brazil , has become an urban-epidemic phenomenon , being transmitted from naturally infected cats to humans [2] . The most affected population is characterized by having poor socioeconomic backgrounds and low access to health services . In this zoonotic scenario of sporotrichosis transmission , female patients with a median age of 39 years predominate , and most of them acquire the disease through bite or scratches from infected cats [2] . In this context , women in childbearing age are an at-risk population to acquire this mycosis . Sporotrichosis in pregnancy is a therapeutic challenge . Pregnant women should not receive azole therapy due to the potential teratogenic effects , as well as potassium iodide saturated solution ( SSKI ) , because of its toxicity to the fetal thyroid . Although terbinafine is classified by the US Food and Drug Administration ( FDA ) as a category B drug , there is no sufficient clinical experience in pregnancy . Besides that , terbinafine passes into the breast milk , which could have an effect on a nursing baby . For severe sporotrichosis cases that need to be treated during pregnancy , amphotericin B is recommended [3–5] . Since systemic treatment is hardly possible , local alternative treatment plays an important role in pregnancy . Thermotherapy is the most reported therapeutic option described in this group of patients entailing weeks of daily self-application of heat to the lesions , and requires a faithfully application with a certain caution to avoid skin burns [3–7] . Cryosurgery is an effective and safe method , when applied by well-trained staff , being a useful therapeutic resource for many infectious skin diseases [8 , 9] . Regarding the treatment of cutaneous sporotrichosis , it has already been reported as an effective adjuvant therapy when associated with oral antifungals [9–11] . However , to the best of our knowledge , has not yet been evaluated in pregnant women . The authors report four cases of pregnant women with cutaneous sporotrichosis that were successfully treated with cryosurgery .
The study was approved by the Ethical Committee of the INI , Fiocruz , Rio de Janeiro , Brazil ( CAAE 55348416 . 5 . 0000 . 5262 ) . The patients’ data were anonymized/de-identified to protect patients’ privacy/confidentiality . The authors reviewed the clinical records of pregnant patients diagnosed with cutaneous sporotrichosis who were treated at the cryosurgery outpatient clinic of the Laboratory of Clinical Research in Infectious Dermatology , Evandro Chagas National Institute of Infectious Diseases ( INI ) , Oswaldo Cruz Foundation ( Fiocruz ) from 2006 to 2016 . Briefly , the protocol of pregnant women with sporotrichosis included isolation of Sporothrix spp . in clinical specimens [2] , complete blood count , and biochemical tests . They were instructed to perform thermotherapy with warm compresses for 20 minutes 3 times a day [7] . Subsequent follow-up was scheduled monthly or anytime in case of worsening of the lesions . For non-adherent patients or those who did not desire to perform thermotherapy for sporotrichosis treatment , cryosurgery was offered , and that was the case of the patients included in this work . Patients that received any other type of treatment for sporotrichosis besides cryosurgery were excluded . Cryosurgery sessions were carried out monthly , performed by dermatologists , up to clinical cure . In each session , lesions were treated with two cycles of 10 to 30 seconds of freeze time with liquid nitrogen in spray form . Clinical cure was defined as complete healing of the lesions . In a general way , once Sporothrix spp . was isolated , molecular identification of the species was performed using the T3B PCR fingerprinting method [12] .
From 2006 to 2016 , 218 adult patients diagnosed with sporotrichosis , by fungal isolation in culture , were treated with cryosurgery . From these 218 patients , 8 were pregnant women , and 4 of them were treated exclusively with cryosurgery . These 4 patients were at the second trimester of pregnancy and their age ranged from 18 to 34 years . All of them lived in Rio de Janeiro state , Brazil . Two of them worked with domestic duties . The patients presented ulcerovegetative or nodular ulcerovegetative lesions ( Fig 1 ) . Due to technical reasons , molecular identification of the agent was feasible in two patients ( cases 3 and 4—Table 1 ) , and the isolates were identified as S . brasiliensis . Complete blood count , and biochemical tests performed before , during , and after the treatment were within the normal ranges . Cryosurgery was well tolerated with no need for local anesthesia . The number of cryosurgery sessions ranged from 1 to 3 . All the patients were discharged after a complete cure . No adverse reactions were observed during the treatment as well as no relapses were documented after delivery . Epidemiological , clinical , and therapeutic data of the patients are detailed in Table 1 .
Despite the low number of pregnant women affected and the benign clinical course , the treatment of sporotrichosis in pregnancy can be always considered challenging . In these cases , topical alternative therapeutic resources are safer than systemic drugs and should be considered whenever possible . Since the 1950s , thermotherapy was reported as the unique topical method for the treatment of sporotrichosis in pregnancy , with a strength of recommendation and quality of evidence considered as BIII [4] . Cryosurgery emerges as a useful tool for many infectious skin diseases , with effects of local cellular and humoral inflammatory response induction in the tissue , with its necrotic effect and , consequently , destructive for the infectious agents [8 , 9] . Cryosurgery has been used as an adjuvant treatment in sporotrichosis , especially in residual lesions or in cases of ulcerovegetative or nodular ulcerovegetative thick lesions since it allows a good penetration of liquid nitrogen in spray form [13] . In other subcutaneous mycoses such as chromoblastomycosis , cryosurgery has been indicated as an isolated method or associated to systemic antifungal agents with good results [14] . Some authors have warned about the risk of lymphatic dissemination with invasive methods performed without systemic drugs in cases of chromoblastomycosis [15 , 16] . In contrast with other procedures , cryosurgery is not only an ablative technique but also promotes an immune response , what could reduce this risk . A recent study with murine model found that cryosurgery was responsible for an increase in antigen-presenting dendritic cells ( DCs ) , neutrophils and macrophages in subcutaneous tissue , as well as migration of DCs to regional lymph nodes [17] . Cryosurgery is contraindicated for patients who are sensitive to cold ( cold urticaria , cryoglobulinemia , or cryofibrinogenemia ) and should be avoided in extensive lesions or flexor surfaces due to the risk of fibrosis [18] . Until now , cryosurgery for sporotrichosis treatment has been poorly explored and documented especially considering cases that involve a supposed more virulent phylogenetic species such as S . brasiliensis . All patients herein reported came from hyperendemic areas of sporotrichosis in Rio de Janeiro state , and become infected during pregnancy . None referred prior trauma with plants , but only contact and/or trauma with cats , in agreement with the zoonotic epidemic profile reported in the literature [2] . Although S . brasiliensis , could be identified in only two cases , it is well known that it is the main species involved in Rio de Janeiro epidemic . All patients presented cutaneous-limited clinical forms on the extremities , similar to previous publications [3 , 5] , in contrast with other mycoses , which can be more aggressive during pregnancy [19] . This work suggests that cryosurgery is a safe and well-tolerated method , easy to perform , being a promising alternative in the treatment of cutaneous sporotrichosis in pregnant women . Further studies with a larger number of patients are necessary to confirm efficacy of cryosurgery for sporotrichosis in pregnant patients .
|
Sporotrichosis is a cosmopolitan disease , considered the most important subcutaneous mycosis in Latin America . Since 1998 , there is an ongoing cat-transmitted zoonotic epidemic of sporotrichosis occurring in Rio de Janeiro , Brazil . Pregnant women are a vulnerable population occasionally affected that require special attention regarding sporotrichosis treatment . Antifungal drugs should be avoided because of their potential risks to the fetus , unless in severe cases when amphotericin B ( an intravenous antifungal drug ) can be indicated . In this context , local measures are the treatment of choice . Cryosurgery consists in local application of intense cold using liquid nitrogen to destroy some infectious , tumoral and inflammatory cutaneous diseases . It is scarcely reported in the literature for the treatment of sporotrichosis , especially in pregnant women for whom local heat is most used . This works aims to describe the clinical response and outcome of cryosurgery for the treatment of sporotrichosis in four pregnant women . All patients reached clinical cure after one to three sessions . These results suggest that cryosurgery can be a well-tolerated , safe , and efficient method for the treatment of sporotrichosis in pregnancy .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"antimicrobials",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"maternal",
"health",
"obstetrics",
"and",
"gynecology",
"drugs",
"geographical",
"locations",
"microbiology",
"antifungals",
"surgical",
"and",
"invasive",
"medical",
"procedures",
"women's",
"health",
"sporotrichosis",
"signs",
"and",
"symptoms",
"pregnancy",
"pharmacology",
"fungal",
"diseases",
"infectious",
"diseases",
"mycology",
"south",
"america",
"amphotericin",
"lesions",
"cryosurgery",
"brazil",
"people",
"and",
"places",
"diagnostic",
"medicine",
"microbial",
"control",
"biology",
"and",
"life",
"sciences"
] |
2018
|
Cryosurgery for the treatment of cutaneous sporotrichosis in four pregnant women
|
The ability of specific neurons to regenerate their axons after injury is governed by cell-intrinsic regeneration pathways . In Caenorhabditis elegans , the JNK and p38 MAPK pathways are important for axon regeneration . Axonal injury induces expression of the svh-2 gene encoding a receptor tyrosine kinase , stimulation of which by the SVH-1 growth factor leads to activation of the JNK pathway . Here , we identify ETS-4 and CEBP-1 , related to mammalian Ets and C/EBP , respectively , as transcriptional activators of svh-2 expression following axon injury . ETS-4 and CEBP-1 function downstream of the cAMP and Ca2+–p38 MAPK pathways , respectively . We show that PKA-dependent phosphorylation of ETS-4 promotes its complex formation with CEBP-1 . Furthermore , activation of both cAMP and Ca2+ signaling is required for activation of svh-2 expression . Thus , the cAMP/Ca2+ signaling pathways cooperatively activate the JNK pathway , which then promotes axon regeneration .
The ability of a neuron to regenerate following injury is dependent on both its intrinsic growth capacity and the extracellular environment . When an axon is injured , intracellular levels of calcium ( Ca2+ ) and cyclic adenosine monophosphate ( cAMP ) increase [1] . The increase in cAMP levels activates protein kinase A ( PKA ) , which in turn activates the axon regeneration-promoting transcription factor CREB . PKA also promotes remodeling of the cytoskeleton , which is necessary for the formation and maintenance of the growth cone , a specialized structure necessary to initiate regeneration . Upon axon severance , regeneration signals are retrogradely transported from sites of damage and imported into the nucleus , where they induce the up-regulation of several transcription factors and drive the increased synthesis of proteins involved in neurite outgrowth [2 , 3] . Manipulation of these processes can improve the chances for successful axon regeneration . Nonetheless , our understanding of the intrinsic signaling pathways that promote this regenerative ability remains limited . The nematode Caenorhabditis elegans has recently emerged as a genetic model for studying the molecular control of axon regeneration [4 , 5] . Recent genetic studies have demonstrated that C . elegans axon regeneration is regulated by the p38 and JNK MAP kinase ( MAPK ) pathways , which consist of DLK-1 ( MAPKKK ) –MKK-4 ( MAPKK ) –PMK-3 ( MAPK ) and MLK-1 ( MAPKKK ) –MEK-1 ( MAPKK ) –KGB-1 ( MAPK ) , respectively [6–8] . The p38 MAPK signaling pathway promotes axon regeneration by activating the MAP kinase-activated protein kinase MAK-2 , which functions to stabilize the mRNA encoding the C/EBP family transcription factor CEBP-1 [7] . CEBP-1 in turn promotes axon regeneration , although the specific targets that mediate this response remain unknown . MAPK cascades can be inactivated by members of the MAPK phosphatase ( MKP ) family [9] . In C . elegans , the vhp-1 gene encodes a MKP that negatively regulates both the DLK-1–MKK-4–PMK-3 and MLK-1–MEK-1–KGB-1 MAPK pathways [8 , 10] . vhp-1 mutant animals are arrested during larval development , due to hyperactivation of the MAPK pathways . Furthermore , axon regeneration is enhanced in vhp-1 mutants [8] . In a previous effort to identify additional components involved in MAPK-mediated signaling , we isolated a number of svh ( suppressor of vhp-1 ) genes , which function as suppressors of vhp-1 larval lethality [11] . Two of these , svh-1 and svh-2 , encode a growth factor and its cognate receptor tyrosine kinase , respectively . SVH-1–SVH-2 signaling mediates the activation of the JNK cascade following axonal injury , a molecular event essential for neuronal regeneration but not for neuronal development . This specific effect on axon regeneration is determined by svh-2 gene expression , which is induced following axon injury in severed neurons ( Fig 1A ) [11] . In the present study , we investigated ets-4 and cebp-1 , two genes that emerged from our previous screen , and examined their potential roles in the regulation of axon regeneration . We demonstrate that ETS-4 and CEBP-1 act as transcriptional activators of svh-2 expression in response to axon injury . ETS-4 and CEBP-1 function as downstream effectors of the cAMP and Ca2+–p38 MAPK pathways , respectively . Our results indicate that the cAMP and Ca2+–p38 MAPK pathways , induced in response to axon injury , converge through the formation of an ETS-4–CEBP-1 transcription factor complex to transactivate svh-2 gene expression and SVH-2 receptor expression , which in turn activates the JNK pathway . Thus , the cAMP and Ca2+–p38 MAPK signaling pathways induce a transcription factor complex that ultimately up-regulates the JNK pathway and promotes axon regeneration .
To identify the transcription factors involved in axon injury-induced activation of svh-2 expression , we asked if any of the svh genes encode transcription factors . Among our svh genes we identified svh-5 , which encodes a member of the Ets transcription factor family and contains an Ets DNA binding domain and a PNT domain , a protein-protein interaction domain conserved in a subset of Ets proteins ( Fig 1B , 1C and 1D ) [12] . We therefore renamed svh-5 as ets-1 . To examine the effect of ets-1 on axon regeneration , we assayed regrowth after laser axotomy in γ-aminobutyric acid ( GABA ) -releasing D-type motor neurons , which extend their axons from the ventral to the dorsal nerve cord ( Fig 2A ) [4 , 6] . In young adult wild-type animals , laser-severed axons were able to initiate regeneration within 24 hr ( Fig 2A and 2B and S1 Table ) . Although the ets-1 ( ok286 ) deletion mutation ( Fig 1B ) slightly inhibited axon regeneration , this effect was not statistically significant ( Fig 2B and S1 Table ) . The C . elegans genome contains ten ets genes , among which PNT domains are found in only ETS-1 and ETS-4 ( Fig 1B and 1C ) [13 , 14] . We found that in contrast to ets-1 , the frequency of axon regeneration in ets-4 ( ok165 ) deletion mutants ( Fig 1B ) was reduced significantly ( Fig 2A and 2B and S1 Table ) . The morphology of D-type motor neurons was normal in ets-4 mutants . These results suggest that ets-1 and ets-4 are involved in vhp-1-mediated larval development and axon regeneration , respectively . To test whether ETS-4 can act in a cell-autonomous manner , we expressed the ets-4 cDNA from the unc-25 or mec-7 promoters in ets-4 mutants . The ets-4 defect was rescued by expression of ets-4 in D-type motor neurons by the unc-25 promoter but not by expression in sensory neurons by the mec-7 promoter ( Fig 2C and S1 Table ) . These results demonstrate that ETS-4 functions cell autonomously in D-type motor neurons . Since expression of svh-2 is induced by axon injury [11] , we next examined whether ETS-4 is involved in svh-2 expression in response to axon injury . For this purpose we used the transgene Psvh-2::nls::venus , which consists of the svh-2 promoter driving the fluorescent protein VENUS fused to a nuclear localization signal ( NLS ) [11] . In wild-type animals , Psvh-2::nls::venus expression was induced in D-type neurons in 52% of the animals following laser surgery ( Fig 3A and 3B and S1 Fig ) . In contrast , we found that the ets-4 ( ok165 ) mutation abolished Psvh-2::nls::venus induction in response to laser surgery in D-type neurons ( Fig 3A and 3B ) . These results suggest that ETS-4 is a transcription factor required for axon injury-induced up-regulation of svh-2 expression . If ETS-4 is required for axon regeneration through activation of svh-2 gene expression , the ets-4 defect should be suppressed by constitutive expression of svh-2 . We replaced the svh-2 promoter with the unc-25 promoter to generate the Punc-25::svh-2 transgene and introduced this as an extrachromosomal array into an ets-4 ( ok165 ) mutant . As expected , this construct was able to rescue the ets-4 ( ok165 ) animals ( Fig 2C and S1 Table ) . These results support the idea that ETS-4 acts as a transcription factor regulating svh-2 expression in response to axon injury . Another gene obtained in our svh screen [11] was the svh-8/cebp-1 gene , which encodes a homolog of mammalian C/EBP ( CCAAT/enhancer-binding protein ) ( Fig 1B ) , and which indeed is known to be involved in axon regeneration [7] . We confirmed that animals having the cebp-1 ( tm2807 ) mutation ( Fig 1B ) are defective in axon regeneration in D-type motor neurons ( Fig 2C and S1 Table ) . Therefore , we examined whether CEBP-1 is required for expression of the svh-2 gene in response to axon injury . We found that in cebp-1 ( tm2807 ) mutants , laser surgery was unable to induce the expression of Psvh-2::nls::venus in D-type neurons ( Fig 3B ) . Yan et al . showed that axotomy-induced signaling via the DLK-1–p38 MAPK pathway promoted the local translation and stabilization of cebp-1 transcripts , accumulation of which is required for axon regeneration [7] . Since CEBP-1 functions downstream of the DLK-1–p38 MAPK pathway in axon regeneration , we investigated whether the DLK-1 pathway is involved in axon injury-induced expression of svh-2 . We found that expression of the svh-2 reporter in D-type neurons was not induced by axon injury in dlk-1 ( km12 ) null mutants ( Fig 3B ) . These results support the possibility that the DLK-1–p38 MAPK pathway is involved in axon injury-induced expression of the svh-2 gene . If the svh-2 gene is the only transcriptional target of CEBP-1 , the cebp-1 defect should be suppressed by constitutive expression of svh-2 . However , in contrast to ets-4 , the Punc-25::svh-2 transgene was unable to rescue the defect associated with the cebp-1 ( tm2807 ) mutation ( Fig 2C and S1 Table ) . Partial rescue of the phenotype may be expected if CEBP-1 has additional transcriptional targets that act in parallel to promote regeneration . Consistent with this , the Punc-25::svh-2 transgene weakly suppressed the regeneration defect caused by a different , stronger allele of cebp-1 ( u819 ) ( Fig 2C and S1 Table ) . These results suggest that CEBP-1 has other target ( s ) in addition to svh-2 that function in regeneration after axon injury . To determine the region in the svh-2 promoter important for axon regeneration and gene induction following axon injury , we generated a series of svh-2 promoter deletions ( Fig 4A ) . We have previously demonstrated that a 6 . 2 kb region of the svh-2 promoter , driving the svh-2 gene , is sufficient to rescue defective axon regeneration in svh-2 mutants ( Fig 4B and S1 Table ) . We observed here that a 2 . 6 kb region of the promoter upstream of the translational start site is also sufficient to rescue this defect , but a 0 . 5 kb promoter region is not ( Fig 4B and S1 Table ) . Consistent with this , we confirmed that expression of a Psvh-2::nls::venus construct in D-type neurons after exposure to axon injury could be induced by the 2 . 6 kb promoter region , but not by the 0 . 5 kb region ( Fig 4C ) . These results indicate that the promoter region of the svh-2 gene between 0 . 5 kb and 2 . 6 kb upstream of the translational start site is important for the transcriptional induction of the svh-2 gene and for axon regeneration . To assess whether ETS-4 and CEBP-1 directly regulate svh-2 expression , we searched the svh-2 promoter region for Ets- and C/EBP-binding sites . Mammalian Ets binds the consensus sequence , 5’-GGAA/T-3’ [15] , and the promoter region of the svh-2 gene between 0 . 5 kb and 2 . 6 kb contains several possible Ets-binding motifs . This promoter region also has two C/EBP-binding motifs ( 5’-TTGNNCAA-3’ ) [16] . Of particular note , there is an Ets- and a C/EBP-binding site located in close proximity to one another at 1370 and 1378 base pairs upstream of the translational start site , respectively ( Fig 4A ) . To determine whether these binding sites are required for axon regeneration and axon injury-induced expression of the svh-2 gene , we converted the Ets consensus GGAA sequence to TTAA and the C/EBP consensus TTGGCCAA to CCGGCCAA ( Fig 4A ) . We found that svh-2 gene constructs carrying either of these point mutations in their promoter failed to rescue the svh-2 defect in axon regeneration ( Fig 4B and S1 Table ) . Furthermore , we found that axon injury-induced Psvh-2::nls::venus expression in D-type neurons was abolished by alteration of either of the Ets or C/EBP binding motif ( Fig 4C ) . These results suggest that ETS-4 and CEBP-1 bind to the svh-2 promoter via their respective binding site , and therein drive svh-2 expression in response to axon injury . Thus , svh-2 promoter activity appears to depend primarily on this combined Ets–C/EBP motif , raising the possibility that ETS-4 and CEBP-1 may physically interact at this site to drive svh-2 gene expression . In a yeast two-hybrid assay we found that ETS-4 and CEBP-1 could indeed interact ( Fig 5A and 5B ) , suggesting that ETS-4 may cooperate with CEBP-1 on the svh-2 promoter to activate transcription . How is ETS-4 regulated in axon regeneration ? The functions of mammalian Ets transcription factors are regulated by phosphorylation [12 , 17] . As ETS-4 contains a protein kinase A ( PKA ) phosphorylation consensus sequence ( Arg-Arg-Xxx-Ser ) at Ser-73 near its PNT domain ( Fig 6A ) , we asked whether PKA phosphorylates ETS-4 at this residue . We performed in vitro kinase assays with active PKA and immuno-purified HA-tagged ETS-4 and confirmed that PKA phosphorylated HA-ETS-4 ( Fig 6B ) . To determine if PKA can phosphorylate ETS-4 on Ser-73 , we generated a mutant form of ETS-4 [ETS-4 ( S73A ) ] , in which Ser-73 is mutated to alanine . In vitro kinase assays showed that the S73A mutation abolished phosphorylation of ETS-4 by PKA ( Fig 6B ) . These results demonstrate that PKA phosphorylates Ser-73 of ETS-4 in vitro . We next addressed the biological importance of ETS-4 Ser-73 phosphorylation . When the phosphorylation-defective mutant ETS-4 ( S73A ) was expressed under the control of the unc-25 promoter in ets-4 ( ok165 ) null mutants , the defect in axon regeneration was not rescued ( Fig 6C and S1 Table ) . In contrast , expression of a phospho-mimetic form of ETS-4 , ETS-4 ( S73E ) , in ets-4 ( ok165 ) mutants by the unc-25 promoter was able to rescue the regeneration defect ( Fig 6C and S1 Table ) . Thus , phosphorylation of Ser-73 is important for the function of ETS-4 in the activation of the regeneration pathway . We next asked how PKA-mediated phosphorylation might regulate ETS-4 in this regeneration pathway . Since Ser-73 is located near the PNT domain , which is a protein-protein interaction domain , we examined the effect of ETS-4 Ser-73 phosphorylation on its interaction with CEBP-1 . We found that a non-phosphorylatable ETS-4 ( S73A ) mutant form lost the ability to associate with CEBP-1 , whereas a phosphorylation mimicking mutant ETS-4 ( S73E ) was able to interact with CEBP-1 ( Fig 5A and 5B ) . Furthermore , the interaction of ETS-4 ( S73E ) with CEBP-1 was stronger than that of wild-type ETS-4 ( Fig 5B ) . These results suggest that PKA-mediated phosphorylation of ETS-4 Ser-73 promotes the formation of an ETS-4–CEBP-1 complex . PKA is activated by cAMP , and cAMP signals have been implicated in axonal regeneration in many systems [18–21] . The C . elegans acy-1 gene encodes the neuronal adenylyl cyclase , and we observed that animals expressing a loss-of-function mutant , acy-1 ( nu329 ) , were defective in axon regeneration ( Fig 6C and S1 Table ) [18] . We also found that in animals carrying acy-1 ( nu329 ) , Psvh-2::nls::venus was not induced in D-type neurons in response to axon injury ( Fig 7A ) . To examine whether ETS-4 functions downstream of cAMP in axon regeneration , we tested the effects of the phosphor-mimetic ets-4 mutation on acy-1 phenotypes . Expression of ETS-4 ( S73E ) by the unc-25 promoter failed to suppress the regeneration defect observed in acy-1 ( nu329 ) mutants ( Fig 6C and S1 Table ) . This result is consistent with the fact that cAMP is known to be important for regeneration and regulates many pathways . Thus , ETS-4 is not the only target of cAMP signaling that functions in axon regeneration . In contrast to axon regeneration , we found that expression of ETS-4 ( S73E ) , but not wild-type ETS-4 , was able to induce svh-2 expression in acy-1 ( nu329 ) mutants ( Fig 7A ) . However , this induction by ETS-4 ( S73E ) was not constitutive and was not observed in the absence of axon injury . This result suggests that induction of svh-2 expression requires the injury-dependent activation of both ETS-4 and CEBP-1 and that activation of ETS-4 alone is not sufficient to induce svh-2 expression . Thus , ETS-4 is required for the induction of svh-2 transcription in response to axonal injury , and this occurs downstream of cAMP signaling through PKA-mediated phosphorylation . Axotomy induces an intracellular increase in Ca2+ via the action of voltage-gated Ca2+ channels , and this can promote axon regeneration in a manner dependent on the DLK-1-p38 MAPK pathway [18 , 22] . Previous studies have shown that a gain-of-function ( gf ) mutation in a subunit of one voltage-gated Ca2+ channel , egl-19 ( ad695gf ) , enhances Ca2+ influx in PLM neurons after axotomy and , subsequently , drives the formation of an active DLK-1 homomeric protein complex [22] . We therefore investigated whether activation of the p38-CEBP-1 pathway by the egl-19 ( ad695gf ) mutation affects expression of the svh-2 gene . We found that the egl-19 ( ad695gf ) mutation did not cause constitutive expression of the Psvh-2::nls::venus reporter in D-type neurons ( Fig 7A and 7B and S2 Fig ) . Since the ETS-4 transcription factor is regulated by the cAMP-PKA pathway , we examined the effect of activation of the cAMP pathway on the transcriptional induction of the svh-2 gene . Treatment of animals with forskolin is expected to cause an increase in cAMP levels by activating adenylyl cyclase [18] . Forskolin treatment of wild-type animals not subjected to axon injury failed to induce Psvh-2::nls::venus expression in D-type neurons ( Fig 7A ) . Thus , activation of either the p38-CEBP-1 or cAMP-ETS-4 pathway is not sufficient to induce transcription of the svh-2 gene . We therefore speculated that simultaneous activation of both pathways might be required . Consistent with this , we found that when egl-19 ( ad695gf ) mutants were treated with forskolin , 28% of the animals expressed the Psvh-2::nls::venus reporter in D-type neurons , even in the absence of axon injury ( Figs 7A and 7B and S2 ) . We next examined whether the effect of the egl-19 gf mutation on svh-2 expression is mediated by the DLK-1 pathway . Activated DLK-1 kinase is targeted for degradation by the E3 ubiquitin ligase RPM-1 , thereby modulating the duration of signaling [7] . Constitutive activation of the DLK-1 pathway induces developmental defects that mimic rpm-1mutants . Moreover , rpm-1 mutants display a MAPK-dependent improvement in axon regeneration [8] . We found that the rpm-1 mutation caused constitutive expression of the svh-2 gene in D-type neurons when cultured in the presence of forskolin ( Fig 7A ) . Taken together , these results suggest that induction of the svh-2 gene is dependent on activation of both cAMP and Ca2+-DLK-1 signaling through an Ets-C/EBP transcription factor complex .
MAPK signaling cascades are evolutionally conserved in eukaryotes from yeast to mammals and play key roles in many aspects of neuronal development and function [23] . Recent genetic studies have shown that the DLK-1–MKK-4–PMK-3 p38 MAPK and the MLK-1–MEK-1–KGB-1 JNK pathways regulate axon regeneration in C . elegans [6–8] . The DLK MAPKKKs are required for axon regeneration in both Drosophila melanogaster and mice [24–26] . Similarly , JNK-mediated activation of c-Jun is important for axonal outgrowth of neurons in axotomized rat nodose and dorsal root ganglia ( DRG ) and mouse DRG [27 , 28] . These discoveries suggest that the core machinery that regulates axon regeneration is conserved from worms to mammals . In C . elegans , the DLK-1 p38 MAPK pathway promotes mRNA stability and local axonal translation of the bZip transcription factor CEBP-1 through the MAPKAP kinase MAK-2 [7] . Although the precise steps by which DLK-1 is activated in axon regeneration remain unknown , a Ca2+-dependent mechanism of activation has been recently described [22] . The JNK MAPK pathway is activated following axonal injury by growth factor-like SVH-1 engagement of its cognate receptor tyrosine kinase SVH-2 ( Fig 8 ) [11] . SVH-1 belongs to the HGF/plasminogen family and SVH-2 is homologous to the HGF receptor Met , suggesting that SVH-1–SVH-2 functions as a ligand–receptor pair in axon regeneration . The svh-1 gene is constitutively expressed in ADL sensory neurons in the head and SVH-1 acts on injured neurons . In contrast , expression of svh-2 is induced by axonal injury [11] . SVH-1-SVH-2 signaling does not affect axon development per se , but rather is specific to axon regeneration . This specificity is determined by the injury-induced expression of the svh-2 gene . In this study , we found that the transcription factor CEBP-1 is required for the injury-induced transcriptional activation of svh-2 expression . Although there remains the possibility that EGL-19–DLK-1– CEBP-1 pathway may activate svh-2 expression through another MAPK , it is known that CEBP-1 acts in the p38 MAPK pathway [7] . Since SVH-2 functions in the JNK MAPK pathway [11] , these results suggest the p38 MAPK pathway functions upstream of JNK MAPK activation in the response to axon injury ( Fig 8 ) . Enforced expression of svh-2 in cebp-1 mutants failed to efficiently suppress the defect in axon regeneration seen in these mutants , indicating that CEBP-1 transcriptionally regulates other targets in addition to the svh-2 gene in response to p38 MAPK activation . ETS-4 was also identified as another transcription factor involved in the activation of svh-2 expression in response to axon injury . A previous study demonstrated that ETS-4 functions in lipid transport , lipid metabolism and innate immunity [14] . ETS-4 is a transcriptional regulator of aging , and shares transcriptional targets with the GATA and FOXO transcriptional regulators . An expression profiling study identified seventy ETS-4-target genes , 54 of which have identifiable Ets-binding consensus motifs in their promoter regions [14] . However , this analysis did not identify the svh-2 gene as an ETS-4 target , presumably because ETS-4-dependent svh-2 expression occurs only in severed neurons and not under normal conditions . We show that ETS-4 acts downstream of cAMP signaling , and that transcriptional up-regulation of svh-2 by ETS-4 requires PKA-dependent phosphorylation at Ser-73 . However , a mutant of ETS-4 that mimics constitutive phosphorylation of the Ser-73 residue failed to rescue the defect in axon regeneration in acy-1 mutants defective in cAMP production . These results suggest that cAMP activates other pathways , in addition to the SVH-2–JNK MAPK pathway , that are required for axon regeneration . We demonstrate that phosphorylation of ETS-4 by PKA leads to the formation of a complex containing ETS-4 and CEBP-1 , and that this protein-protein interaction is necessary for activation of svh-2 expression in response to axon injury . It has been shown for mammals that direct interactions between C/EBPs and specific Ets family members are important for eosinophil lineage determination [29] . In this case , the physical interaction between Ets-1 and C/EBPα proteins is mediated by their DNA-binding domains , and these complexes transactivate their target genes by high-affinity binding to combined motifs in the target promoters . Similarly , the svh-2 promoter region has an Ets- and a C/EBP-binding site located in close proximity to one another and these binding sites are required for axon injury-induced expression of the svh-2 gene . It is tempting to speculate that the high affinity of the CEBP-1–ETS-4 interaction is important for synergy between the two factors . We conclude that CEBP-1 is a biologically important and relevant interacting partner of ETS-4 that is involved in the transcriptional activation of svh-2 gene expression . Similar to vertebrate neurons , increased Ca2+ and cAMP facilitate axon regeneration in severed C . elegans neurons [18] . It is likely that some effects of elevated Ca2+ or cAMP on axon regeneration are mediated at the transcriptional level [30] . Interestingly , CEBP-1 acts downstream of the Ca2+–p38 MAPK pathway and ETS-4 functions downstream of the cAMP signaling pathway . Based on our findings , we propose a model wherein axon injury initiates cAMP signaling and the Ca2+–p38 MAPK pathway , which together function to induce the formation of an ETS-4–CEBP-1 transcription factor complex . This complex binds to the svh-2 promoter to induce svh-2 expression , and SVH-1 signaling via SVH-2 then activates the JNK pathway ( Fig 8 ) . Thus , two injury-signaling pathways ( p38 MAPK and cAMP/PKA ) converge to regulate expression of the svh-2 gene after injury , which ultimately promotes axon regeneration . The identification and characterization of the signaling pathways that regulate axon regeneration in C . elegans should yield numerous insights into the mechanisms used by nervous systems to regulate similar processes across metazoans .
C . elegans strains used in this study are listed in S2 Table . All strains were maintained on nematode growth medium ( NGM ) plates and fed with bacteria of the OP50 strain , as described previously [31] . Axotomy and microscopy were performed as described previously [11] . All animals were subjected to axotomy at the young adult stage . The imaged commissures that had growth cones or small branches present on the proximal fragment were counted as “regenerated” . The proximal fragments that showed no change after 24 hr were counted as “no regeneration” . A minimum of 20 individuals with 1–3 axotomized commissures were observed for most experiments . Standard fluorescent images of transgenic worms were observed under a Zeiss Plan-APOCHROMAT 63 X objective of a Zeiss Axioplan II fluorescent microscope and photographed with a Hamamatsu 3CCD camera . Confocal fluorescent images were taken on an Olympus FV500 confocal laser scanning microscope with a 40 X objective . Punc-25::ets-4 and Pmec-7::ets-4 were generated by inserting the ets-4 cDNA isolated from a cDNA library into the pCZ325 vector and pPD52 . 102 vectors , respectively . Punc-25::ets-4 ( S73A ) and Punc-25::ets-4 ( S73E ) were made by oligonucleotide-directed PCR using Punc-25::ets-4 as a template and the mutations were verified by DNA sequencing . pLexA-ETS-4 , pLexA-ETS-4 ( S73A ) and pLexA-ETS-4 ( S73E ) plasmids were constructed by inserting the wild-type and mutagenized ets-4 cDNAs into the pBTM116 vector . The Psvh-2 ( 6 . 2 kb ) ::nls::venus plasmid has been described previously [11] . For the construction of Psvh-2 ( 2 . 6 kb ) ::nls::venus and Psvh-2 ( 0 . 5 kb ) ::nls::venus plasmids , Psvh-2 ( 6 . 2 kb ) ::nls::venus was digested with either HindIII or PsiI , respectively , and then self-ligated . Psvh-2 ( Ets-bm ) ::nls::venus and Psvh-2 ( C/EBP-bm ) ::nls::venus plasmids were made by oligonucleotide-directed PCR using Psvh-2 ( 2 . 6 kb ) ::nls::venus as a template and the mutations were verified by DNA sequencing . These mutated promoters were then used to make Psvh-2 ( 6 . 2 kb ) ::svh-2 , Psvh-2 ( 2 . 6 kb ) ::svh-2 , Psvh-2 ( 0 . 5 kb ) ::svh-2 , Psvh-2 ( Ets-bm ) ::svh-2 and Psvh-2 ( C/EBP-bm ) ::svh-2 , respectively . The pCMV-HA-ETS-4 plasmid was made by inserting the ets-4 cDNA into the pCMV-HA vector . The pACTII-CEBP-1 plasmid was made by inserting the cebp-1 cDNA , isolated from a C . elegans cDNA library by PCR , into the pACTII vector . The lin-15 plasmid is a gift from Dr . S . Takagi ( Nagoya University ) . Other plasmids , including Pttx-3::gfp , Pmyo-2::dsredm , Punc-25::cfp and Punc-25::svh-2 have been described previously [11 , 32] . Transgenic animals were obtained as described [33] . Psvh-2 ( 6 . 2 kb ) ::svh-2 ( 25 ng/μl ) , Psvh-2 ( 2 . 6 kb ) ::svh-2 ( 25 ng/μl ) , Psvh-2 ( 0 . 5 kb ) ::svh-2 ( 25 ng/μl ) , Psvh-2 ( Ets-bm ) ::svh-2 ( 25 ng/μl ) and Psvh-2 ( C/EBP-bm ) ::svh-2 ( 25 ng/μl ) plasmids were used in kmEx529 [Psvh-2 ( 6 . 2 kb ) ::svh-2 + Pmyo-2::dsredm] , kmEx530 [Psvh-2 ( 2 . 6 kb ) ::svh-2 + Pmyo-2::dsredm] , kmEx531 [Psvh-2 ( 0 . 5 kb ) ::svh-2 + Pmyo-2::dsredm] , kmEx532 [Psvh-2 ( Ets-bm ) ::svh-2 + Pmyo-2::dsredm] , kmEx533 [Psvh-2 ( C/EBP-bm ) ::svh-2 + Pmyo-2::dsredm] , respectively . Punc-25::ets-4 ( 25 ng/μl ) , Pmec-7::ets-4 ( 25 ng/μl ) , Punc-25::ets-4 ( S73A ) ( 25 ng/μl ) and Punc-25::ets-4 ( S73E ) ( 25 ng/μl ) plasmids were used in kmEx534 [Punc-25::ets-4 + Pmyo-2::dsredm] , kmEx535 [Pmec-7::ets-4 + Pmyo-2::dsredm] , kmEx536 [Punc-25::ets-4 ( S73A ) + Pmyo-2::dsredm] , kmEx537 [Punc-25::ets-4 ( S73E ) + Pmyo-2::dsredm] , kmEx544 [Punc-25::ets-4 + Pttx-3::gfp] and kmEx545 [Punc-25::ets-4 ( S73E ) + Pttx-3::gfp] , respectively . Punc-25::cfp ( 50 ng/μl ) , Psvh-2 ( 6 . 2 kb ) ::nls::venus ( 75 ng/μl ) , Psvh-2 ( 2 . 6 kb ) ::nls::venus ( 75 ng/μl ) , Psvh-2 ( Ets-bm ) ::nls::venus ( 75 ng/μl ) and Psvh-2 ( C/EBP-bm ) ::nls::venus ( 75 ng/μl ) plasmids were used in kmEx538 ( Punc-25::cfp + lin-15 ) , kmEx539 [Psvh-2 ( 6 . 2 kb ) ::nls::venus + Pmyo-2::dsredm] , kmEx540 [Psvh-2 ( 2 . 6 kb ) ::nls::venus + Pmyo-2::dsredm] , kmEx541 [Psvh-2 ( 0 . 5 kb ) ::nls::venus + Pmyo-2::dsredm] , kmEx542 [Psvh-2 ( Ets-bm ) ::nls::venus + Pmyo-2::dsredm] and kmEx543 [Psvh-2 ( C/EBP-bm ) ::nls::venus + Pmyo-2::dsredm] , respectively . pBTM116-ETS-4 ( WT , S73E , S73A ) and pACTII-CEBP-1 plasmids were cotransformed into the Saccharomyces cerevisiae reporter strain L40u [MATa trp1 ura3 leu2 his3 LYS2:: ( lexAop ) 4-HIS3] and allowed to grow on SC-Leu-Trp plates . Transformants grown on these plates were streaked out onto SC-Leu-Trp-His plates containing 80 mM 5-aminotriazole and incubated at 30℃ for 4 days . For β-galactosidase assays , the NMY51 [MATa trp1 ura3:: ( lexAop ) 8-lacZ leu2 his3 LYS2:: ( lexAop ) 4-HIS3 ade2:: ( lexAop ) 8-ADE2 GAL4] strain ( Dualsystems Biotech ) was used as the host strain . The β-galactosidase assay was performed as described previously [34] . To examine nuclei of D neuron cells , we fixed worms in 4% paraformaldehyde for 1 hr and permeabilized with methanol for 5 min . Next , the worms were stained with 0 . 1 μg/ml of the DNA-binding dye 4 , 6-diamidino-2-phenyl-indole and mounted on 2% agarose slides for viewing using fluorescence imaging . Expression of VENUS fluorescence was quantified using the ImageJ program ( NIH ) . The cell bodies of severed D neurons were outlined with closed polygons and the fluorescent intensities within these areas were determined ( Is ) . The cell bodies of unsevered D neurons in the same animal were analyzed similarly as controls ( Iu ) . To determine the background intensity of each cell , the same polygon was placed in the area neighboring the cell body and fluorescence measured ( Ibs and Ibu , respectively ) . The relative signal intensity ( Ir ) was calculated as ( Is-Ibs ) / ( Iu-Ibu ) . Cells having an Ir>5 were categorized as “expressed” . Transfection of transgenes into COS-7 cells , preparation of the cell lysates and immunoblotting procedures have been described previously [10] . Anti-phospho-PKA substrate ( RRX ( p ) S/ ( p ) T ) [ ( p ) , phosphorylated] rabbit monoclonal antibody ( 100G7E ) , was purchased from Cell Signaling Technology . Statistical analyses were carried out as described previously [11] . Briefly , confidence intervals ( 95% ) were calculated by the modified Wald method and two-tailed P values were calculated using Fisher’s exact test ( http://www . graphpad . com/quickcalcs/contingency1/ ) . The Welch’s t -test was performed by using t-test calculator ( http://www . graphpad . com/quickcalcs/ttest1/ ) .
|
An axon’s ability to regenerate after injury is governed by cell-intrinsic regeneration pathways . In C . elegans , the JNK and p38 MAPK pathways play an important role in axon regeneration . The JNK pathway is activated by growth factor SVH-1 , which signals through its receptor SVH-2 . It is known that expression of the svh-2 gene is induced in response to axonal injury , however the molecular mechanisms underlying this induction have been unknown . Here , we demonstrate that induction of svh-2 expression in response to axon injury involves the transcription factors ETS-4 and CEBP-1 , which function downstream of the cAMP and Ca2+–p38 MAPK pathways , respectively . Our results suggest that these two injury-signaling pathways converge to regulate expression of the svh-2 gene and thereby promote axon regeneration .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Axon Regeneration Is Regulated by Ets–C/EBP Transcription Complexes Generated by Activation of the cAMP/Ca2+ Signaling Pathways
|
Cyclophilin B ( CyPB ) , encoded by PPIB , is an ER-resident peptidyl-prolyl cis-trans isomerase ( PPIase ) that functions independently and as a component of the collagen prolyl 3-hydroxylation complex . CyPB is proposed to be the major PPIase catalyzing the rate-limiting step in collagen folding . Mutations in PPIB cause recessively inherited osteogenesis imperfecta type IX , a moderately severe to lethal bone dysplasia . To investigate the role of CyPB in collagen folding and post-translational modifications , we generated Ppib−/− mice that recapitulate the OI phenotype . Knock-out ( KO ) mice are small , with reduced femoral areal bone mineral density ( aBMD ) , bone volume per total volume ( BV/TV ) and mechanical properties , as well as increased femoral brittleness . Ppib transcripts are absent in skin , fibroblasts , femora and calvarial osteoblasts , and CyPB is absent from KO osteoblasts and fibroblasts on western blots . Only residual ( 2–11% ) collagen prolyl 3-hydroxylation is detectable in KO cells and tissues . Collagen folds more slowly in the absence of CyPB , supporting its rate-limiting role in folding . However , treatment of KO cells with cyclosporine A causes further delay in folding , indicating the potential existence of another collagen PPIase . We confirmed and extended the reported role of CyPB in supporting collagen lysyl hydroxylase ( LH1 ) activity . Ppib−/− fibroblast and osteoblast collagen has normal total lysyl hydroxylation , while increased collagen diglycosylation is observed . Liquid chromatography/mass spectrometry ( LC/MS ) analysis of bone and osteoblast type I collagen revealed site-specific alterations of helical lysine hydroxylation , in particular , significantly reduced hydroxylation of helical crosslinking residue K87 . Consequently , underhydroxylated forms of di- and trivalent crosslinks are strikingly increased in KO bone , leading to increased total crosslinks and decreased helical hydroxylysine- to lysine-derived crosslink ratios . The altered crosslink pattern was associated with decreased collagen deposition into matrix in culture , altered fibril structure in tissue , and reduced bone strength . These studies demonstrate novel consequences of the indirect regulatory effect of CyPB on collagen hydroxylation , impacting collagen glycosylation , crosslinking and fibrillogenesis , which contribute to maintaining bone mechanical properties .
Type I collagen , the most abundant protein component of the extracellular matrix of skin , tendon and bone , is a heterotrimer consisting of two α1 ( I ) and one α2 ( I ) chains encoded by the COL1A1 and COL1A2 genes , respectively . The pro-alpha chains of type I collagen contain an uninterrupted helical region consisting of 338 repeats of the Gly-Xaa-Yaa triplet . Biosynthesis of procollagen is a complex process that requires several co- and post-translational modifications within the endoplasmic reticulum , including formation of disulfide bonds within the propeptide extensions , isomerization of peptidyl-prolyl bonds , hydroxylation of Yaa lysyl and prolyl residues , and glycosylation of hydroxylysines [1] . The post-translational modifications occur before , and to a major extent stabilize , collagen helical folding . After secretion into the pericellular space and processing of propeptide extensions , the mature collagen heterotrimer is incorporated into heterotypic collagen fibrils . The fibrils are then stabilized by intermolecular aldehyde-derived crosslinks formed from specific collagen lysyl and hydroxylysyl residues by lysyl oxidases ( LOX ) [2] , [3] . Dominant mutations in COL1A1 or COL1A2 cause classical osteogenesis imperfecta ( OI ) , with susceptibility to fractures from minimal trauma and growth deficiency [4] . Glycine substitutions in the collagen alpha chains delay folding and increase exposure to modifying enzymes , resulting in collagen overmodification . Some OI cases ( ≈10% ) have recessive inheritance , caused by deficiency of proteins that interact with collagen for folding or post-translational modification [5] . Most commonly , recessive OI involves the collagen prolyl 3-hydroxylation complex , consisting of prolyl 3-hydroxylase 1 ( P3H1 , encoded by LEPRE1 , leucine- and proline-enriched proteoglycan 1 ) , cartilage-associated protein ( CRTAP ) and cyclophilin B ( CyPB , encoded by PPIB , peptidyl-prolyl cis-trans isomerase B ) . This complex is responsible for 3-hydroxylation of the Xaa position α1 ( I ) P986 residue in types I and II collagen [6]–[8] . Loss of individual components abrogates collagen P986 3-hydroxylation [8]–[11] . CRTAP and P3H1 are mutually supportive in the complex; deficiency of either component causes severe to lethal OI with rhizomelia , classified as OI types VII and VIII , respectively [12] . However , loss of the CRTAP/P3H1 complex does not decrease the level of CyPB , and , conversely , the CRTAP/P3H1 complex is only partially decreased in the absence of CyPB . CyPB is an ER-localized member of the immunophilin family of proteins with peptidyl-prolyl cis-trans isomerase ( PPIase ) activity [13] , [14] . CyPB plays a key role as a member of several foldase and chaperone complexes , including BiP , GRP94 , PDI and calreticulin , for example , to facilitate folding of multiple substrates within the ER [15] , [16] . Cis-trans isomerization of peptidyl-prolyl bonds is especially important for type I collagen folding because prolines constitute approximately one-fifth of its primary sequence . Studies published over twenty years ago demonstrated that exposure of cells to the cyclophilin inhibitor cyclosporin A ( CsA ) slows the rate of collagen folding and results in overmodification of lysyl residues [17] . Thus , CyPB is thought to be the major , and possibly unique , collagen peptidyl-prolyl cis-trans isomerase [18] . However , collagen biochemical data from the few reported patients with CyPB deficiency is inconsistent . Two patients with PPIB mutations causing moderate OI have normal levels of collagen prolyl 3-hydroxylation [19] , [20] , while in two lethal OI cases of CyPB deficiency , α1 ( I ) P986 3-hydroxylation decreases to 30% of normal [11] , [20] . Overall collagen overmodification was detected in both lethal cases , but in only one moderately severe patient . In an attempt to clarify the inconsistencies among the human cases of PPIB/CyPB deficiency , we generated a Ppib knockout mouse model . We demonstrate a major , but not unique , role for CyPB in collagen folding and extend the previously reported role of CyPB in lysyl hydroxylation [21] . Biochemical studies reveal novel cell- and tissue-specific dysregulation of collagen helical lysyl hydroxylation and glycosylation in the absence of CyPB , independent of impaired collagen folding . Furthermore , reduced hydroxylation of specific collagen helical lysine residues led to a shift in the pattern of intermolecular crosslinks in bone tissue , and reduced collagen deposition into matrix . These studies establish novel functions for CyPB in regulating collagen biosynthesis and post-translational modification .
Ppib-null mice were produced from an ES cell line carrying a gene trap insertion in intron 1 of Ppib . Two ESC lines , RST059 and RST139 , were screened by real-time RT-PCR . Expression of Ppib in RST059 and RST139 was decreased to 76±5% and 54±3% , respectively , of wild-type levels ( Figure S1A ) . Because RST139 ES cells were more likely to have one null Ppib allele , we proceeded with generation of a mouse line in the C57BL/6 background using these cells . Genomic DNA from homozygous mutant mice was amplified by heminested PCR ( Figure 1A ) and sequenced to demonstrate insertion of the gene trap vector 123 bp from the 5′-end of the1066 bp Ppib intron 1 . The insertional mutation interrupts transcription of Ppib in cells and tissues isolated from Ppib−/− mice . Expression of Ppib in primary fibroblast ( FB ) and osteoblast ( OB ) cultures from newborn mice , as well as in dermal and femoral tissues of 8 week-old mice , was reduced in cells and tissues heterozygous for the gene-trapped allele , and completely absent from homozygous cells and tissues ( Figure 1B ) . Thus , Ppib expression is completely blocked in the gene trap mutant allele . Heterozygous mice were bred into the C57BL/6 line . In surviving homozygous offspring of F5 matings , growth deficiency became apparent soon after weaning . Knockout mice weighed about 25% less than wild-type and heterozygous littermates from 3 to 24 weeks of age ( Figure 1D ) . At 8 weeks of age , Ppib−/− femoral and tibial lengths were reduced 7% and 10% versus wild-type mice and heterozygotes ( p<0 . 00004 and p<0 . 02 , respectively ) ( Figure 1C ) . However , in contrast to Crtap and P3H1 null mice [8] , [22] , Ppib−/− mice do not have rhizomelia . The ratio of femoral to tibial length is comparable to wild-type ( 0 . 886±0 . 020 vs 0 . 861±0 . 035 , p = 0 . 07 ) . Skeletal abnormalities of Ppib−/− mice include decreased mineralization and abnormal shape of calvaria , shortened limbs and a deformed and flared rib cage ( Figure 2A ) ; these features are accentuated in lethal null pups . The genotype distribution among offspring of F4 and F5 het x het crosses deviated from the Mendelian ratio , with 30 and 50% lethality of homozygous pups from F4 and F5 matings , respectively , likely due to respiratory insufficiency from abnormal rib cage structure ( Figure 2A ) . Heterozygous mice are essentially normal . At 2 months of age , the rib cage of both heterozygous and homozygous mice has a narrow apex and drooping ribs at the base , providing limited space under the ribs for abdominal contents , which puff out the abdomen ( Figure 2B ) . Homozygous mice also have kyphosis . Adult knockout mice are osteoporotic; they have decreased aBMD of femora ( p = 0 . 001 ) and vertebrae ( p = 0 . 02 ) ( Figure 2C ) . Femora of 2-month male Ppib−/− mice display altered cortical and trabecular structure on μCT analysis ( Figure 3A; Table S1 ) . Their trabecular bone volume is half of wild-type , with reduction of both trabecular number and thickness . Femoral cortical bone is thinner , with decreased cortical area and a modestly enlarged marrow space . Femora were mechanically tested to assess fragility . Ppib−/− femora have reduced stiffness , yield load , and ultimate load , requiring 48% less total energy to fracture than wild-type controls ( Figure 3B ) . Femoral post-yield displacement and plastic energy were reduced 77% and 89% respectively , demonstrating a more brittle phenotype characteristic of OI . Importantly , while stiffness was reduced 37% in Ppib−/− mice ( p<0 . 01 ) , cortical bending moment of inertia , as measured by microCT , was only modestly reduced ( −11% , p = 0 . 14 ) suggesting there may be significant changes in bone material properties at levels unaccounted for by changes in cortical geometry . We examined the components of the collagen 3-hydroxylation complex in CyPB-deficient cells . In Ppib−/− OB but not FB , complete absence of CyPB was associated with a 40–55% reduction in P3H1 ( Figure 4A ) . CRTAP levels were not altered in either cell type . Although CyPB was decreased in heterozygous cells , P3H1 and CRTAP were unchanged . Despite the persistence of P3H1 and CRTAP in Ppib−/− cells , 3-hydroxylation of α1 ( I ) P986 residues was severely reduced . In Ppib−/− mice , only 5–11% of OB and FB and 1–2% of dermal and bone tissue collagen α1 ( I ) P986 residues were modified , in contrast to wild type and heterozygous cultures and tissues in which there is nearly complete collagen 3-hydroxylation ( Table 1 ) . The secondary prolyl 3-hydroxylation site was also undermodified; Ppib-null OB and femora have half the α2 ( I ) P707 3-hydroxylation of wild-type and heterozygous samples . Type I collagen alpha chains from Ppib−/− FB and OB have delayed and broadened gel migration , consistent with delayed folding and overmodification of lysyl residues ( Figure 4B ) . Collagen overmodification in the absence of a structural defect is expected to increase thermal stability [10] , [23] . However , type I collagen secreted by wild-type , heterozygous and homozygous null cells had equivalent melting temperatures ( Tm ) by differential scanning calorimetry ( Figure 4C ) . In agreement with Tm results , amino acid analysis of type I collagen secreted by Ppib−/− FB and OB revealed normal proportions of hydroxylated lysine and proline residues ( Table 1 ) . To resolve these apparent discrepancies , we utilized a direct intracellular collagen folding assay . Accumulation of intracellular protease-resistant collagen was slower in Ppib-null FB than in wild-type cells ( Figure 5A ) , with a greater percent of folded trimers in wild-type cells throughout the independent time course experiments . In Ppib−/− OB , the delay in type I collagen folding compared to wild-type was nearly double that in FB ( Figure 5B ) . Notably , addition of the inhibitor CsA to both fibroblasts and osteoblasts further delayed the rate of collagen folding in Ppib−/− as well as wild-type cells . These data suggest that , although CyPB is the primary PPIase involved in catalyzing folding of the type I collagen helix , other PPIases may also be capable of supporting this function . Furthermore , since collagen folds more slowly in Ppib−/− than wild-type cells both with and without CsA treatment , two PPIases and/or a CyPB-dependent protein may be involved in collagen folding . The apparent discrepancy between slower folding rate and normal total lysyl hydroxylation of collagen from Ppib−/− mice led to a more detailed analysis of collagen hydroxylation and subsequent glycosylation in null cells and tissues . Biochemical analysis of Ppib−/− skin tissue showed a substantial decrease in lysyl hydroxylation ( 18% of wild-type ) ( Figure 6A ) and glycosylation ( 40–50% of wild-type ) ( Figure 6B ) in collagen . A site-specific analysis of skin and fibroblast lysine modification and crosslink patterns will be presented elsewhere . Decreased modification of skin-derived type I collagen resulted in slightly faster gel migration ( Figure 6C ) . However , Ppib−/− bone collagen lysyl hydroxylation was slightly increased in heterozygous ( p = 0 . 0002 ) and homozygous ( p = 0 . 005 ) mice ( Figure 6D ) , with a significant increase in galactosylhydroxylysine ( p<0 . 001 ) in homozygous mice ( Figure 6E ) and subtly broadened electrophoretic mobility ( Figure 6F ) . In addition , Ppib−/− bone extracts show a substantial decrease in type V collagen alpha chains . Detailed characterization of lysine residues in collagen from OB cultures and bone tissue showed site-specific changes in lysine hydroxylation and glycosylation , and differences between cultured cells and tissues ( Table 2; Figure 7 ) . In Ppib−/− OB collagen , only α1 ( I ) K87 , α2 ( I ) K87 and K174 are underhydroxylated , with about 20% of K87 residues in Ppib−/− collagen unhydroxylated versus <1% of wild-type . However , hydroxylation of other helical lysines from Ppib−/− OB collagen was normal or subtly increased , i . e . α1 ( I ) K603 and K756 hydroxylation were increased 14% and 9% , respectively ( Table 2 ) . Bone tissue collagen from Ppib−/− mice has an even more striking increase in the proportion of lysine residues involved in crosslink formation that are unhydroxylated . Hydroxylation of α2 ( I ) and α1 ( I ) K87 residues is decreased 30–40% , respectively , and α2 ( I ) K933 hydroxylysine content is decreased 38% , compared to wild-type bone ( Table 2 ) . Decreased hydroxylation is also evident at α2 ( I ) K174 , but other residues do not have significant changes . Glycosylation patterns are also altered in Ppib−/− collagen . In wild-type OB collagen , a major helical cross-linking residue , α1 ( I ) K87 , has the highest proportion of diglycosylation , while α2 ( I ) K87 is mainly non-glycosylated ( Table 2 and Figure 7 ) , as we previously reported [24] , [25] . In Ppib−/− collagen , mono- and di-glycosylation of α1 ( I ) K87 is decreased in tissue and OB culture collagen , and α2 ( I ) K87 di-glycosylation is only slightly increased in OB culture collagen , despite slow helical folding ( Figure 7 ) . In OB and bone-derived Ppib−/− collagen , all other lysine residues assayed have substantial increases in glycosylation , corresponding with delayed intracellular collagen folding . We observed increased glycosylation at α1 ( I ) K174 , K564 , K603 , K756 and α2 ( I ) K219 in Ppib−/− OB collagen , and to a lesser extent in Ppib−/− bone-derived collagen , ( Table 2 ) , in agreement with the increased total lysyl hydroxylation and broad gel migration ( Figure 6D–F ) . Collagen helical lysine modifications are catalyzed primarily by lysyl hydroxylase 1 ( LH1 ) for hydroxylation , GLT25D1 , for galactosylation of hydroxylysine [26] , and LH3 , which harbors both lysyl hydroxylase and glucosyltransferase activities [24] , [27]–[29] . By real-time RT-PCR , we found a modest increase in transcript levels for all three enzymes in Ppib−/− FB cultures , but not in skin , OB or femoral tissue ( Figure 8A–C ) . Importantly , the protein levels of these enzymes were normal on western blots of FB and OB lysates ( Figure 8D ) . These data verify that expression levels of the modifying enzymes do not account for the alterations in lysine modification demonstrated in collagen from Ppib−/− cells and tissues . The combined effects of CyPB absence , slower collagen folding and abnormal modification , impact procollagen synthesis . Pulse-chase experiments show increased ( nearly double ) total collagen synthesis per cell by CyPB-deficient OB , while fibroblasts produce about half the amount of collagen per cell as wild-type cells ( Figure 9 , upper ) . However , the kinetics of collagen secretion is only slightly delayed for Ppib−/− FB and OB , ( Figure 9 , lower ) , but of uncertain physiological significance , since nearly the entire pulse of labeled collagen is secreted in the same timeframe as wild-type cells . This finding is similar to our previous report on FB with null mutations in LEPRE1 [10] , but distinct from the reported increased rate of collagen secretion in FB with null mutations in CRTAP [8] . To determine the consequences of altered collagen post-translational modification on extracellular matrix , we analyzed collagen deposition into matrix in culture . Sequential extraction of the incorporated labeled collagen revealed that collagen deposited into insoluble matrix by OB ( Figure 10A , left ) and FB ( Figure S1B ) was decreased 80% and 70% , respectively , compared to wild-type cultures . Raman spectroscopy of OB-derived matrix corroborated the biochemical analysis ( Figure 10A , right ) , indicating a two-thirds reduction in matrix collagen content compared to wild-type . Consistent with the marked reduction in hydroxylation of the helical crosslinking lysine in bone in the absence of CyPB , there was a significant difference in the crosslinking pattern . For divalent crosslinks , there was a nearly three-fold increase in the helical lysine-involved crosslink , hydroxylysinonorleucine ( HLNL ) , in homozygous humeral bone ( p = 1 . 9×10−8 ) , while the helical hydroxylysine-involved dihydroxylysinonorleucine ( DHLNL ) is comparable to wild-type bone ( Figure 10C ) . The increase in HLNL crosslinks decreases the DHLNL/HLNL ratio in Ppib−/− humeri ( p = 0 . 0004 ) . For the trivalent mature crosslinks , hydroxylysylpyridinoline ( HP ) crosslinks were unchanged in Ppib−/− bone relative to wild-type . However , the helical lysine-involved crosslink , lysylpyridinoline ( LP ) , was markedly increased by four to five-fold in Ppib−/− humerii ( p = 2 . 8×10−9 ) and femoral tissue ( p = 0 . 0001 ) , respectively ( Figure 10D ) . The resulting HP/LP ratio in CyPB-deficient bone is decreased 4 . 25-fold in humeral ( p = 0 . 00004 ) and 5 . 6-fold in femoral bone tissue ( p = 0 . 003 ) . It is also quite noteworthy , in the context of both the reduced collagen deposition into Ppib−/− osteoblast matrix and the increase in non-hydroxylated forms of crosslinks in tissue , that total bone collagen crosslinks were increased in null mice versus wild-type at 2 months of age ( DHLNL+HLNL , p = 0 . 002; HP+LP , p = 0 . 00004 and 0 . 003 for humerus and femora , respectively ) . The abnormal collagen modification and crosslinking affects the structure and organization of collagen fibrils . In dermal fibrils of Ppib−/− mice , we noted the presence of disorganized aggregate forms , as well as a 25% increase in the average fibril diameter ( p = 1 . 7×10−15 ) , and a broader distribution of fibril diameters ( p = 4 . 0×10−8 ) ( Figure 10B ) . In contrast to dermal fibrils , collagen fibrils from femoral tissue were less densely packed and had visibly decreased diameters compared to wild-type . Bone fibril diameters were not quantitated because multiple bundle orientations in each field does not allow a clear analysis of fibril cross-sections .
To investigate the role of CyPB/PPIB in post-translational modification and folding of type I collagen , we generated a CyPB knockout mouse using a Ppib gene-trapped ES cell line . We confirmed complete absence of Ppib transcripts in cultured cells , as well as dermal and bone tissue of homozygous null mice . Furthermore , CyPB was undetectable in immunoblots of both fibroblast and osteoblast cultures . The resulting murine phenotype reproduces the clinical findings in patients with PPIB deficiency , including growth deficiency with bone deformities , reduced bone mineral density , decreased bone volume and strength . Interestingly , the impairment of mechanical properties of Ppib−/− femora exceeded the extent expected from bone structural parameters , suggesting that bone material properties have also declined . This study recapitulates some features previously reported in a CyPB-deficient mouse , including a moderately severe skeletal phenotype with reduced bone geometry , absence of α1 ( I ) P986 3-hydroxylation and delayed collagen migration on gel electrophoresis . The prior study focused on the role of CyPB in the P3H complex , while we investigated CyPB's function in collagen synthesis , folding and helical modification . We found that collagen folding is delayed in Ppib−/− cells , and further folding delay by CsA inhibition provides support for an additional collagen PPIase . Furthermore we identified striking differences in the pattern of collagen post-translational modification in cells versus tissues . Unexpectedly , CyPB deficiency also results in tissue- and site-specific alterations of post-translational hydroxylation and glycosylation of type I collagen , which consequently affect fibril assembly , crosslinking in matrix , and bone mineralization . Thus our study offers novel mechanisms with the potential to explain the unique features of CyPB deficiency in humans , compared to the consequences of P3H1/CRTAP defects . In our Ppib KO mice , α1 ( I ) P986 3-hydroxylation is nearly absent from dermal and bone collagen , demonstrating that CyPB is required in vivo for 3-hydroxylation complex activity in tissues . This is consistent with the 3-hydroxylation status previously reported in murine Ppib-null bone and skin collagen , as well as with <1% modification of P986 residues in α1 ( I ) , α1 ( II ) and α2 ( V ) chains from bone , skin and cartilage of Crtap-null mice [8] , [30] , [31] . P986 3-hydroxylation of type I collagen was also absent in tail tendon , bone and skin of P3H1-null mice [22] , [32] . At the A2 site ( α2 ( I ) P707 ) , we observed a 50% reduction of 3-hydroxylation in collagen from skin and bone , suggesting that P3H1 or other P3H isoforms can modify this site without requiring CyPB for activity . For other sites and/or collagen alpha chains , the requirement for CyPB may be quite stringent . Delayed electrophoretic migration of collagen alpha chain , caused by increased post-translational modification , is considered to be a reliable surrogate assay for slower helix folding . Steady-state collagen from Ppib−/− cultured fibroblasts and osteoblasts has delayed gel mobility , consistent with the significant delay in helical folding of type I collagen demonstrated in direct intracellular folding assays in both cell types . Collagen folding in Ppib−/− fibroblasts and osteoblasts was delayed 5 to 8 minutes , respectively . By comparison , fibroblast cell lines from OI patients harboring typical structural defects in type I collagen demonstrated 5 to 60 minutes delay in formation of the collagen helix versus control cells , depending on the glycine substitution [33] . Treatment of control fibroblast and osteoblast cultures with cyclosporin A ( CsA ) , a non-specific cyclophilin inhibitor , resulted in a 10–12 minute delay in collagen folding , similar to published experiments [17] , [34] . Unexpectedly , CsA treatment of CyPB-deficient cells further increased the collagen folding delay by 6–8 minutes beyond that seen in untreated cells . These data suggest that another CsA-sensitive PPIase is capable of partial compensation for CyPB deficiency , and may be involved in collagen folding under normal conditions . Furthermore , even in CsA treated cells , folding remains slower in Ppib−/− than WT cells , suggesting that a CyPB-dependent , CsA-resistant function may also be supporting collagen folding . We previously reported a type IX OI patient with total absence of CyPB due to a start codon substitution [19] . Fibroblast collagen from this patient had normal helical modification suggestive of normal collagen folding , prompting us to propose redundancy of collagen cis-trans isomerases . The identity of this PPIase is open to speculation . Only one cyclophilin , but at least six FK506 binding proteins ( FKBPs ) , occur in the rER [13] , [14] . Until recently , FKBP65 , which binds gelatin and is partially inhibited by CsA in vitro , was an attractive candidate [35] . However , addition of FK506 to cell cultures delays collagen folding only slightly [34] , and absence of FKBP65 in patients with OI , Bruck syndrome and Kuskokwim disease has negligible effects on collagen helical folding [36]–[38] , limiting FKBP65 to at most a compensatory role in the absence of CyPB . Similarly , mutations in FKBP14 , which encodes FKBP22 , were identified in an Ehlers-Danlos like-syndrome similar to the kyphoscoliotic type of EDS ( EDS VIA ) caused by LH1 deficiency , but their types I , III and V collagens showed normal electrophoretic migration [39] . One of the striking findings of this study was the difference in post-translational modification in collagen isolated from cell cultures versus dermis and bone . Ppib−/− fibroblast and osteoblast steady-state type I collagen has significant electrophoretic delay , but normal total hydroxylysine content , which suggested that increased glycosylation delayed gel mobility . In contrast , collagen extracted from dermal tissue of CyPB-deficient mice has an 80% decrease in total helical hydroxylysine content , leading to more rapid gel migration of alpha chains than collagen from wild-type skin . The decreased hydroxylation of collagen in dermal tissue of the Ppib−/− mouse is similar to the dermal tissue data from American quarter horses with hyperelastosis cutis caused by a homozygous Ppib missense mutation that does not impair its isomerase function in vitro . The significantly decreased hydroxylysine content of dermal collagen in the horse is proposed to result from loss of a CyPB-LH1 interaction required for LH1 activity [21] . Type I collagen extracted from humeri of Ppib−/− mice had different post-translational modification than both osteoblasts and dermal tissue . Bone-derived collagen alpha chains demonstrated subtle broadening and backstreaking on PAGE analysis , consistent with the increased global hydroxylation and glycosylation of lysyl residues determined by amino acid analysis . However , detailed mass spectrometric analysis revealed site-specific alterations in lysine post-translational modification of bone tissue collagen . We found dramatic underhydroxylation of lysine residues involved in the formation of intermolecular crosslinks in tissue , specifically at α1 ( I ) K87 ( 57% Hyl in null vs 98% in WT ) , α2 ( I ) K87 ( 45% Hyl in null vs 77% in WT ) , and α2 ( I ) K933 ( 62% Hyl in null vs 100% in WT ) . Significant decreases in hydroxylation of these residues are also seen in CyPB-deficient osteoblast collagen . It is generally accepted that hydroxylation of helical lysine residues is catalyzed by LH1 . However , LH1 stability does not appear to be affected in Ppib−/− tissues based on western blot analysis ( Figure 8D ) . Whether CyPB contributes to LH1 folding , or stabilizes its activity by binding , are among the possibilities to be investigated . Thus absence of CyPB from bone has critical direct and indirect effects on collagen , affecting both collagen folding and the activity of a major collagen-modifying enzyme , respectively . This gives CyPB a pivotal role in determining the structure of secreted collagen . The site-specific changes in lysine hydroxylation seen in CyPB-deficient cells and tissues raises the possibility of an additional regulatory mechanism in bone , where there is apparent functional redundancy of LH activity directed at helical lysines not involved in crosslinking . Primary LH1 and LH2 deficiency demonstrated that LH1 hydroxylates lysines in the collagen triple helix while LH2 functions as the telopeptidyl lysyl hydroxylase [40]–[43] . However , collagen demonstrated variable decreases in Hyl content in bone from Ehlers-Danlos Type VIA patients ( 10–43% of normal ) and a Plod1 KO mouse ( 75% of wild-type ) , although specific helical lysines involved in crosslinks were always underhydroxylated based on decreased HP/LP crosslink ratios [44]–[46] . Thus , although LH1 appears to be required as the primary hydroxylase for helical domain crosslinking lysines in bone collagen , the relative roles of LH1 , LH2 and LH3 at other helical sites and in other tissues is less well understood . LH3 , which has LH , galactosyltransferase and glucosyltransferase activity in vitro and in culture [24] , [27] , [47] , [48] , could be the source of helical Hyl and even increased lysine glycosylation , in Ppib−/− bone [49] . Both a KO murine model for LH3 and a mouse with inactivated LH3 hydroxylation have faster gel migration of fibroblast type I collagen [50] . In the single human case of LH3 deficiency reported , the disaccharide derivative of pyridinoline crosslinks was absent in the patient's urine [28] . The intracellular hydroxylation of collagen helical K87 , K930/933 and telopeptidyl lysine residues determines the collagen crosslink pathway . Crosslinks between collagen molecules in extracellular matrix are crucial to skeletal function because they contribute to matrix stability , bone strength and ductility . In both dominant OI caused by primary collagen defects and recessive OI caused by CRTAP and P3H1 deficiency , the overhydroxylation of collagen helical lysines leads to increased divalent DHLNL/HLNL and trivalent HP/LP collagen crosslink ratios [51]–[55] . In contrast , in Ppib−/− mice the underhydroxylation of K87 residues in tissue collagen results in substantial decreases in DHLNL/HLNL and HP/LP ratios in bone . Notably , these changes are similar to findings in primary LH1 deficiency , in which bone-derived urinary peptides reflect decreased hydroxylation of collagen lysine residues involved in crosslink formation [40] , and decreased HP/LP ratios and increased total pyridinoline crosslinks were observed [40] , [56] , [57] . The cause of increased total crosslinks seen in Ppib−/− bone collagen is not clear . Possibly , decreased K87 hydroxylation and glycosylation may favor binding or activity of lysyl oxidase at this site [25] , [58] , [59] . Eyre and colleagues proposed that the collagen prolyl 3-hydroxylation modification supports matrix supramolecular assembly by fine-tuning the intermolecular alignment of collagen molecules to facilitate crosslink formation [55] . Our finding of severe reduction in collagen deposition into matrix is consistent with this hypothesis . However , we now understand that collagen secreted by CyPB-deficient osteoblasts has site-specific alterations in hydroxylation and glycosylation as well as absent P986 3-hydroxylation , which could also alter matrix assembly in addition to changes in collagen crosslinking and fibril structure . Furthermore , although this investigation focused on type I collagen , extracts from Ppib−/− bone show a substantial decrease in the quantity of type V collagen alpha chains ( Figure 6F ) . The effect of altered collagen crosslinking in Ppib−/− bone on mineralization remains to be explored . Bone from OI caused by collagen structural defects or CRTAP deficiency , whose crosslink pattern is opposite to this Ppib−/− mouse , is paradoxically hypermineralized . Collagen crosslinking and bone mineral crystallinity were strongly correlated in long bones of several congenic mouse strains [60] . On the other hand , increased HP/LP ratios correlate with the ultimate compressive strength of trabecular bone , but are independent of BMD [61] , [62] . Comparison of mineralization in Ppib−/− and classical OI bone will provide insight into the interaction of bone crosslinks and mineralization , and could point to collagen modification as an intracellular pathway by which osteoblasts can actively influence bone mineralization . Our initial goal in generating this mouse model of CyPB deficiency was to address the inconsistent findings of type I collagen lysyl and P986 hydroxylation among patients with type IX OI . Our analysis has revealed features of collagen biochemistry that have yet to be addressed in patients with type IX OI , in which all four sets of analyses were limited to collagen from fibroblast cultures . First , the extent and distribution of collagen helical modification in murine Ppib−/− samples differed between cells and tissues , with the overmodification of type I collagen from cultured cells due to increased glycosylation of a normal number of hydroxylysine resides . Future site-specific examination of tissue samples from type IX OI will be required to determine whether LH1 function is compromised and whether compensatory hydroxylation occurs . Second , type I collagen 3-hydroxylation differs between humans and mice with CyPB deficiency . In cultured cells of type IX OI patients , α1 ( I ) P986 3-hydroxylation is normal in two moderately severe cases and reduced to about 30% in two lethal infants , while tissues and cells from both murine models have totally abolished P986 3-hydroxylation . Thus , our data suggests that the redundancy for CyPB's role in collagen folding may not apply to its role in 3-hydroxylation of the A1 site of type I collagen in mice . Tissue studies of type IX OI probands will be critical to determine whether the role of CyPB in 3-hydroxylation by the CRTAP/P3H1 complex is fully rescued by redundancy in human cells with total absence of CyPB , but only partially rescued in the presence of truncated CyPB .
ES cell gene trap lines with a β-geo reporter construct inserted in intron 1 of the Ppib gene were produced by BayGenomics ( UCSF , CA ) and obtained from the Mutant Mouse Regional Resource Center ( Davis , CA ) [63] . The gene trap construct contains intron 1 and a portion of exon 2 to include splice acceptor sequence from the mouse En2 gene , followed by a β-galactosidase/neomycin ( β-geo ) reporter-selection cassette and SV40 polyadenylation signal . Cells were expanded for isolation of mRNA and quantitation of Ppib expression . A heterozygous clone from cell line RST139 was shown to have half-normal Ppib expression by real-time RT-PCR , and was injected into C57BL/6 blastocysts . Founders were generated by mating chimeric males with 129/P2/OlaHsd females to retain the 129 background of the ES cell line . A second line was generated for experiments by backcrossing F1 mice into C57BL/6 for 5 generations . Genomic DNA for genotyping was isolated using the Red Extract-n-Amp tissue PCR kit ( Sigma ) and amplified by hemi-nested PCR using a sense primer located in Ppib exon 1 ( 5′-TGCCCGGAC CCTCCGTGGCCAACGATAAGA-3′ ) , an antisense primer corresponding to intron 1 of En2 ( 5′-GGCATCTCCCCTTCAGTCTTCCTGTCCAGG-3′ ) , and an antisense primer downstream of the inserted construct internal to Ppib intron 1 ( 5′-GGGGGGCTGGGGGAGTCTGGGTTATTCTCT-3′ ) . Complete absence of Ppib transcripts was verified by real-time RT-PCR of mRNA isolated from femoral tissue and skin of F2 Ppib homozygous knock-out mice . Animal care and experiments were performed in accordance with a protocol approved by the NICHD ACUC committee . For skeletal staining , skin and viscera were removed from dead P1 newborn pups . Pups were fixed in 95% ethanol for 7 days and stained with 0 . 3% Alcian Blue 8GS and 0 . 1% Alizarin Red S [64] . Excess stain was removed with 1% KOH and increasing concentrations of glycerol . For growth curves , mice were fed regular rodent chow and weighed weekly from 3 to 24 weeks of age . Skeletal characteristics of F6 wild-type , heterozygous and homozygous Ppib-null littermates were analyzed at age 8 weeks . Radiographs were performed by Faxitron ( 30 kV for 1 min ) . Areal bone mineral density ( aBMD ) scans of mouse femurs were acquired using a GE Lunar PIXImus2 ( GE Healthcare ) and internal calibration standards . To determine femoral length , femora were dissected and cleaned of soft tissue , leaving the epiphyses intact . Femurs were measured from the proximal head to the distal end of the medial and lateral condyles with a digital caliper . Left femora of 8 week-old mice were analyzed by μQCT for both structural and mineral parameters using a SkyScan1174 compact micro-CT scanner ( MicroPhotonics ) operating at 50 kV with an X-ray source current of 800 µA , according to manufacturer directions . BMD was calibrated with hydroxyapatite phantoms . Morphometric analyses of trabecular and cortical regions was performed using CTAn software ( v . 1 . 13 ) and 3D images were generated using CTvol software ( v . 2 . 2 ) . The trabecular region of interest ( ROI ) was located just proximal to the distal femoral growth plate and extended 10% total femoral length . The diaphyseal cortical ROI centered on the femur midpoint , spanning 15% total femoral length . Femora were loaded to failure in four-point bending as previously described [65] . Femora were loaded to failure in four-point bending at 0 . 5 mm/s in the anterior-posterior direction with the posterior surface under tension using a servohydraulic testing device ( MTS 858; MiniBionix; MTS Systems Corporation ) . Force was recorded by a 50lb load cell ( Sensotec ) and vertical displacement by an external linear variable differential transducer ( Lucas Schavitts ) at 2000 Hz . Load-displacement curves were used to calculate stiffness , yield load , yield displacement , ultimate load , failure displacement , post-yield displacement , and energy to failure . Primary fibroblast ( FB ) and calvarial osteoblast ( OB ) cultures were derived from 3-day old pups by standard procedures [66] . Cells from digestions 3–5 were plated at a density of 5 , 000 cells/cm2 and cultured in αMEM with 10% FBS , 2 mM glutamine , 1% pen-strep and 8% CO2 . FB cultures were derived from dermal tissue dissected from the abdomens of newborn pups . FB were allowed to grow out from dermal samples for 2 weeks , released by trypsin digestion and cultured in DMEM containing 10% fetal bovine serum , 2 mM glutamine , 1% pen-strep and 5% CO2 . Total RNA was extracted from cell cultures or tissues dissected from F6 mice using TriReagent ( Molecular Research Center ) according to the manufacturer's protocol . Total RNA was treated with DNA-free ( Life Technologies ) , then reverse-transcribed using a High Capacity cDNA Archive Kit ( Life Technologies ) . Real-time RT-PCR was performed using Taqman Assays on Demand ( Life Technologies , Ppib , Mm00478295_m1; Plod1 , Mm01255760_m1; Plod3 , Mm00478798_m1; Glt25d1 , Mm00600638_m1; Glt25d2 , Mm01290012_m1; Gapdh , Mm99999915_g1 ) . Relative expression of genes of interest was measured in triplicate , normalized to Gapdh transcripts , and quantified relative to wild-type calvarial OB . Two independent fibroblast ( FB ) and osteoblast ( OB ) cultures for each genotype were lysed in RIPA buffer ( 150 mM NaCl , 1% NP-40 , 0 . 5% Na-deoxycholate , 0 . 1% SDS , 50 mM Tris , pH 7 . 4 ) supplemented with protease inhibitor cocktail ( Sigma ) . Protein concentration was determined using the BCA Protein Assay Kit ( Thermo Scientific ) . Samples ( 15 µg protein ) were subjected to SDS-PAGE on 10% acrylamide gels under denaturing conditions and electroblotted onto nitrocellulose membranes . The membranes were blocked overnight in 5% non-fat milk in TBST . After washing in TBST , membranes were incubated overnight at 4°C in TBST containing 2 . 5% non-fat milk and primary antibody ( diluted 1∶1000 ) . After washing in 1× TBST , membranes were incubated with the corresponding IRDye infrared secondary antibody ( diluted 1∶10 , 000 ) ( LI-COR Biosciences ) . Proteins were visualized using an Odyssey Infrared Imaging System ( LI-COR Biosciences ) . Quantitation of proteins was performed using the Odyssey 3 . 0 . 30 software with normalization to Actin levels . The analysis was repeated to ensure reproducibility . The primary antibodies used include anti-P3H1 ( Abnova ) , rabbit anti-actin ( Santa Cruz Biotechnology ) , rabbit anti-CyPB ( Abnova ) , rabbit anti-PLOD1 ( Santa Cruz Biotechnology ) , rabbit anti-PLOD3 ( Proteintech ) , and goat anti-GLT25D1 ( Santa Cruz Biotechnology ) . Anti-mouse CRTAP antibody was a generous gift from Dr Brendan Lee , Baylor College of Medicine . Steady-state collagen analysis was performed as previously described [67] . Collagens were prepared by pepsin digestion ( 50 µg/ml ) of procollagen samples , separated on 6% SDS-urea- polyacrylamide gels and visualized by autoradiography . For analysis of collagen modification in cell culture , confluent FB and OB were stimulated for collagen synthesis in DMEM or αMEM containing 0 . 1% FBS and 100 µg/ml ascorbate for three days , with daily collection . Collected medium was buffered with 100 mM Tris-HCl , pH 7 . 4 , and cooled to 4°C . Protease inhibitors were added to the following final concentrations: 25 mM EDTA , 0 . 02% NaN3 , 1 mM phenylmethylsulfonylfluoride , 5 mM benzamidine , and 10 mM N-ethylmaleimide . Procollagens were precipitated from media with ammonium sulfate overnight at 4°C . Procollagen was collected by centrifugation , resuspended in 0 . 5 M acetic acid and digested with 0 . 1 mg/ml pepsin at 4°C overnight . Selective salt precipitation of collagen with 0 . 9 M NaCl in 0 . 5 M acetic acid was performed twice . Purified collagen samples were resuspended in 0 . 5 M acetic acid , dialyzed against 5 mM acetic acid overnight and lyophilized before further analyses . Differential Scanning Calorimetry ( DSC ) scans were performed as previously described [68] . Thermograms were recorded in 0 . 2 M sodium phosphate , 0 . 5 M glycerol , pH 7 . 4 , from 10 to 50°C at 0 . 125 and 1°C/min heating rates in a Nano III DSC instrument ( Calorimetry Sciences Corporation ) . Analysis of tissue-derived collagens was performed using skin , femora and humeri from 2 month old mice ( n = 5 ) . After marrow was flushed with cold PBS , bone samples were pulverized in liquid N2 , demineralized with EDTA at 4°C for 2 weeks , and lyophilized . Skin dissected from the backs of mice was minced and lyophilized after removal of hair , fat and muscle . Two mg of the dried samples were then reduced with standardized NaB3H4 , hydrolyzed with 6N HCl and subjected to amino acid and cross-link analyses as previously described [69]–[71] . The extent of Lys hydroxylation in collagen was calculated as hydroxylysine ( Hyl ) /hydroxyproline ( Hyp ) ×300 ( i . e . ∼300 residues of Hyp/collagen ) . The cross-link precursor aldehydes and reducible cross-links were measured as their reduced forms and all cross-links were quantified as moles/mole of collagen . Collagen 3-hydroxylation was analyzed by in-gel tryptic digestion of SDS-PAGE-purified type I collagen alpha chains . Electrospray mass spectrometry was performed on the tryptic peptides using an LCQ Deca XP ion-trap mass spectrometer equipped with in-line liquid chromatography ( Thermo Finnigan ) using a C8 capillary column ( 300 µm×150 mm; Grace Vydac 208MS5 . 315 ) eluted at 4 . 5 µl per min . Site-specific modification of lysyl residues was determined by LC/MS/MS on a Waters Q-Tof Premier mass spectrometer coupled to a nanoACQUITY UPLC system ( Waters Corporation ) as reported [25] . Tryptic peptides containing Lys residues , their hydroxylated and/or glycosylated forms were identified from the LC/MS/MS analyses using manual interpretation of the MS/MS spectra . Relative quantitation of lysine ( Lys ) , hydroxylysine ( Hyl ) , galactosyl-Hyl ( G-Hyl ) and glucosylgalactosyl-Hyl ( GG-Hyl ) at a specific glycosylation site was performed by dividing the total ion abundance determined for each species by the sum of the ion abundances of all observed species containing that particular site . An intracellular collagen folding assay was performed as described [17] . Confluent cells were stimulated overnight in media with 10% FBS and 100 µg/ml ascorbic acid , and then incubated in serum free media containing 100 µg/ml ascorbic acid for 2 hr . Cells were pulsed with 1 . 4 µCi/ml 14C-proline for 15 min to label procollagen chains , followed by collection of the cell layer every 5 min . Each sample was digested for 2 min at 20°C with 0 . 2% Triton X-100 , 100 µg/ml trypsin , and 250 µg/ml chymotrypsin in PBS ( Sigma ) . Digestions were stopped by addition of 1 mg/ml soybean trypsin inhibitor ( Sigma ) . Samples were precipitated overnight , collected by centrifugation , electrophoresed on 3–8% Tris-acetate gels ( Life Technologies ) and quantitated by densitometry of autoradiograms . Assays were performed in duplicate on two independent cultures for each genotype . For pulse-chase assays , performed as described [72] , wild-type and homozygous Ppib-null FB were grown to confluence . For each cell line , two wells were used for cell counts . Procollagens were harvested at the indicated times , digested with pepsin and precipitated . Samples were loaded for equivalent cell number on 3–8% Tris-acetate gels . Collagen alpha chains were quantitated by densitometry and expressed as the percent secreted at each time point , as determined by ( media ) / ( cell + media ) ×100 . Wild-type and homozygous Ppib-null FB and OB were grown to confluence and stimulated every other day for 14 days with fresh DMEM ( fibroblasts ) or αMEM ( osteoblasts ) containing 10% FBS and 100 µg/ml ascorbic acid , as described [73] . Matrix collagens were sequentially extracted at 4°C , with neutral salt for newly incorporated collagen , then acetic acid for collagens with acid-labile cross-links , and , finally , by pepsin digestion for collagens with mature cross-links [74] . All fractions were electrophoresed on 6% polyacrylamide-urea-SDS gels . Samples were loaded for equivalent densitometry signal; the total signal for each fraction was calculated by adjusting the gel signal by the total volume of that fraction . In separate experiments , quantitation of matrix deposited in culture was performed using Raman microspectroscopy . Cultures were fixed in 1% paraformaldehyde and analyzed as previously described [36] . Matrix collagen:cell organics ratios were evaluated from decomposition of corrected spectra of collagen-free cytoplasm and purified collagen in the amide III spectral region . A dermal biopsy was obtained from abdominal skin of wild-type and homozygous null mice , then processed as described [75] . Representative areas of the stained grids were photographed in a Zeiss EM10 CA transmission electron microscope ( JFE Enterprises ) .
|
Osteogenesis imperfecta ( OI ) , or brittle bone disease , is characterized by susceptibility to fractures from minimal trauma and growth deficiency . Deficiency of components of the collagen prolyl 3-hydroxylation complex , CRTAP , P3H1 and CyPB , cause recessive types VII , VIII and IX OI , respectively . We have previously shown that mutual protection within the endoplasmic reticulum accounts for the overlapping severe phenotype of patients with CRTAP and P3H1 mutations . However , the bone dysplasia in patients with CyPB deficiency is distinct in terms of phenotype and type I collagen biochemistry . Using a knock-out mouse model of type IX OI , we have demonstrated that CyPB is the major , although not unique , peptidyl prolyl cis-trans isomerase that catalyzes the rate-limiting step in collagen folding . CyPB is also required for activity of the collagen prolyl 3-hydroxylation complex; collagen α1 ( I ) P986 modification is lost in the absence of CyPB . Unexpectedly , CyPB was found to also influence collagen helical lysyl hydroxylation in a tissue- , cell- and residue-specific manner . Thus CyPB facilitates collagen folding directly , but also indirectly regulates collagen hydroxylation , glycosylation , crosslinking and fibrillogenesis through its interactions with other collagen modifying enzymes in the endoplasmic reticulum .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biochemistry",
"medicine",
"and",
"health",
"sciences",
"model",
"organisms",
"genetics",
"biology",
"and",
"life",
"sciences",
"research",
"and",
"analysis",
"methods",
"clinical",
"genetics"
] |
2014
|
Abnormal Type I Collagen Post-translational Modification and Crosslinking in a Cyclophilin B KO Mouse Model of Recessive Osteogenesis Imperfecta
|
The human gut microbiota comprise a complex and dynamic ecosystem that profoundly affects host development and physiology . Standard approaches for analyzing time-series data of the microbiota involve computation of measures of ecological community diversity at each time-point , or measures of dissimilarity between pairs of time-points . Although these approaches , which treat data as static snapshots of microbial communities , can identify shifts in overall community structure , they fail to capture the dynamic properties of individual members of the microbiota and their contributions to the underlying time-varying behavior of host ecosystems . To address the limitations of current methods , we present a computational framework that uses continuous-time dynamical models coupled with Bayesian dimensionality adaptation methods to identify time-dependent signatures of individual microbial taxa within a host as well as across multiple hosts . We apply our framework to a publicly available dataset of 16S rRNA gene sequences from stool samples collected over ten months from multiple human subjects , each of whom received repeated courses of oral antibiotics . Using new diversity measures enabled by our framework , we discover groups of both phylogenetically close and distant bacterial taxa that exhibit consensus responses to antibiotic exposure across multiple human subjects . These consensus responses reveal a timeline for equilibration of sub-communities of micro-organisms with distinct physiologies , yielding insights into the successive changes that occur in microbial populations in the human gut after antibiotic treatments . Additionally , our framework leverages microbial signatures shared among human subjects to automatically design optimal experiments to interrogate dynamic properties of the microbiota in new studies . Overall , our approach provides a powerful , general-purpose framework for understanding the dynamic behaviors of complex microbial ecosystems , which we believe will prove instrumental for future studies in this field .
The human gut harbors a dense and complex microbial ecosystem . Our ability to extensively characterize the microbiota has greatly increased in the last several years , due to lower costs and technical improvements in both DNA sequencing [1] and bioinformatics tools [2] , [3] . High-throughput sequencing-based studies of the microbiota generally analyze regions of the conserved 16S ribosomal subunit gene [4] , or use shotgun sequencing to sample the entire repertoire of genes present in a complex , mixed population of microbes [5]–[8] . These studies , of either human subjects or animal models , have uncovered intriguing associations between the composition of the gut microbiota and various diseases , including obesity [8] , inflammatory bowel disease [7] , [9] , and multiple sclerosis [10] . Longitudinal studies of the microbiota are particularly important for further advancing the field [3] , [5] , [6] , [11]–[13] . The majority of such longitudinal studies have been observational , monitoring the composition of the flora in healthy individuals over time at various body sites [3] , [5] , [6] , [11] . Such observational studies are valuable for understanding natural variations in commensal communities , as well as capturing rarer events such as onset and resolution of acute disease in the host . Additionally , interventional studies have been performed , in which subjects were intentionally exposed to agents that perturb the microflora , with subsequent evaluation of changes in host ecosystems over time [12] , [13] . Such interventional studies hold promise for discovering the mechanisms by which microbes interact with one another and the host , and to define how sub-communities of micro-organisms may cause or protect against disease . To date , longitudinal studies of the microbiota have largely employed static analysis techniques that do not adequately capture the dynamic nature of the data . The most common types of analyses employed involve either computation of diversity measures , such as the Shannon-Weaver diversity index [14] , at each data point , or measures such as Unifrac [15] or Bray-Curtis dissimilarity [16] , which characterize pair-wise relationships between data points . These techniques have proven useful for uncovering certain trends in longitudinal data [3] , [6] , [8] , [11] , [12] . However , these techniques treat longitudinal data as a collection of static snapshots , and ignore inherent ordering and other temporal dependencies . We developed a probabilistic model and inference algorithm , called Microbial Counts Trajectories Infinite Mixture Model Engine ( MC-TIMME ) , which provides a unified framework for analyzing the dynamic behavior of the microbiota captured via high-throughput sequencing data . Our framework models time-varying counts of microbial taxausing exponential relaxation processes . Each relaxation process is characterized by a transient effect level ( the amplitude of the process immediately after a perturbation ) , an equilibrium level ( the amplitude of the process approached as time tends to infinity ) , and a relaxation time constant ( the time required for the amplitude of the process to reach 33% of the transient effect level plus 67% of the equilibrium level , as measured on a logarithmic scale ) . Observed abundances of taxa are assumed to arise from an infinite mixture of prototype signatures . Each prototype signature is composed of a set of relaxation processes that model a response to multiple perturbations . Each reference operational taxonomic unit ( refOTU ) in the ecosystem ( s ) analyzed is probabilistically assigned to a prototype signature . Adaptive Bayesian techniques are used to model the dimensionality of prototype signatures and the extent of sharing of prototype signatures among refOTUs within and across ecosystem ( s ) . The time-series of observed counts for a particular refOTU in an ecosystem is modeled through a generative process , in which the prototype signature to which the refOTU has been assigned is customized to an individual signature by addition of refOTU and time-point specific offset terms . Data is then generated from a discrete-valued noise model parameterized by the individual signature . MC-TIMME enables several new types of analysis . The first type , termed Signature Diversity ( SD ) analysis , measures the variety of time-dependent microbial responses to perturbations to the host ecosystem ( s ) . SD utilizes time-varying information , and is thus distinct from traditional static measures of ecological diversity [17] , which characterize the complexity of a microbial community in terms of its constituent members at a single point in time . The second type , Relaxation Time Distribution ( RTD ) analysis , estimates the distribution of times required for members of the host ecosystem ( s ) to equilibrate after a perturbation event . This analysis summarizes the kinetics of responses , which is useful for understanding the phasing of changes within microbial ecosystems . The third type , Consensus Signature Group ( CSG ) analysis , identifies sub-communities of microbes within the larger ecosystem ( s ) that exhibit coordinated responses to a set of perturbations . This latter analysis provides information about which microbes may form functional sub-populations that affect the host or other microbial populations over time . Finally , MC-TIMME enables automated design of longitudinal studies of the microbiota . Our method couples information theoretic and Bayesian approaches to estimate from prior data the optimal sets of time-points to be sampled in future experiments . In this manner , data from a pilot study can be leveraged to develop optimized experimental designs for larger studies . MC-TIMME has some similarities to previously published methods for analyzing other types of high-throughput data . Several studies have employed continuous-time models [18]–[21] or infinite mixture models [22] , [23]to analyze time-series microarray data . Methods for optimal experimental design for time-series microarray data have also been described [24] . However , methods designed for analyzing gene expression data do not model dynamics inherent in complex microbial ecosystems , such as equilibrium reverting behavior . Further , these methods generally assume observed data are continuously valued and normally distributed , which is reasonable for microarray data , but not sequencing data , which consist of counts . Extensive statistical literature has documented that , for data consisting of discrete counts , direct modeling of the data yields superior results as compared to transforming data to continuous values ( see e . g . , [25]–[27] ) . This issue has been recognized for RNA Seq data , and several methods have been developed that use discrete-valued noise models [28]–[30] . Recently , Holmes et al . [31]presented a method for modeling microbial metagenomics counts data that also uses discrete-valued distributions . However , in contrast to MC-TIMME , these methods for analyzing RNA Seqor metagenomics count data use only static models . To gain new understanding of dynamic changes in the human gut microbiota caused by antibiotic exposure , we applied MC-TIMME to data from a longitudinal study by Dethlefsen et al . [12] . Despite the profound effects antibacterial agents presumably have on commensal species in vivo , remarkably little is known about the rates at which these complex ecosystems recover or remain altered after cessation of a course of antibiotics . Further , it remains poorly understood how antibiotic-induced changes in the microbiota affect underlying host physiology , including enhanced susceptibilities to other pathogens [32] , disease states such as allergic and auto-immune responses [33] , [34] , and acute or chronic effects on host diet and metabolism [35] . To date , the Dethlefsen et al . study provides the longest time-series systematically monitoring the effects of antibiotics on human gut commensals . In this study , human subjects were given two spaced five day courses of oral ciprofloxacin , a broad-spectrum antibiotic that targets the DNA gyrase and topoisomerases of many microbial species [36]–[40] . Subjects' gut microbiota was monitored at 50+ time-points over nearly a year , by sequencing 16S rRNA gene signatures present in stool samples . The remainder of the manuscript is organized as follows . First , we provide additional background on the Dethlefsen et al . dataset that we re-analyzed . Second , we describe the MC-TIMME framework , including our model of dynamics , inference algorithm , and automated experimental design method . Third , we apply MC-TIMME to the Dethlefsen et al . data to demonstrate the utility of Signature Diversity ( SD ) , Relaxation Time Distribution ( RTD ) , and Consensus Signature Group ( CSG ) analyses , and experimental design methods . Our results provide new quantitative insights into how antibiotic exposures affect the human gut microbiota , and how these dynamic alterations may influence the host .
Dethlefsen et al . [12] examined the microbiota from stool samples of three human subjects over a 10 month period , using Roche 454 high-throughput sequencing of PCR amplicons spanning the V1 , V2 and V3 regions of 16S rRNA genes , producing a total of approximately 5 million reads . During the study , each subject received two separate 5 day courses of oral ciprofloxacin . Stool samples were collected daily during , one week prior , and one week after the antibiotics courses , but otherwise were collected less regularly throughout the study . Dethlefsen et al . combined the sequencing reads from all subjects to produce 2 , 582 reference operational taxonomic units ( refOTUs ) . However , they found that only a few hundred refOTUs were present at consistently detectable levels in at least one sample per subject . In order to focus our analysis on refOTUs above the threshold of detection of the sequencing assay , we required that refOTUs have ≥5 counts for ≥10 time-points . This resulted in a total of 218 refOTUs for subject D , 261 for subject E , and 277 for subject F . Dethlefsen et al . taxonomically labeled refOTUs using the Silva 100 Small Subunit Reference database [41] , and UClust [42] with a minimum best hit similarity of 95% . Data files containing counts and taxonomic labels for refOTUs were downloaded from the website linked to the original publication; DNA sequences for each refOTU were not available . Our approach is based on a Bayesian information theoretic formulation of the experimental design problem ( see e . g . , [46]–[48] ) . To define notation , suppose we are given a joint probability distribution p ( Θ , A ( T ) ) over model parameters Θ and possible data A ( T ) collected at a set of time-points T . Suppose we then perform experiments , which allow us to collect a dataset denoted a . This results in a gain in Shannon information that is given by: Here , H{•} denotes the differential entropy . Thus , we see that the gain in information is due to the difference in entropy between prior and posterior distributions . The objective of our automated experimental design algorithm is then to choose the sampling times T that maximize the expected information gain over all possible data that could be observed at those time-points: This is a high dimensional integral that is in general intractable . However , for a linear model with Gaussian noise , the integral can be written as [46] , [49] , [50]: Here , IM denotes the Fisher information matrix , and |•| the determinant of the matrix . The integral can be approximated with a function g ( • ) of samples Θ ( j ) from the priorp ( Θ ) , yielding: In the case of a Generalized Linear Model , a linear approximation can be calculated to yield a local Bayesian D-optimality measure [50] . We use this measure , as each prototype signature in MC-TIMME is a Generalized Linear Model if we condition on the appropriate parameters ( see Protocol S1 ) . We estimate samples from p ( Θ ) , the prior probability distribution over model parameters for future experiments , using a model learned from previously observed data . Specifically , we use 500 MCMC samples obtained from the posterior distribution of the MC-TIMME model conditioned on a set of observed data . We then use a greedy optimization algorithm with the Bayesian D-optimality function g ( • ) defined above , to generate experimental designs . See Protocol S1 for complete details .
MC-TIMME analyzed the complete Dethlefsen et al . dataset , consisting of 3 subjects with 50+ time-points each , in approximately 12 hours on an Intel Xeon E5507 2 . 27 GHz core . Figure S2 provides examples of individual signatures for refOTUs inferred by MC-TIMME . The subsequent sections detail our biological findings based on these analyses . We also ran additional analyses to evaluate the sensitivity of our results to key model assumptions or features of the data . First , we tested the robustness of the model's dimensionality adaptation capability , which is a critical component of Signature Diversity scores . These tests showed no significant differences in our results when relevant model parameters were varied . Second , we tested the robustness of our results to noise . Because an equivalent gold standard experimental dataset does not exist , we generated simulated data for use in testing . For these simulations , we used all prototype signatures estimated by MC-TIMME from the full Dethlefsen et al . dataset as our gold standard , and then generated test datasets with varying amounts of added noise . When the amount of noise equaled that in the original dataset ( coefficient of variation of ≈60% for counts ) , MC-TIMME recovered Signature Diversity scores with <≈10% error , and relaxation time constant estimates with ≈25% error for the first post-antibiotic exposure interval , and ≈40% error for the second interval; measures of consistency of assignment of refOTUs to prototype signatures showed ≈20% reduction in quality . Third and finally , we tested the sensitivity of our results to exclusion of each experimental subject . These tests showed error rates comparable to those from our simulations when noise levels were equal to those in the original dataset . See Protocol S1 for complete details . Overall , our model performance tests demonstrate that our results are robust to changes in relevant parameter settings , noise , and exclusion of experimental subjects . To characterize the diversity of responses of the microbiota to repeated antibiotic treatments , we calculated three types of Signature Diversity ( SD ) scores . As shown in Figure 2 , each SD score ( SD1 to SD3 ) measures the dynamic behavior of micro-organisms in the host ecosystem ( s ) at a different level of resolution . These scores take into account the responses over time for refOTUs , and thus provide new information about dynamic properties of the ecosystems studied , as compared to traditional static measures of ecological diversity [17] . The intra-signature diversity ( SD1 ) scores for all three subjects in the Dethlefsen et al . study were >≈50% ( Figures 3A , 3B ) , indicating that the majority of micro-organisms in these host ecosystems exhibited changes in equilibrium levels or relaxation times after one or both antibiotic treatments . As shown in Figures 3A and 3B , the SD1 score has two components: ( 1 ) SD1μ , which measures the expected fraction of refOTUs with changes in equilibrium levels between pre-treatment and at least one post-antibiotic treatment interval , and ( 2 ) SD1λ , which measures the expected fraction of refOTUs with changes in relaxation time constants between the two antibiotic treatment intervals . The intra-ecosystem signature diversity ( SD2 ) score was ≈8–20 expected equivalent signatures per 100 refOTUs ( Figure 3C ) , indicating that many micro-organisms within each host ecosystem exhibited similar responses to the antibiotic treatments . Stated another way , a typical refOTU in a subject's gut ecosystems shared a similar response with ≈5–12 other refOTUs . These analyses indicated that subject E's gut microbiota exhibited fewer long-term shifts in abundance levels and responded overall more uniformly to the antibiotic exposures . That is , subject E had significantly lower intra-signature and intra-ecosystem Signature Diversity scores , with an SD1μscore of 50% and SD2 score of 10 , as compared with the other two subjects with SD1μ scores >70% and SD2 scores ≈20 . This differential behavior of subject E's microbiota was not discernible in the original analysis performed by Dethlefsen et al . , as they did not use techniques that quantified diversity of temporal responses . Our Signature Diversity analysis thus provides additional information about the functional diversity of subject E's microbiota , suggesting that this subject may have harbored a more ciprofloxacin-resistant flora prior to the experiments . Of note , subjects in the study had not received antibiotics in the past year before the experiments , but their antibiotic exposure history prior to this point was unknown . The inter-ecosystem signature diversity ( SD3 ) score for the 3 subjects was 48% ( p-value<10−6using a permutation test with null hypothesis of independent ecosystems ) , indicating that there were substantial similarities in the time-dependent responses of the subjects' microbiota to the antibiotic treatments . As shown in Figure 3D , the SD3 score is a ratio of two SD2 scores: ( 1 ) SD2D , which is computed on a hypothetical combined ecosystem , in which refOTUs from different subjects probabilistically share prototype signatures , and ( 2 ) SD2I , which is a weighted average of independent SD2 scores computed separately on each subject . The SD3 score of 48% indicates that approximately as many prototype signatures were shared among the subjects as were unique to each subject . Thus , although subjects' microbiota did exhibit varied responses to the antibiotic treatments , as reported in the Dethlefsen et al . study , our findings indicate that there were substantial commonalities among responses . These commonalities could not have been found using the analysis techniques of the original study , which relied on calculations at individual time-points , in part because different sampling times for each subject made point-wise comparisons impossible . In contrast , MC-TIMME uses a continuous-time model of dynamics that leverages information from multiple time-points to estimate key dynamical properties , allowing comparisons across subjects on a common time-scale . We generated Relaxation Time Distribution ( RTD ) plots using data from all three subjects , to investigate common trends in the rates at which the microbiota attained equilibrium levels after repeated antibiotic exposures ( Figure 4 ) . These plots depict estimated smoothed probability distributions of relaxation time constants , in units of days , for all refOTUs across all subjects . As shown in Figure 4 , the Relaxation Time Distribution for the first post-antibiotic exposure interval is multi-modal . A large peak in the distribution at ≈1–3 days indicates that many refOTUs quickly reached equilibrium levels , while a subsequent broader peak suggests waves of microbial succession events among subpopulations that took longer to equilibrate . Interestingly , after the second antibiotic exposure , the relaxation time distribution became simpler and more concentrated , with more refOTUs exhibiting relaxation times around ≈1–3 days . This finding suggests that the first antibiotic treatment shifted gut ecosystems toward more rapidly equilibrating states , possibly by selecting for more antibiotic resistant organisms or for sub-communities that more quickly and stably established themselves in relevant niches . To further our understanding of the differential responses of microbial sub-communities to antibiotic exposures , we next generated Consensus Signature Groups ( CSGs ) , which represent groups of refOTUs that consistently covary in terms of relative abundances over time . Combining data from all subjects , MC-TIMME identified 125 CSGs . Interestingly , many of the CSGs contained refOTUs that are phylogenetically related or are common to all subjects . To assess the phylogentic relationships among refOTUs within each CSG , we calculated an enrichment p-value for taxonomic labels at the order , family and genus levels , based on a hypergeometric distribution null hypothesis . Approximately 61% of refOTUs belonged to CSGs significantly enriched for at least one taxonomic label at these levels , with most such CSGs shared across subjects . These results provide evidence that MC-TIMME detected biologically relevant sub-communities of organisms based only on evaluation of time-varying behaviors of refOTUs . MC-TIMME additionally discovered refOTUs that exhibited consistent behavior across all subjects: 88refOTUs were present in all 3 subjects , and of these 88 refOTUs , ≈25% were assigned to common CSGs . These sets of refOTUs , which consistently covary across all subjects , could serve as candidate biomarkers in future studies of antibiotic treatments or other perturbations to the gut flora . We created a time-line of the largest and best taxonomically defined Consensus Signature Groups ( Figure 5A–M ) , to gain insight into the specific responses and successive equilibration of gut commensal sub-populations after antibiotic exposures . For purposes of visualization , we ordered the CSGs according to their relaxation time constants after the first antibiotic pulse , as these relaxation times exhibited the most variation . To facilitate interpretation , we included only CSGs containing at least one refOTU shared among all subjects , and significantly enriched ( false discovery rate<0 . 05 , hypergeometric tests ) for at least one taxonomic label at the family or genus level . These criteria yielded 13 CSGs , containing ≈50% of all refOTUs . Of the 13 CSGs , 8 exhibited decreases in relative abundance during the first antibiotic pulse ( Figure 5A–H ) , and 5 respectively exhibited increases ( Figure 5I–M ) . Among Consensus Signature Groups showing decreases in relative abundance during the first antibiotic pulse , those containing refOTUs in the genus Bacteroides ( Figure 5A , C ) , or refOTUs in the related family Porphyromonadaceae ( Figure 5D ) , were among the first groups to equilibrate after cessation of antibiotics . These groups of refOTUs had relaxation times <≈2 weeks , and returned to the same or higher relative abundances as compared to those prior to the first antibiotic pulse . The Bacteroides [51]are obligate anaerobes . Members of this genus , in particular B . fragilis , are known to have greater resistance to ciprofloxacin , mediated in part by several bacterial factors , coupled with reduced activity of the antibiotic under anaerobic conditions [40] , [52] . The Bacteroides can also be opportunistic pathogens and are capable of developing resistance to multiple classes of antibiotics after repeated exposures [53] , [54] . These characteristics may explain why the Bacteroides were among the first genera found by MC-TIMME to recover post-antibiotic treatment . Interestingly , another CSG significantly enriched for refOTUs of genus Bacteroides ( Figure 5M ) exhibited an increase in relative abundance during antibiotic treatment , with a very slow return to equilibrium levels ( relaxation time ≈42 days ) . This finding suggests that subjects consistently harbored antibiotic resistant Bacteroides , echoing concerns that members of this genus could serve as reservoirs of resistance genes for more frankly pathogenic bacteria [53] , [54] . MC-TIMME also identified another quickly equilibrating sub-community that contained refOTUs belonging to acetate [55] and butyrate [56] producing genera ( Figure 5I ) . This sub-community showed increases in relative abundance during antibiotic treatments and was significantly enriched for refOTUs belonging to the genera Blautia , Faecalibacterium , or Roseburia . Many Blautia species are acetogens , producing acetate from H2 and CO2 through the acetyl-CoA pathway [55] . Acetate has known downstream effects on the microbial production of butyrate [57] . Butyrate , a 4-carbon short chain fatty acid , has important roles in maintaining colonic health of the host , providing a luminal source of energy to the epithelial barrier [57] , while limiting the degree of autophagy in host colonocytes and reducing the host's susceptibility to agents that might otherwise promote damage to the colonic mucosa [58] . Members of the genera Faecalibacterium and Roseburia are prominent butyrate producers in the human gut [56] . Our CSG analysis suggests that the identified Blautia , Faecalibacterium , and Roseburia refOTUs may operate as a functional multi-species community in the gut , one demonstrating relative resilience tociprofloxacin's effects . This finding highlights MC-TIMME's ability to identify and track over time and across multiple subjects , bacterial sub-communities with potential benefits to the host . Interestingly , MC-TIMME discovered a second CSG containing Blautia refOTUs , but not the butyrate producing genera ( Figure 5B ) , and that exhibited a different response pattern , with a decrease in relative abundance during the first antibiotic pulse , and relatively rapid return to pre-antibiotic relative abundances . Of note , certain Blautia species , such as B . hydrogenotrophica , use a broader range of substrates for acetogenesis than other species in the genus [55] . The presence of such Blautia species with greater metabolic flexibility in the CSG depicted in Figure 5B could explain the lack of butyrate producers in this consensus signature group . Several Consensus Signature Groups contained refOTUs belonging to the family Ruminococcaceae ( Figure 5F , G , H ) . These CSGs showed decreases in relative abundance during the first antibiotic pulse and equilibrated slowly thereafter . In fact , one group of these organisms ( Figure 5H ) , become undetectable after the first antibiotic pulse , and another group ( Figure 5F ) declined to very low relative abundance levels after the second pulse . The Ruminococcaceae are overall obligately anaerobic , fastidious organisms that may require substrates produced as by-products of metabolism by earlier colonizers in gut luminal food-webs [57]–[59] . Thus , the delay in which these consensus groups of Ruminococcaceae recovered may be due to high degrees of dependence on activities of other organisms in the ecosystem . MC-TIMME also discovered a number of Consensus Signature Groups containing refOTUs from the family Lachnospiraceae ( Figure 5E , J , K , L ) . The majority of these CSGs showed increases during the antibiotic pulses , with fairly long relaxation times to pre-antibiotic relative abundance levels . The Lachnospiraceae are a large family of difficult to cultivate organisms , some of which have been found in close association with the mucous layer over the distal colonic epithelium [60] . Although little is known about the antibiotic susceptibilities of these organisms , it has been hypothesized that they may have evolved special mechanisms to survive the higher concentrations of endogenously produced host anti-microbial peptides present in this niche [60] . Our CSG analysis identified distinct groups of refOTUs from the Lachnospiraceae family that exhibited prolonged increases in relative abundance after ciprofloxacin exposure . These findings provide new information about this poorly understood bacterial family , which could be used to guide future studies to evaluate potential mechanisms of innate antibiotic resistance among Lachnospiraceae sub-communities . Application of metagenomic techniques to diagnostic medicine will require human clinical trials across many subjects to ascertain time-dependent effects and responses of the microbiota relative to a defined perturbation or clinical course of disease . The complicated logistics and expense of such trials highlight the need for computational techniques to optimize sampling across subjects . We developed an algorithm for automated experimental design , and applied it to the Dethlefsen et al . dataset to explore how the experimental design of a longitudinal study of the microbiota could be improved for future , larger trials . Our algorithm uses the data from a set of previously performed experiments to estimate an initial model of prototype signatures . This initial model is used to find a set of time-points that maximize the information that can be gained from future , hypothetical experiments . In general , our algorithm prioritizes time-points for sampling in future experiments around time-points in the original experiment that had the highest degree of uncertainty in the model . The selected time-points may differ from those in the original experiments , and thus indicate when increased sampling could better estimate dynamics of the ecosystems under study , or when reduced sampling still yields sufficient information . Figure 6 depicts the optimized experimental design produced by our algorithm for each subject . To facilitate comparisons among designs , we restricted our algorithm to choose the same numbers of time points as were used in the original design . Our algorithm generated an optimal design that consistently differed from the original design , over certain temporal intervals . On the pre-antibiotic interval , the optimized design required more uniform sampling , reflecting the modeling assumption that host ecosystems were at equilibrium prior to antibiotic exposure . For the period immediately after the first antibiotic exposure , the optimized design required additional frequent sampling beyond the one week of the original design . This increased sampling requirement is consistent with the relaxation time distribution on the first post-antibiotic treatment interval shown in Figure 4 , which indicates the presence of transient effects beyond one week for many refOTUs . However , the optimal design required fewer subsequent samples on the first post-antibiotic treatment interval , suggesting that MC-TIMME can effectively leverage earlier time-points obtained while the ecosystem is still equilibrating to estimate later behavior near steady-state . During the period immediately after the second antibiotic exposure , the optimal design generally required less frequent sampling than did the original design . This reduced sampling requirement reflected the shorter relaxation time distribution on the second post-antibiotic interval as shown in Figure 4 . Finally , the optimal design required considerably more sampling at the end of the time-series; the original experiments clearly under-sampled after the second antibiotic exposure , as is evident from high variability in model parameter estimates on this interval ( see Figure 5 and Figure S2 ) . To assess the predictive accuracy of our experimental design algorithm , we evaluated its ability to find a set of experiments to best estimate a model to predict held-out data ( Figure S3 ) . We used root mean square error ( RMSE ) to measure predictive accuracy . RMSE is the square root of the sum of squared differences between actual and predicted sequencing counts , averaged over refOTUs and time-points . Evaluation of predictive accuracy is important to assess the degree to which a probabilistic model generalizes to new data , without over-fitting features particular to one dataset . To perform this evaluation , we estimated optimal experimental designs for each subject , using three design strategies . For the first strategy , a sequential design , we gave the experimental design algorithm data for all refOTUs observed at a subset of time-points , and asked the algorithm to estimate additional time-points to sample in the same subject . For the second strategy , a cross-subject design , we gave the experimental design algorithm all observed data from one subject , and asked the algorithm to estimate time-points to sample for a different subject . In the third strategy , a dispersed design , we did not use the experimental design algorithm , and simply chose time-points to sample that were as evenly spaced on the study interval as possible . The dispersed design uses no information from observed data , and thus served as a baseline with which to compare the other two design strategies . The two experimental design strategies ( sequential and cross-subject ) that use prior information improved on the uninformative dispersed strategy by an average of 13% , as measured by reduction in prediction accuracy ( RMSE ) . Of the two informative strategies , neither consistently dominated the other . However , the cross-subject strategy did substantially outperform the sequential strategy for subject D . This subject exhibited the highest Signature Diversity equilibrium level ( SD1μ ) score , meaning that many refOTUs in this subject changed equilibrium levels subsequent to one or both antibiotic exposures . Consequently , equilibrium levels for refOTUs in subject D were harder to predict from prior equilibrium levels . Thus , the sequential design strategy , which uses only partial time-series data as input to the design algorithm , suffered in performance . In contrast , the cross-subject strategy that uses complete data from another subject , performed particularly well for subject D , because it leveraged prototype signatures predicted from subject E or F that were substantially similar to those in subject D .
We presented MC-TIMME , a unified computational framework for inferring dynamic signatures of the microbiota from high-throughput sequencing time-series datasets , and applied our framework to discover new features of the in vivo response of human gut microbes to antibiotic treatments . Our work represents both biologically and computationally significant advances . From the biological perspective , our study provides new insights into the differential and dynamic effects of antibiotic treatments on commensal bacteria in the human gut . Antibiotics disrupt the commensal flora , which can contribute to overgrowth and pathogenic effects of organisms such as Clostridium difficile , the cause of pseudomembranous colitis [61] . However , antibiotic effects have primarily been studied in pathogens; the range of effects on complex commensal populations remains largely unknown . Our results provide evidence , consistent across multiple human subjects , that sub-groups of commensals exhibit distinct temporal responses to treatment with a broad spectrum antibiotic . These results illustrate the staged dynamics of responses among sub-populations of commensals , and will enable future experimental studies to characterize the underlying molecular mechanisms behind these differential responses . From the computational standpoint , our study provides a robust , general-purpose framework for extracting fundamental information on ecosystem dynamics from massive sequencing datasets . Our framework may be applied to types of data other than 16S phylotypes , such as metagenomics or RNA Seq data , as well as model systems in animal or plant hosts , or studies in soil or marine environments . MC-TIMME employs probabilistic models of dynamics and associated measures of their properties , which yield important functional information that standard techniques for analyzing microbial communities cannot . Our use of adaptive Bayesian methods not only increases the strength of statistical inferences , but also provides signatures of microbial responses that are robust across multiple experimental subjects . Additionally , MC-TIMME enables optimal design of new time-series experiments , which will provide a strong foundation for future longitudinal studies of the microbiota . Our results on automated experimental design strategies have implications for how future longitudinal studies of the microbiota should be designed . An automated cross-subject design strategy generally performed comparably to or better than a sequential design strategy . A cross-subject design strategy uses all data from one subject to predict a future experimental design for a second subject . In contrast , a sequential design strategy uses limited samples from one subject to predict a future experimental design for the same subject . In the past , when the costs of experimentally interrogating samples were high , strategies using automated design were advocated in which samples would be over-collected , frozen , and then sequential design methods would be used to select the next samples to interrogate [24] . Our findings suggest that , in an era of plummeting sequencing costs , automated experimental designs based on pilot studies with small cohorts may prove more effective , particularly for clinical trials in which sample collection costs and logistics can be the limiting factors . However , confirmation of this hypothesis will require larger numbers of subjects and more detailed information about the heterogeneity of cohorts , with respect to factors such as demographics and environmental exposures . Additionally , we used a general-purpose , information theoretic utility function as a basis for selecting optimal experimental designs . This utility function has proven useful in many prior studies ( see e . g . , [46] ) , and performed well in our analyses . However , our framework for experimental design could readily be extended , by using utility functions that explicitly include financial or other costs involved in performing experiments . Analysis of host microbial ecosystems solely by 16S phylotyping has inherent limitations . Sequencing based methods suffer from various biases , due to factors such as the DNA extraction method and sequencing platform utilized [62] . Additionally , from 16S phylotyping data , it is only possible to infer abundances of taxa relative to other members of the microbial ecosystem detected with sequencing , and not the actual biomasses of individual taxa relative to the input mass of material analyzed . Thus , from our relaxation time analysis , one cannot infer the time required for taxa to equilibrate in terms of their absolute biomasses in vivo . Nonetheless , many studies have shown that relative abundances of organisms serve as important ecological indicators [2] . Relative abundances reflect differential abilities of organisms to compete for and effectively use limited resources , and thus provide insights into the roles of sub-communities within larger host ecosystems . However , the most profound limitation to 16S phylotyping data is that it is only useful for identifying which bacteria are present , not what they are doing . Ultimately , targeted or high-throughput functional studies [63] are essential for following up hypotheses generated based on 16S phylotyping . MC-TIMME can be extended with alternate models of dynamics for analyzing other time-series datasets . The key components of the model , including the infinite mixture model for prototype signatures and the noise model for counts data , employ general-purpose inference techniques that would not need to be modified to accommodate different models of dynamics . However , the Reversible Jump MCMC techniques we used for inference of intra-signature dimensionality changes require model-specific moves; in future work , more general techniques such as sparse priors [64] , [65]could be employed for inference of this portion of the model . For the experimental system we modeled , with defined antibiotic administrations , we assumed that perturbations to the microbiota start at known time-points . This model of dynamics could be extended for analyzing observational studies , in which naturally occurring perturbations may occur , by adding latent variables that automatically determine switch-points in dynamics . Another direction for extension would be building more elaborate relaxation time models . We used a relaxation time model based on ordinary differential equations , which assumes instantaneous transitions to different dynamic regimes and monotonic exponential decay to equilibrium levels . A more detailed model could allow smooth transitions between regimes , and richer kinetics of decay to equilibrium , to capture more subtle or chronic responses to perturbations . Additionally , more complex temporal correlations could be captured using a stochastic differential equation model such as the Ornstein-Uhlenbeck process [66] . Finally , longitudinal covariates , such as subject diet , could be added to the model of dynamics to capture exogenous factors that may affect the microbiota . Such extensions to MC-TIMME will enable increasingly sophisticated longitudinal studies , to expand our knowledge of the role of the microbiota in human health or disease .
|
Microbes colonize the human body soon after birth and propagate to form rich ecosystems . These ecosystems play essential roles in health and disease . Recent advances in DNA sequencing technologies make possible comprehensive studies of the time-dependent behavior of microbes throughout the body . Sophisticated computer-based methods are essential for the analysis and interpretation of these complex datasets . We present a computational method that models how human microbial ecosystems respond over time to perturbations , such as when subjects in a study are treated with a drug . When applied to a large publicly available dataset , our method yields new insights into the diversity of dynamic responses to antibiotics among microbes in the human body . We find that within an individual , sub-populations of microbes that share certain physiological roles also share coordinated responses . Moreover , we find that these responses are similar across different people . We use this information to improve the experimental design of the previously conducted study , and to develop strategies for optimal design of future studies . Our work provides an integrated computer-based method for automatically discovering patterns of change over time in the microbiota , and for designing future experiments to identify changes that impact human health and disease .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"small",
"intestine",
"clinical",
"research",
"design",
"biota",
"statistics",
"microbiology",
"algorithms",
"metagenomics",
"mathematics",
"gastroenterology",
"and",
"hepatology",
"genetics",
"and",
"genomics",
"computer",
"inferencing",
"ecological",
"metrics",
"ecosystems",
"medical",
"microbiology",
"statistical",
"methods",
"computing",
"methods",
"biology",
"ecosystem",
"functioning",
"microbial",
"ecology",
"longitudinal",
"studies",
"computer",
"science",
"ecosystem",
"modeling",
"computer",
"modeling",
"ecology",
"genomics",
"colon",
"computational",
"biology",
"modeling"
] |
2012
|
Inferring Dynamic Signatures of Microbes in Complex Host Ecosystems
|
We have sequenced the complete mitochondrial genome of the extinct American mastodon ( Mammut americanum ) from an Alaskan fossil that is between 50 , 000 and 130 , 000 y old , extending the age range of genomic analyses by almost a complete glacial cycle . The sequence we obtained is substantially different from previously reported partial mastodon mitochondrial DNA sequences . By comparing those partial sequences to other proboscidean sequences , we conclude that we have obtained the first sequence of mastodon DNA ever reported . Using the sequence of the mastodon , which diverged 24–28 million years ago ( mya ) from the Elephantidae lineage , as an outgroup , we infer that the ancestors of African elephants diverged from the lineage leading to mammoths and Asian elephants approximately 7 . 6 mya and that mammoths and Asian elephants diverged approximately 6 . 7 mya . We also conclude that the nuclear genomes of the African savannah and forest elephants diverged approximately 4 . 0 mya , supporting the view that these two groups represent different species . Finally , we found the mitochondrial mutation rate of proboscideans to be roughly half of the rate in primates during at least the last 24 million years .
An accurate and well-supported phylogeny is the basis for understanding the evolution of species . With the appropriate and adequate amount of data , it is possible not only to determine relationships among species , but also to date divergence events between lineages . In turn , divergence events can be correlated to environmental changes recorded in the fossil record to help understand mechanisms driving evolution . The power of these correlations , and arguments for particular environmental mechanisms driving evolutionary change , increases when the pattern is repeated across multiple taxa . Sequencing complete mitochondrial genomes has become a powerful tool in the investigation of phylogenetic relationships among animal groups , principally mammals , birds , and fishes . Despite this potential , and the proliferation of ancient mitochondrial DNA ( mtDNA ) research , geneticists so far have succeeded in sequencing complete mitochondrial genomes from ancient DNA for only a few extinct species of moas [1 , 2] and the woolly mammoth [3 , 4] . For both groups , the complete sequences resolved long-standing evolutionary questions , which argues for an extension of such analyses to other species . The living elephants comprise the last survivors of the Elephantidae , a once-flourishing sub-group of the Order Proboscidea that lived throughout much of Africa , Eurasia , and the Americas [5] . The evolution of Elephantidae , which includes the recently extinct mammoths , has been extensively studied in recent years using both modern and ancient sequences of mtDNA and nuclear DNA ( nuDNA ) . For example , nuDNA sequences have been used to argue that the African forest elephant is a valid species ( Loxodonta cyclotis ) , distinct from the African savannah elephant ( L . africana ) [6] , with the two species having diverged by 2 . 6 million years ago ( mya ) , though this view is disputed [7] . Two recent studies reported complete mtDNA genomes from the woolly mammoth ( Mammuthus primigenius ) [3 , 4] that provided strong evidence that mammoths were more closely related to Asian elephants than to African elephants . However , another study analyzing several nuDNA segments cautioned that the lack of a closely related outgroup is a problem in phylogenetic analyses of Elephantidae and argued that the relationship between mammoth and the living elephants is still unresolved [8] . To date , all genetically based analyses of elephant phylogenies have used dugong ( Dugong dugon ) and hyrax ( Procavia capensis ) as outgroups . These are the nearest living relatives of elephants , but they diverged from proboscideans some 65 mya [5 , 9] , which severely limits their power as effective outgroups for assessing elephant genetic data . In contrast , ancestors of the American mastodon , Mammut americanum , diverged from the Elephantidae lineage no earlier than 28 . 3 mya [10] , which would make mastodon a much more appropriate outgroup for the Elephantidae . Moreover , mastodon fossils preserved in permafrost and dating to the late Pleistocene have been recovered in eastern Beringia ( Alaska and Yukon ) . Their young age and preservation in permafrost means Beringian mastodon are excellent candidates for ancient DNA analyses [11 , 12] . The good biochemical preservation of mastodon samples was appreciated early in the study of ancient DNA . In 1985 , Shoshani and colleagues determined the immunological distances of albumins among mammoth , mastodon , and African and Asian elephant [13] . We report here what is to our knowledge the first complete mitochondrial genome for mastodon . The sequence was derived from a tooth collected in northern Alaska , where Pleistocene bones are well-preserved in permafrost . Collagen from the root of the tooth was radiocarbon-dated to more than 50 , 000 y before present ( BP ) , but the geological provenance suggests it is not older than 130 , 000 y ( see Materials and Methods ) . This extremely old and complete sequence is of interest in its own right , but it also can be used to help resolve existing debates concerning the phylogeny of the Elephantidae and allows us to evaluate rates of molecular evolution within the Proboscidea , because it finally provides an appropriate outgroup for such analyses . Here , we use the mastodon mtDNA sequence to resolve relationships among African elephants , Asian elephants , and mammoths and to more accurately date their divergence times .
To determine the mastodon mtDNA genome sequence , we used 78 primer pairs , separated into two sets of 39 pairs each , on DNA extract from a mastodon molar originating from northern Alaska ( Figure 1 ) and performed quadruplicate amplifications for each primer pair . In the first round of amplification , 48 primer pairs yielded at least one positive result , and for 36 primer pairs all four attempts were positive . To obtain the remaining fragments , we used four consecutive rounds of redesigning primers ( see Protocol S1 for details ) . In each round , whenever possible , we used the mastodon sequences from flanking fragments as the basis for primer design . With three primer pairs , PCR controls or extraction controls yielded products of the correct size . All three cases of contamination occurred with primers for which it was not possible to design the primers in a way that they selected against the amplification of human DNA . The resulting products were cloned and sequenced and turned out to be exclusively of human origin . Two out of six clones obtained from a single amplification from mastodon extract using one of these primer pairs were also of human origin . We did not replicate part of the mastodon sequence in an independent laboratory , following the argument we have made previously [14] , and in contrast to Cooper and Poinar [15] , who argued that this is necessary in all ancient DNA studies . None of the determined sequence fragments was identical to any known sequence either from GenBank or determined previously in our laboratory . Moreover , to ensure authenticity of each individual fragment , apart from extensive internal replication , we compared each PCR fragment individually to all published sequences from GenBank and noted the similarity to both Elephantidae and hyrax and dugong mtDNA sequences . All amplified fragments showed between 78 . 1% and 98 . 6% identity to Elephantidae sequences and between 45 . 9% and 92 . 6% identity to hyrax and dugong sequences . Moreover , for individual fragments , the average identity to Elephantidae was 6 . 2% to 40% higher than that to hyrax and dugong . These results are consistent with the phylogenetic position of the mastodon and difficult—if not impossible—to explain by contamination . Therefore , we conclude that our results indeed represent authentic mastodon sequences . The length of the mastodon mitochondrial genome is 16 , 469 bp . Thus , it is about 300–400 bp smaller than the mitochondrial sequences of the other proboscideans published to date . This difference in length is most likely an artifact due to the failure to obtain an amplification product that covers the complete tandem repeat in the control region . The mastodon mitochondrial genome contains 13 protein coding genes , 22 tRNA genes , two rRNA genes , and the control region , as expected for a placental mammal . Stop and start codons are shared with at least one of the other proboscideans for each of the protein coding genes , except ND3 , ND4 , and ND6 . For these three genes , the mastodon sequences start with ATT , ATG , and GTG , whereas the corresponding genes of all other proboscideans start with ATC or ATA , GTG , and ATG , respectively . Differences in the annotation were observed in the other sequences notably because of stop codons being created by polyadenylation , a phenomenon widely present in mitochondrial genomes of other vertebrates [16] . The early stop of ND4 of the woolly mammoth sequences was not observed , and none of the proteins of the mastodon seems to be truncated . With a divergence time of at least 24 million years from any living relative , the mastodon sequence shows that the multiplex PCR approach can also be applied to taxa without sequence information from closely related species . Partial DNA sequences that are claimed to be derived from mastodon are available from four studies . The sequences of Yang et al . [17] ( 228 bp of CYTB ) and Joger et al . [18] ( 294 bp of CYTB ) show substantially less divergence from mammoth and elephants than our sequence . The numbers of substitutions between their mastodon sequences and the Elephantidae are comparable to differences within the Elephantidae ( results not shown ) . In contrast , our sequence clearly differs from the elephantid sequences . The results of Yang et al . have been recently questioned [19 , 20] . Debruyne et al . [19] concluded that Yang et al . 's mammoth sequence is a probable chimera of African and Asian elephant . Our study also rejects the results of Yang et al . for mastodon because the fragment of their CYTB mastodon sequence clusters closely within Elephantidae ( see Figure S1A ) , as expected by the number of substitutions . For the same reason , the partial sequence of Joger et al . also seems to be derived from an Asian/African elephant contaminant ( see Figure S1B ) . Finally , the unpublished 16S rDNA sequences obtained by Park et al . and Goldstein et al . ( GenBank accession numbers AF279699 and AY028924 , respectively ) show 97% and 99% identity to primates using Blastn [21] , which shows that these two sequences also were very likely derived from contamination . In order to further investigate proboscidean evolution , we used the six existing whole mitochondrial genomes of Elephantidae ( two African elephants , L . africana; two Asian elephants , Elephas maximus; and two woolly mammoths , Mammuthus primigenius ) together with our newly obtained Mammut americanum sequence . As the mastodon lineage is evolutionarily much closer to Elephantidae than to extant outgroup species such as dugong or hyrax , multiple substitutions should not represent such a pronounced problem as is the case with the latter species . A comparison of the seven proboscidean mitochondrial genomes to those from hyrax and dugong shows that the number of substitutions separating mastodon from the Elephantidae is less than half that separating either hyrax or dugong from the Elephantidae ( Table S1 ) . Consistent with this result , the rates of synonymous ( ds ) and nonsynonymous ( dn ) substitutions are much lower when using mastodon as the outgroup ( ds = 0 . 4; dn = 0 . 05 ) . Using dugong or hyrax as the outgroup , ds is 1 . 2 , showing substitution saturation [4] , and dn is still 0 . 15 . The transition/transversion ratio can also be used as a measure of substitution saturation [3 , 22] . Including the whole mitochondrial genome of the cow ( Bos taurus ) and the ostrich ( Struthio camelus ) reveals that dugong and hyrax have reached a plateau for both number of substitutions and transition/transversion ratio when compared to Elephantidae , whereas the mastodon has not ( Figure 2 ) . Thus , as previously noted [23] , the mastodon represents a much better outgroup for inferring Elephantidae evolution than either dugong or hyrax . The greater suitability of mastodon as an outgroup to the Elephantidae is further demonstrated by the results of the phylogenetic analyses . Using the mastodon sequence as an outgroup , we obtained higher support values for a sister group relationship of mammoth and Asian elephant than previous studies [3 , 4] and obtained the same tree topology from different methods of phylogenetic inference ( Table 1 ) . The bootstrap values for this relationship were between 94% ( neighbor joining [NJ] ) and 99% ( maximum likelihood [ML] ) with a Bayesian posterior probability of 1 . 00 ( Table 1 ) . These bootstrap values do not vary significantly among substitution models and are not dependent on whether the data are partitioned or unpartitioned ( results not shown ) . Thus , at least with regard to mtDNA sequences , the relationship among mammoth and the living elephant species can no longer be seen as equivocal , as argued by some authors [8] . This result indicates that mastodon would also provide an excellent outgroup for phylogenetic analyses of Elephantidae using nuclear sequences [8] if it became possible to recover nuclear mastodon DNA sequences . Using our complete mtDNA mastodon sequence , we were able to employ gene-by-gene phylogenetic analyses to explain why several earlier studies found a sister group relationship between African elephants and mammoths . The reconstructed phylogeny of the Elephantidae varied widely when we used each of the 13 protein coding genes and the two rRNAs individually . We recovered the mammoth–Asian elephant topology for the majority of the genes , but with lower support values ( 44%–90% for bootstraps and 0 . 42–1 . 00 for posterior probabilities ) . Other genes supported different tree topologies , sometimes with high bootstrap values or Bayesian posterior probabilities ( up to 90% or 1 . 00; see Figure 3 and Table S2 ) . In fact , when considering NJ trees alone , the majority ( eight of 15 ) of the single-gene analyses in fact supported an incorrect topology . Some single-gene analyses resulted in different , yet well supported topologies when hyrax and dugong were used as the outgroup instead of mastodon [4] . These results indicate that studies based on a single gene can be misleading , and long sequences may often be necessary to obtain correct phylogenies ( Figures 3 and S2; Protocol S3; see also , e . g . , [24–27] ) . Unlike the sequences of the nearest living outgroups of the Elephantidae , the dugong and the hyrax , the mastodon sequence has not yet reached substitution saturation ( Figure S3 ) . Consequently , the mastodon mtDNA sequence provides better estimates of the dates of divergence within the Elephantidae . Moreover , fossil evidence constrains the divergence of hyrax and dugong from proboscideans to earlier than ∼60 mya , the date for the oldest proboscidean fossil genus , Phosphatherium [28] . Previous attempts to date the Elephantidae divergence either used the two very distant outgroups , with limited success because the sequences violated the molecular clock assumption [4] , or used the divergence of African elephant as calibration point [3] , which yields more reliable relative divergence dates within Elephantidae , but does not provide independent estimates of divergence times within this group . Thus , the mastodon sequence allows independent dating of Elephantidae divergence times based on the sequence of a well-calibrated outgroup . Using the best-fitting topology , we found the TN93 [29] model of substitution to be the simplest that fitted the data ( results not shown ) . Using this substitution model , we were not able to reject the assumption of a molecular clock for the whole mtDNA genome sequences of the mastodon and the three Elephantidae species ( 2Δℓ = 6 . 8; p = 0 . 24 ) . Using the Bayesian approach of Yang and Rannala [30] , we evaluated the saturation level of our data by plotting the posterior means versus the width of the 95% credibility intervals ( CIs ) of the divergence times . The relationship is almost linear , and the coefficient of correlation of the linear regression has a value of 0 . 85 ( Figure 4 ) . Although a correlation of 0 . 85 suggests that the sequence data are highly informative it is still possible that the accuracy of divergence time estimates could be further improved by additional sequence data . When we used the paleontologically determined divergence date of 24–28 mya for mastodon [10 , 31] , the divergence time of the African elephant turned out to be older than the 6 mya previously assumed [32] . In fact , we calculated it to have a posterior mean of 7 . 6 mya ( 95% CI 6 . 6 to 8 . 8 mya ) when using the whole mitochondrial genome . The divergence between mammoth and Asian elephant also moves back in time to 6 . 7 mya ( CI 5 . 8 to 7 . 7 mya; Figure 5; Table 2 ) . Both dates are concordant with the presence of African elephant and Asian elephant fossils by 5 . 4–7 . 3 and 5 . 2–6 . 7 mya [33] , respectively . Only about a million years separates the two divergence events , which is less than in humans , chimpanzees , and gorillas [34] , for which extensive lineage sorting has been found [35 , 36] . Hence , we expect lineage sorting for the nuclear genome to also be problematic for Elephantidae , as claimed by Capelli et al . [8] . Our new estimates make the divergence of mammoth and African and Asian elephants even closer in time to the divergence of humans , chimpanzees , and gorillas , and other mammalian taxa . A number of environmental changes were occurring globally at that time , including the spread of grasslands and an increase in C4 plant biomass [37 , 38] . Further efforts should be given to studying the relationship between environmental changes and these phylogenetic events during the late Miocene . Finally , our revised divergence dates also have bearing on the status of the African forest elephant . Based on nuDNA sequences , Roca et al . [6] argued that the African forest elephant represents its own species , L . cyclotis , distinct from the savannah elephant , L . africana . Using nuDNA sequence data and a divergence time between African and Asian elephants of 5 mya , they estimated the divergence between the nuDNA sequences from African savannah and forest elephants to be 2 . 63 ± 0 . 94 mya , and argued that both the deep divergence and the reciprocal monophyly between forest elephants and savannah elephants with regard to nuDNA support the distinction of the two forms as different species , a view also supported by microsatellite analyses [39] . However , based on extensive mtDNA analyses , both Eggert et al . [40] and Debruyne [7] disputed this view , as they found forest and savannah elephants being polyphyletic with respect to mtDNA sequences . In an extension of their earlier analysis , using both X and Y chromosomal and mtDNA sequences , Roca et al . [41] confirmed their view of a distinction on the species level and argued for unidirectional gene flow of mtDNA from forest elephants to savannah elephants . Although , complete mtDNA sequences for the two African elephant are not available , we can use our results to recalculate the divergence time between forest and savannah elephant inferred from the nuclear sequences . Using our estimate of 7 . 6 mya for the initial Loxodonta divergence increases the estimated divergence time to 4 . 0 mya . This date is older than the divergence times of many species pairs and hence supports the classification of African savannah and forest elephants as different species as proposed by Roca et al . [6 , 41] . The initial sequencing of the mammoth mitochondrial genome indicated that the substitution rate within Elephantidae is much lower than in humans and the African great apes [3] . We can use the mastodon sequence to determine whether this arose recently or in the more distant past . We compared the substitution rates of the mammoth , elephants , and mastodon , which had a most recent common ancestor 24–28 mya , to those of human , chimpanzee , gorilla , and baboon , which had a most recent common ancestor about 33 mya [42] . The substitution rate for the whole mtDNA genome was found to be more than twice as high in the four species of primates than in the four species of proboscideans . However , the distribution of rates among sites within the genome was almost identical ( Table S3 ) . It is not clear what causes the difference in rates . A possible explanation could be differences in body size , which in turn influence metabolic rates [43 , 44] , but further studies on more species would be necessary to evaluate whether this is the major , or only , cause . The sequence of a complete mtDNA genome obtained from a mastodon tooth extends the time frame for large-scale sequencing of ancient DNA substantially . Inclusion of this sequence in phylogenetic analyses confirms mammoth and Asian elephants as sister taxa and provides evidence for earlier divergences between Elephantidae species . The similarity of the divergence dates between Elephantidae species and between humans and African great apes suggests that a change in environmental conditions triggered speciation in African mammals beginning some 7 . 5–8 mya . Finally , we found further evidence that the mitochondrial substitution rate in proboscideans is considerably lower than in primates , and this difference manifested by at least 24 mya . Although permafrost environment is especially suitable for long-term DNA survival , DNA sequences about 130 , 000 y old have also been reported from non-permafrost remains . Moreover , the age limit for preservation of plant DNA in permafrost environment is even older , currently at 300 , 000–500 , 000 y [45] . Even if the DNA from specimens is more fragmented than in the sample used in this study , the two-step multiplex procedure would allow reconstruction of long , continuous sequences such as complete mitochondrial genomes , substantially enlarging the possibilities of ancient DNA analyses .
For DNA extraction , we cut a ∼200-g sample along the tooth root of a mastodon molar ( United States Bureau of Land Management–Alaska collection IK-99–237 ) collected in 1999 on the Ikpikpuk River ( 69° 22′ 10′′ N , 154° 40′ 46′′ W; Figure 1 ) , which drains the central arctic coastal plain of northern Alaska . The tooth was a detrital find collected on a point bar along with numerous other large mammal bones of late Pleistocene affinity . These bones are eroded from cutbanks and deposited in large quantities on point bars along a number of rivers that drain the arctic coastal plain . The fluvial and alluvial sediments in these systems were mostly deposited during the late Pleistocene and early Holocene , but contain reworked bones of Pleistocene age [46–48] . There is a rare occurrence of sediments dating to the last interglacial period ( ∼130 , 000–100 , 000 y BP ) . To date , an assemblage of ∼3 , 000 bones has been collected , 312 of which have been radiocarbon-dated [48 , 49] . Most are within the radiocarbon range ( <50 , 000–40 , 000 y ) , but about 20% have returned nonfinite dates ( i . e . , “greater than” ages ranging from 50 , 000 to 40 , 000 y BP ) . Radiocarbon analyses of a collagen extract from mastodon tooth IK-99–237 yielded a nonfinite 14C age of >50 , 000 y BP ( Lawrence Livermore National Laboratory number CAMS91805 ) , which places a lower limit on its age . In terms of its maximum age , there are only two bones in the entire assemblage from a taxon ( Praeovibos ) that dates to the middle Pleistocene , and even those two bones were re-worked into late Pleistocene sediments . Consequently , it is conservatively estimated that nonfinite-age bones in the assemblage date to between approximately 50 , 000 y BP and the end of the penultimate glaciation ( about 150 , 000 y ago ) . Since the mastodon is principally an inhabitant of interglacial periods in Alaska , it is most likely IK-99–237 dates to the last interglacial period ( i . e . , its maximum age is probably ∼130 , 000 y BP ) , though we have no way of absolutely ruling out an age between 100 , 000 and 50 , 000 y BP . Bones in this assemblage are extremely well-preserved because they were entombed in permafrost sediments before being exposed by river erosion . This accounts for the exceptional preservation of DNA in IK-99–237 . Part of the tooth root was cut into small pieces and ground to a fine powder in a freezer mill ( freezer mill 6750 , Spex SamplePrep , http://www . spexcsp . com/sampleprep/ ) using liquid nitrogen . About 25 g of the powder was incubated overnight under constant agitation at room temperature in 700 ml of extraction buffer consisting of 0 . 45 M EDTA ( pH 8 . 0 ) and 0 . 25 mg/ml proteinase K . After centrifugation the supernatant was concentrated to ∼50 ml using the Vivaflow 200 system ( Vivascience , http://www . vivascience . com ) with a polyethersulfone membrane with a molecular weight cut-off of 30 , 000 [50] . DNA was bound to silica using 40 ml of guanidinium thiocyanate buffer [51] and 100 μl of silica suspension for each of the five 10-ml aliquots of concentrated extraction buffer , with the pH adjusted to 4 . 0 using hydrochloric acid . After incubation for 3 h under constant agitation the silica was pelleted by short centrifugation and washed once with 1 ml of guanidinium thiocyanate buffer and twice with 1 ml of wash solution ( 51 . 3% ethanol , 125 mM NaCl , 10 mM Tris , and 1 mM EDTA [pH 8 . 0] ) . DNA was eluted using 50 μl of TE buffer for each of the five aliquots , resulting in ∼250 μl of extract . An extraction blank was carried alongside throughout all steps of extraction to monitor for possible contamination . We designed 78 primer pairs using published sequences of African ( Loxodonta1 ) and Asian ( Elephas1 ) elephants , mammoth ( Mammuthus1 ) , and dugong ( Dugong ) . The length of the targeted fragments varied between 139 and 334 bp ( including primer ) , covering the entire mitochondrial genome of the mastodon except a repeat sequence in the control region . Concordant with the initial paper describing the multiplex approach for DNA amplification from ancient samples [3] , we divided the primer pairs for the first step into two sets to avoid amplification of the short overlapping fragments between adjacent amplification products [52] . In the second step , each primer pair was used individually , and to increase the specificity of amplification , in all except four cases , one primer per pair was “nested” compared to the first step , resulting in fragments between 139 and 324 bp long ( including primers ) . Wherever possible , primers were designed to exclude amplification of human DNA . Four primary amplifications ( the multiplex step ) were done for each of the two primer sets using two times 5 μl of 1:5 and 1:10 dilutions , respectively , of the extract , resulting in altogether eight primary amplifications . PCRs were conducted in a final volume of 20 μl consisting of 1× GeneAmp PCR Buffer II ( Applied Biosystems , http://www . appliedbiosystems . com ) , 4 mM MgCl2 , 1 mg/ml BSA , 250 nM of each dNTP , 2 U AmpliTaq Gold ( Applied Biosystems ) , and 150 nM each of 78 primers ( 39 primer pairs per set ) . Initial denaturation and activation of the polymerase was done for 9 min at 94 °C followed by 30 cycles at 94 °C for 20 s , at 48 , 50 , or 52 °C for 30 s , and at 72 °C for 30 s . Aliquots of 5 μl of a 1:20 dilution of this primary amplification were used for the two sets of 39 individual secondary amplifications . These secondary PCRs were done in a final volume of 20 μl consisting of the same reagent concentrations as described above except that only 0 . 25 U of AmpliTaq Gold was used and the primer concentration was raised to 1 . 5 μM . Cycling conditions were the same as above except that between 30 and 40 cycles were performed . Extraction and water controls were carried along during all steps . As not all amplifications were successful in the first attempt , several primers were redesigned after sequencing the flanking regions obtained from successful amplifications , and more primary amplifications were performed ( see Results/Discussion and Protocol S1 ) . Amplification products were visualized on 2 . 0% agarose gels , and products of correct lengths were cloned using the Topo TA Cloning Kit ( Invitrogen , http://www . invitrogen . com ) . In cases where amplifications showed visible primer dimers in addition to products of the correct length , the products were isolated from the gel and purified using the QIAquick Gel Extraction Kit ( Qiagen , http://www . qiagen . com ) . After colony PCR [53] and purification of the products using the QIAquick PCR Purification Kit ( Qiagen ) and a Biorobot , a minimum of three clones per amplification ( see Protocol S1 ) were sequenced on an ABI 3730 capillary sequencer ( Applied Biosystems ) using M13 universal primers . All alignments were made using ClustalW [54] with default parameters . The D-loop sequences were removed from all genome sequences prior to analyses . We aligned the mastodon sequence to previously published sequences of two mammoths ( Mammuthus1 and Mammuthus2 ) , two African elephants ( Loxodonta1 and Loxodonta2 ) , and two Asian elephants ( Elephas1 and Elephas2 ) . For this alignment , we verified that the start and stop codons were aligned and that the open reading frames were preserved . For the concatenation of the protein coding genes , all genes were aligned individually and subsequently concatenated , ignoring the fact that some of them were overlapping . Hence , some nucleotides are duplicated in our concatenation . We examined publicly available mastodon sequences by aligning them with our sequence and constructing NJ trees using MEGA3 . 1 [55] with default parameters ( see Protocol S2 for details on the program settings ) . To test the suitability of our outgroup we estimated the degree of substitution saturation for the concatenated protein coding genes by computing the number of synonymous and nonsynonymous substitutions per site with the Pamilo-Bianchi-Li method implemented in MEGA3 . 1 . We also used MEGA3 . 1 to compute the number of substitutions per site between proboscideans , dugong , and hyrax and to compute the transition/transversion ratio by aligning the seven proboscidean mtDNA genomes with those of dugong , hyrax , cow , and ostrich . To solve the Elephantidae phylogeny we used MEGA3 . 1 to reconstruct trees using NJ and maximum parsimony . Initially , parameters similar to those used in [4] were chosen . We used the default parameters from MEGA3 . 1 with 10 , 000 bootstrap replicates . In addition , we built a Bayesian tree using MrBayes 3 . 1 [56] . The tree with maximum posterior probability was computed using one million iterations ( see Protocol S1 for information on the options used ) . A GTR model [57 , 58] and a HKY85 model [59] of nucleotide substitution , both with gamma distributed rates of substitutions among sites , were used . The ML tree was constructed using Paup* [60] with an exhaustive search . For the ML tree , we performed 1 , 000 bootstrap replicates and chose a GTR model of substitution with gamma distributed rates among sites . Finally , we reconstructed the phylogeny for the seven sequences individually for each protein coding gene using MrBayes 3 . 1 and a HKY85 model of substitutions as well as by NJ as described above . The same analysis was performed for the concatenated sequence of the protein coding genes . However , the result of this analysis did not differ significantly from that of the analysis performed using the complete sequence ( results not shown ) . Because the different reconstructions revealed a single topology , we tested the different models of nucleotide substitutions against each other following the Felsenstein hierarchy in order to choose the best-fitting model . To this end , we compared the different models by likelihood ratio tests with the program baseml in PAML [61] using the alignment of the seven sequences without the D-loop , using gamma distributed rates among sites with eight categories , and considering a non-clock situation . The topology of the tree used for these analyses is shown in Figure 5 and identical to that in previous publications [3 , 4] . We reconstructed the tree once again using the above methods with TN93 , the simplest model of substitution for which none of the more complex models fitted the data significantly better , but no significant difference from the previous results was observed ( results not shown ) . To test whether the seven sequences evolved following a molecular clock we estimated the ML values for the phylogenetic tree of the seven sequences under both a non-clock assumption and a molecular clock assumption and compared the likelihoods of the trees obtained under the two assumptions using a likelihood ratio test ( five degrees of freedom ) . As before , we assumed gamma distributed rates among sites and TN93 as a model of substitution . Given the difficulty of precisely estimating calibration points and hence divergence times [62 , 63] , we used a Bayesian approach and the program mcmctree to estimate the posterior means and the CIs of the divergence times . We used as a calibration point the divergence of the mastodon , estimated to be 24–28 mya [31] applying the lower and upper bound method . Given the unavailability of TN93 , we chose HKY85 , the most complex model implemented in mcmctree . The burn-in was set to 10 , 000 , the number of samples to 100 , 000 , and the sample frequency to five , as in [30] , with four independent chains for each analysis . We chose wide priors for κ , the transition/transversion ratio , r , the substitution rate , and α , the parameter of the gamma distribution of rate variation among sites , to avoid too strong an influence of the priors on the posteriors . The 95% prior intervals were set to ( 0 . 00 , 111 . 24 ) , ( 0 . 02 , 3 . 69 ) , and ( 0 . 00 , 1 . 23 ) for κ , r , and α , respectively . To compute the rate of substitution for primates we used the same approach for a four-taxon tree of baboon , gorilla , chimpanzee , and human . The calibration point was set to a lower and upper bound of 6–8 mya for the chimpanzee–human split [64] and an upper bound of 33 mya for the baboon divergence ( [42] and references therein ) . We used the same disperse priors as before for κ , r , and α and the same options for the Markov chain Monte Carlo . Since the use of a single model of evolution for the whole mtDNA sequence may result in errors , we partitioned the data into five subsets . We separated the first , second , and third positions of the codons , the rRNAs , and the tRNAs . The few noncoding sites were excluded to avoid overparametrization . The protein–protein overlapping fragments were classified under the second codon position , except for the seven sites between ND5 and ND6 that were excluded because the two genes are on different strands . Indeed we inverted the strands according to the transcription process . The phylogenetic tree was constructed using MrBayes 3 . 1 with a GTR model of substitution and one million iterations with unlinked partitions . We also allowed the partitions to have different rates . The divergence times using the partitions were computed using mcmctree with the options described above .
The GenBank ( http://www . ncbi . nlm . nih . gov ) accession numbers for the complete mitochondrial genome sequences of E . maximus , L . africana , and Mammuthus primigenius are NC_005129 . 1 for Elephas1 and NC_005129 . 2 for Elephas2 , NC_000934 . 1 for Loxodonta1 and DQ316069 . 1 for Loxodonta2 , and NC_007596 . 2 for Mammuthus1 and DQ316067 . 1 for Mammuthus2 . For the partial sequences of Mammut americanum the GenBank accession numbers are U23737 . 1 ( denoted Yang Mastodon ) , AY028924 . 1 , and AF279699 . 1; for the partial sequences of Mammuthus primigenius the GenBank accession numbers are U23738 . 1 ( denoted Yang Mammuthus1 ) and U23739 . 1 ( denoted Yang Mammuthus2 ) . The accession number for the whole mitochondrial genome of Mammut americanum determined in this study is EF632344 . For the non-proboscideans , the GenBank accession numbers for the whole genomes used in this study are NC_003314 . 1 for D . dugon ( Dugong ) , NC_004919 . 1 for P . capensis ( Procavia ) , NC_006853 . 1 for B . taurus ( Bos ) , Y12025 . 1 for S . camelus ( Struthio ) , NC_001992 . 1 for Papio hamadryas ( Baboon ) , NC_001645 . 1 for Gorilla gorilla ( Gorilla ) , NC_001643 . 1 for Pan troglodytes ( Chimpanzee ) , and NC_001807 . 4 for Homo sapiens ( Human ) .
|
We determined the complete mitochondrial genome of the mastodon ( Mammut americanum ) , a recently extinct relative of the living elephants that diverged about 26 million years ago . We obtained the sequence from a tooth dated to 50 , 000–130 , 000 years ago , increasing the specimen age for which such palaeogenomic analyses have been done by almost a complete glacial cycle . Using this sequence , together with mitochondrial genome sequences from two African elephants , two Asian elephants , and two woolly mammoths ( all of which have been previously sequenced ) , we show that mammoths are more closely related to Asian than to African elephants . Moreover , we used a calibration point lying outside the Elephantidae radiation ( elephants and mammoths ) , which enabled us to estimate accurately the time of divergence of African elephants from Asian elephants and mammoths ( about 7 . 6 million years ago ) and the time of divergence between mammoths and Asian elephants ( about 6 . 7 million years ago ) . These dates are strikingly similar to the divergence time for humans , chimpanzees , and gorillas , and raise the possibility that the speciation of mammoth and elephants and of humans and African great apes had a common cause . Despite the similarity in divergence times , the substitution rate within primates is more than twice as high as in proboscideans .
|
[
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] |
[
"evolutionary",
"biology",
"mammals"
] |
2007
|
Proboscidean Mitogenomics: Chronology and Mode of Elephant Evolution Using Mastodon as Outgroup
|
Lassa virus ( LASV ) is responsible for a viral hemorrhagic fever in humans and the death of 3 , 000 to 5 , 000 people every year . The immune response to LASV is poorly understood , but type I interferon ( IFN-I ) and T-cell responses appear to be critical for the host . We studied the response of myeloid dendritic cells ( mDC ) to LASV , as mDCs are involved in both IFN-I production and T-cell activation . We compared the response of primary human mDCs to LASV and Mopeia virus ( MOPV ) , which is similar to LASV , but non-pathogenic . We showed that mDCs produced substantial amounts of IFN-I in response to both LASV and MOPV . However , only MOPV-infected mDCs were able to activate T cells . More surprisingly , coculture with T cells completely inhibited the activation of LASV-infected mDCs . These differences between LASV and MOPV were mostly due to the LASV nucleoprotein , which has major immunosuppressive properties , but the glycoprotein was also involved . Overall , these results suggest that mDCs may be important for the global response to LASV and play a role in the outcome of Lassa fever .
Viral hemorrhagic fevers ( VHF ) are an important global health problem with a significant yearly death toll . Among VHFs , Lassa fever ( LF ) is endemic in West Africa , with an estimated 300 , 000 to 500 , 000 cases and 3 , 000 to 5 , 000 deaths every year [1] . Populations living from Senegal to Nigeria are at risk , which is up to 200 million people [2] . Lassa virus ( LASV ) , the causative agent of LF , is listed by the World Health Organization ( WHO ) as one of the emerging pathogens likely to cause severe outbreaks in the near future , and for which there are few or no medical countermeasures [3] . There are currently no approved vaccines against LASV , no treatment efficient in endemic areas , and limited knowledge about the pathogenesis of LF . LASV is an Old-World Arenavirus . This enveloped virus contains two segments of negative strand RNA ( S and L ) coding for four proteins . The nucleoprotein ( NP ) and glycoprotein precursor ( GPC ) are coded by the S segment , the matrix protein ( Z ) and RNA-dependent RNA polymerase ( L ) by the L segment . ORFs are coded in ambisense and separated by a hairpin structure containing termination sites . LF begins as a non-specific febrile illness , indiscernible from multiple diseases more common in West Africa , such as malaria , typhoid , dysentery , arboviruses , etc [4] . LF then progresses to pharyngitis , vomiting , diarrhea , and hemorrhagic symptoms . Hypovolemia , multiple organ failure , and shock syndrome then lead to death [5] . The pathophysiological mechanisms that give rise to these symptoms are still poorly understood , but adverse effects of the immune response are strongly suspected [6] . In humans , only small amounts of neutralizing antibodies are detected late after the disease and the humoral response does not correlate with survival [7] . Recovery from LF appears to be associated with early innate immunity and effective T-cell responses [8] . Studies with non-human primates ( NHPs ) support this hypothesis , in which the survival of LASV-infected NHPs correlates with the early type I interferon ( IFN-I ) response and detection of circulating antigen-experienced T cells [9] . Early innate responses and the initiation of T-cell immunity are two properties of dendritic cells ( DC ) . DCs are also important targets of LASV in vitro and in vivo [10 , 11] . Among DCs , plasmacytoid dendritic cells ( pDC ) are highly potent IFN-I producers [12] , whereas myeloid dendritic cells ( mDC ) are specialized in antigen presentation and the induction of T-cell responses [13] . However , mDCs can also produce IFN-I under certain conditions , which can play a crucial role in in vivo , as they are more numerous that pDCs [14] . In mice infected with Lymphocytic Choriomeningitis Virus ( LCMV ) , mDCs are the main source of IFN-I , and such IFN-I production is required for efficient T-cell priming [15] . mDCs may play an important role in LASV infection , as both IFN-I and T-cell responses are important . We previously showed that LASV replicated in monocyte-derived DCs ( moDC ) but did not activate them [10] . Here we further investigated the relevance of the escape of LASV from the DC response using another arenavirus , Mopeia virus ( MOPV ) [16] . MOPV is phylogenetically very close to LASV and has been isolated from Mastomys natalensis , the main reservoir of LASV [17] . However , MOPV is non-pathogenic in NHPs and can protect against LASV challenge [18] . No human case of MOPV infection has ever been reported [16] . Comparing MOPV and LASV , two very similar viruses but with different pathogenic potential , is a useful approach to identify immune and viral features involved in LASV pathogenesis . This comparison is used to model the differences between fatal and non-fatal LF in vitro . In moDCs , MOPV infection was productive and moderately activated cells [19] . In addition , MOPV-infected moDCs cultured with autologous T cells induced early T-cell activation , strong proliferation , and acquisition of effector , memory , and cytotoxic phenotypes . In contrast , LASV-infected moDCs only induced weak and delayed T-cell responses [20] . These results highlight the differences in the response of moDCs to MOPV and LASV , and their ability to induce T-cell responses . They further emphasize the link between DCs and the global response to LASV . Nevertheless , recent findings have raised doubts concerning the validity of the moDC model . High dimensional mapping of the DC surface phenotype showed that moDCs do not match any in vivo population ( DCs from blood , lymphoid organs , or skin ) [21] . Thus , moDCs are not phenotypically representative of any DC subset in healthy individuals . Studies on influenza A virus also showed different susceptibility to infection between moDCs and primary mDCs [22] . In light of these results , we decided to investigate the role of mDCs during LASV infection using our LASV/MOPV comparison model . We purified primary human mDCs and infected them with MOPV and LASV . mDCs infected with MOPV produced much higher levels of IFN-I than moDCs . Surprisingly , LASV-infected mDCs also produced high levels of IFN-I . Unlike moDCs , mDCs were not productively infected by MOPV or LASV , due to the strong IFN-I response . We further characterized the mDC response using transcriptomic and Luminex-based approaches . These large-scale results showed differences in the global activation of mDCs depending on the virus . We next tested the ability of mDCs to induce T-cell responses . Coculture with autologous T cells modulated the mDC response: MOPV-infected mDCs were more strongly activated when cultured with T cells , whereas activation of LASV-infected mDCs was inhibited . Furthermore , MOPV-infected mDCs induced early activation , proliferation , and cytotoxic responses of T cells . LASV-infected mDCs induced no or little T-cell activation . These results suggest crosstalk between mDCs and T cells , influencing not only T-cell responses but also mDC activation . We generated MOPV/LASV chimeras and tested them in our coculture model to identify viral parameters responsible for the differences between LASV and MOPV .
Previous studies showed that MOPV and LASV induced different IFN-I responses in moDCs [19 , 23] . We first investigated the kinetics of the IFN-I response to MOPV by mDCs . The synthesis of IFNα1 , IFNα2 , and IFNβ mRNAs after MOPV infection started as early as eight hours post-infection ( hpi ) , peaked at 24 hpi , and then decreased to reach the level of that of uninfected cells at 40 hpi ( S1A Fig ) . We detected no TNFα mRNA during MOPV infection . Based on these results , we quantified IFN-I at 24 hpi for the rest of the study . The synthesis of IFNα1 , IFNα2 , and IFNβ , as well as that of IFNα6 and IFNα8 mRNA was significantly higher in MOPV-infected mDCs than in uninfected mDCs ( Fig 1A ) . Surprisingly , LASV-infected mDCs also produced high levels of IFN-I , nearly similar to that of MOPV-infected mDCs . The IFN-I response to both viruses was associated with upregulation of the activation molecules CD40 , CD80 , CD83 , and of TRAIL ( Fig 1B ) . Viral titers in supernatants of MOPV- or LASV-infected mDCs decreased over time ( Fig 2A , white circles ) , indicating that mDC infection by MOPV and LASV was not productive . Neither significant levels of transcription nor replication of either virus was detected in cells , as illustrated by the lack of mCherry synthesis in most cells after infection with mCherry-expressing LASV or MOPV ( Fig 2B , control condition ) . However , we observed expression of mCherry in a few MOPV- and LASV-infected mDCs , indicating a low level of viral replication in a minority of mDCs during the two days following infection . We were also able to detect MOPV and LASV genomes both in the culture medium and in the cell pellet 24 hpi , with very low differences between MOPV and LASV , indicating that both viruses are able to enter into mDCs with a similar efficiency ( S1B and S1C Fig ) . This was confirmed by the low level of MOPV and LASV tagged-Z protein expression detected in mDCs 2 days post-infection ( dpi , S2C Fig ) . Together , these results suggest that there is no significant difference in the ability of LASV and MOPV to infect mDCs and in the lack of productive infection in these cells . We investigated the possible relationship between the IFN-I response and the lack of viral replication . We thus inhibited the autocrine and paracrine effects of the IFN-I response with neutralizing antibodies that target the IFNα/β receptor ( IFNAR ) , IFNα , and IFNβ . We treated mDCs with the anti-IFN cocktail and infected them with mCherry-expressing MOPV or LASV . mCherry was detected in the cells ( Fig 2B , anti-IFN condition ) , indicating that both viruses replicated . We confirmed this result by titrating infectious particles in the culture medium . Under the anti-IFN condition , viral titers increased from one to four dpi ( Fig 2A , black circles ) , with curves similar to those of classical replication kinetics of MOPV and LASV . Therefore , neutralizing the IFN-I response was sufficient to establish productive infection in mDCs . We demonstrated that mDCs are per se permissive to MOPV and LASV infection , but infection is inhibited by the early and robust IFN-I response . Our study of mDC responses to MOPV and LASV showed , so far , no significant differences between the two viruses . We thus used large-scale methods to obtain a deeper insight into these responses . We first focused on molecules released by mDCs in the culture medium using the Luminex assay ( Fig 3A ) . There was significantly more IFNα2 in the supernatants of MOPV-infected mDCs than those of LASV-infected mDCs . Chemokines ( MCP-1 , MCP-3 , and IP-10 ) and growth factors ( VEGF , G-CSF , and GM-CSF ) were produced by MOPV- and LASV-infected mDCs . Many pro-inflammatory cytokines were also upregulated in both MOPV- and LASV-infected mDCs: IL-5 , IL-6 , IL-10 , IL-12 , IL-15 , and TNFα . In our study , IL-6 , IL-15 , VEGF , and MCP-3 were particularly interesting , because they were produced in higher quantities by LASV-infected than MOPV-infected mDCs . Next , we used a transcriptomic approach to obtain an overview of mDC gene transcription . The differential expression of genes in MOPV- or LASV-infected mDCs relative to uninfected mDCs is presented in Fig 3B . The central column ( MOPV/LASV ) corresponds to the differential expression of genes between MOPV-infected mDCs and LASV-infected mDCs . The plotted genes show significant differences in expression between MOPV-infected mDCs and LASV-infected mDCs . Globally , genes up ( or down ) regulated in MOPV-infected mDCs relative to mock-infected mDCs were also up ( or down ) regulated in LASV-infected mDCs . However , the fold change of expression was generally higher for MOPV than LASV . Thus , the genes upregulated by MOPV were statistically more highly induced than those by LASV . As expected , many genes differentially regulated between MOPV- and LASV-infected mDCs were related to the immune response: IFN ( IFNα6 , IFNα8 , IFNα10 , and IFNω1 ) , pro-inflammatory ( IL15 , IL15RA , SIGLEC7 , and AXL ) , and cell-cell communication ( CD101 , CD200 , HLA-W , MS4A4A , and HOMER3 ) . We also identified genes involved in cell growth and death ( Fas , BCL2L14 , PRAME , and PID1 ) , adhesion and mobility ( SYNPO2 , DMD , COL23A1 , and VCAM1 ) , or both ( ANTXR1 , AXL , and SDC2 ) . More surprisingly , metabolic genes were regulated , especially those for lipid metabolism ( ELOVL7 , SCD5 , METRNL , LRP5 , APOE , APOC1 , and FUCA1 ) , oxidative molecule regulation ( SPSB4 , DDAH2 , and NOXA1 ) , and ionic transport ( SLCO5A1 , TMEM150C , CLIC2 , FAM20C , and TNNI2 ) . These differences had an important effect on the global activation state . Indeed , various pathways were regulated in LASV- and MOPV-infected mDCs relative to mock-infected mDCs ( Fig 3C ) . Most of them were related to the immune response , showing that mDCs responded to both viruses . However , there were substantial differences between MOPV- and LASV-infected mDCs at the pathway level . Immunity-related pathways , such as Pattern Recognition Receptors or RIG-I Like Receptor signaling , were upregulated in MOPV-infected mDCs relative to LASV-infected mDCs ( Fig 3C ) . DC maturation was also upregulated by MOPV relative to LASV . This difference was mostly caused by the upregulation of IFN-I , CD40 , CD83 , CCR7 , IL-6 , NFκB ( p50 ) , MDA5 , and Fas genes in MOPV-infected mDCs , but not LASV-infected mDCs ( S1D Fig ) . Overall , these results highlight differences in the activation state of MOPV- and LASV-infected mDCs . DCs are involved in the innate immune response . However , a major role of mDCs in vivo is to induce a specific cellular response . We developed a coculture model with mDCs ( infected or not ) and autologous T cells ( CD4 and CD8 ) . In these cocultures , the amount of MOPV and LASV genomes in the culture medium and in the cells decreased over time , suggesting an absence of viral replication ( S2A and S2B Fig ) . To verify this hypothesis , we stained purified mDCs , purified T cells and mDC-T cocultures for intracellular Z protein . We found no positive T cells ( S3 Fig ) , indicating that viral particles do not enter CD4 and CD8 T cells . mDCs infected by MOPV and LASV had a detectable level of intracellular Z , but only when cultured without T cells ( S2C Fig ) . Considering the low intensity of the Z staining compared to permissive A549 cells , the detected Z proteins more likely come from internalized particles than from new virions . We then investigated the response of mDCs using this coculture model . MOPV induced substantial IFN-I and CXCL10 production by mDCs at 24 hpi , whereas LASV did not ( Fig 4A ) . At 2 dpi , IFN-I and CXCL10 were upregulated under both MOPV and LASV conditions ( S4A Fig ) . MOPV-infected mDCs upregulated the activation markers CD40 , CD80 , CD83 , and CD86 ( 48 hpi ) , but LASV-infected mDCs did not ( Fig 4B ) . These results were surprising , as they did not match those obtained with infected mDCs alone ( Fig 1 ) . Fig 4C shows the IFN-I response at 24 hpi for MOPV- and LASV-infected mDCs ( mDC alone , results from Fig 1A ) and mDCs in coculture ( mDC-T coculture , results from Fig 4A ) . Under uninfected conditions , IFN-I mRNA levels were 10-fold lower in coculture than when mDCs were cultured alone . This difference was expected because mRNA levels were normalized to those of GAPDH ( present in all cells ) and IFN-I is mostly produced by mDCs ( 10% of the cells ) . MOPV infection of mDCs induced a 10- ( IFNα1 ) to 80- ( IFNα2 ) fold increase in IFN-I production relative to that of mock-infected cells when cocultured with T cells . Remarkably , coculture with T cells completely inhibited IFN-I synthesis by LASV-infected mDCs: alone , IFN-I production was comparable to that of MOPV-infected mDCs , and in coculture it was comparable to that of the mock condition . Altogether , these results show that mDC activation and the IFN-I response to MOPV were increased and prolonged in coculture , whereas activation and the IFN-I response to LASV were reduced and delayed . This suggests crosstalk between mDCs and T cells , which modulate the mDC response to MOPV and LASV . Lastly , we studied T cell responses in our model using flow cytometry . We quantified CD69 , a marker of early T cell activation , 2 dpi ( Fig 5A ) . CD69 was upregulated in both CD4 ( Fig 5B ) and CD8 ( Fig 5C ) T cells under conditions of MOPV infection , but not those of LASV infection . By 12 dpi , CD4 and CD8 T cells cultured with MOPV-infected mDCs expressed higher amounts of the cytotoxic molecules perforin and granzyme B ( GrzB , Fig 5B–5D ) . Such a cytotoxic phenotype was not induced by LASV-infected mDCs . At 15 dpi , there was slight upregulation of perforin and GrzB after coculture with LASV-infected mDCs , but it was still lower than that for the coculture with MOPV-infected mDCs ( S4B and S4C Fig ) . Ki67 expression by CD4 and CD8 T cells was also upregulated by MOPV-infected mDCs , suggesting their proliferation . Therefore , MOPV-infected mDCs induced activation , a cytotoxic response , and proliferation of CD4 and CD8 T cells . LASV-infected mDCs poorly activated T cells . We searched for factors that could influence the mDC response at early time points . The MOPV coculture was characterized by upregulation of IFNγ , Fas ligand ( FasL ) , IL-15 , TNFβ , and TRAIL ( Fig 5E ) . None of these molecules were upregulated in the LASV coculture , and they may play a role in the modulation of mDC responses . Interestingly , IL-18 was downregulated under conditions of LASV infection relative to those of mock or MOPV infection . After having investigated the differences between the responses of MOPV- and LASV-infected mDCs in coculture , we assessed the role played by viral factors . We designed MOPV/LASV chimeras by swapping the different viral proteins between the viral backbones ( Fig 6A ) . We obtained MOPV in which its GP , NP , or intergenic regions of the S segment ( IGRS ) were replaced by their LASV counterparts , and the corresponding LASV containing MOPV GP , Z , or IGRS . Successful exchanges of the ORFs were verified by next generation sequencing ( S5 Fig ) and western blot ( S6A Fig ) . We also included LASVNP ExoN , an LASV mutant with a non-functional exonuclease domain in its NP ( described in [24 , 25] ) . As expected , some chimeric viruses displayed a reduced and delayed growth in VeroE6 cells compared to wild type viruses ( S6B Fig ) . The genomes/infectious particles ratio varied between viral stocks , but with the exception of LASVNP ExoN , differences were rather low ( S6B and S6C Fig ) . We tested the chimeras in our mDC-T coculture model . Quantification of the IFN-I response allowed us to segregate the viruses into three groups: MOPV-like , LASV-like , and intermediate . In the MOPV-like group ( MOPVWT , MOPVGP LASV , LASVNP ExoN and LASVGP MOPV ) , infected mDCs produced large amounts of IFN-I mRNA . In the LASV-like group ( LASVWT , LASVZ MOPV and MOPVNP LASV ) , the IFN-I response was as low as under the uninfected condition . MOPVIGRS LASV- and LASVIGRS MOPV-infected mDCs produced small amounts of IFN-I and were designated as “intermediate” . It is worth noticing that mDC activation by the chimeric viruses in mDC-T coculture did not match their attenuation on VeroE6 cells: LASVNP ExoN and MOPVNP LASV were the most attenuated ones , whereas their phenotypes in mDC-T coculture were completely different . These results confirmed that the differences we observed in the responses are due to the activity of viral proteins and not to defects in viral replication and assembly . The viral titers did not detectably increase over time for any of the viruses , even though LASVWT titer seemed to decrease more slowly ( S6D Fig ) . Quantification of intracellular viral RNA at 1 and 2 dpi ( Fig 6B ) showed an increase in viral RNA for only LASV-like viruses . Inhibition of the IFN-I response appeared to be associated with viral replication , even at low levels , without viral particle release . Differences identified in the mDC response were also observed in the T-cell response ( Fig 7 ) . MOPV-like viruses induced over-expression of CD69 , perforin , and GrzB in CD4 ( Fig 7A ) and CD8 ( Fig 7B ) T cells . LASV-like viruses did not induce T-cell activation . The intermediate viruses showed low-level overexpression of CD69 , perforin , and GrzB . The T-cell response correlated with the IFN-I response of mDCs in coculture . This approach provided useful information on the immunogenic/immunosuppressive properties of viral proteins . MOPV Z protein did not alter the ability of LASV to escape from or suppress mDC and T-cell responses . Exchanging the IGRS between LASV and MOPV partially reversed their phenotype , suggesting that MOPV IGRS is more immunogenic than LASV IGRS . However , the immunogenicity of IGRS was not sufficient to explain the differences between MOPVWT and LASVWT . LASV GP did not appear to have immunosuppressive properties , as MOPVGP LASV exhibited a MOPV-like phenotype . In contrast , MOPV GP was sufficient to render LASV as immunogenic as MOPV . This immunogenicity was associated with the lack of viral replication of LASVGP MOPV ( Fig 6B ) . The most striking result of this experiment was the role of the NP . MOPVNP LASV behaved as LASVWT , showing that LASV NP was sufficient to completely abolish MOPV immunogenicity . The fact that LASVNP ExoN was as immunogenic as MOPV suggests that the immunosuppressive properties of LASV NP require a functional exonuclease domain . In summary , we showed that the inhibition of mDC and T-cell activation by LASV requires the expression of a functional NP with exonuclease properties . Our results also suggest that MOPV GP affects the ability of LASV to replicate in cells , influencing its immunogenicity . Finally , MOPV and LASV IGRS had different immunogenic properties .
We investigated the immune response parameters involved in the control of LASV infection or possibly associated with severe disease and death by studying the response of primary human mDCs to LASV and MOPV . We used MOPV infection to model non-fatal LF , as MOPV is very closely related to LASV , but non-pathogenic in humans [10 , 19 , 20 , 23] . We showed that mDCs were activated , acquired a mature phenotype ( CD83 expression ) , and produced high levels of IFN-I mRNA in response to both MOPV and LASV infection . mDC activation was associated with detection of viral genomes on and/or inside the cells , but not with a significant productive infection . These results were surprising , as MOPV is much more immunogenic than LASV . Previous studies showed that moDCs , which are often used as a model for mDCs , produced moderate amounts of IFN-I in response to MOPV , but not LASV infection [10 , 19] . In addition , moDCs were productively infected by MOPV and LASV , whereas mDCs were not . This difference is consistent with the fact that LASV replication in LPS-matured moDC was lower than that in immature cells [10] . We also demonstrated that the IFN-I response of mDCs to both viruses was sufficient to explain the absence of productive infection . We then conducted a large-scale study to exhaustively assess mDC responses to MOPV and LASV . We identified cytokines produced by MOPV- and LASV-infected mDCs , most which are found in patients with terminal VHF [26] . Five of these proteins were produced in different amounts by MOPV-infected and LASV-infected mDCs: IFNα2 , IL-6 , IL-15 , VEGF , and MCP-3 . IFNα2 levels were significantly higher in MOPV-infected mDCs than LASV-infected cells . This result was consistent with the non-significant differences observed in IFNα2 , IFNα6 , and IFNα8 mRNA levels between MOPV- and LASV-infected mDCs . The other four proteins were produced in higher amounts during LASV infection . IL-6 is known to be associated with severe cases of VHF [27] and IL-15 production has been correlated with severe cases in CCHF patients [28] . VEGF has been associated with plasma leakage and severity in dengue hemorrhagic fever [29] , and MCP-3 has recently been described in Ebola virus-infected NHPs [30] . These factors could all have a role in the pathogenesis of LF , even though DCs are not necessarily the main cell type producing them in vivo . MOPV infection strongly altered the transcriptomic state of mDCs . Most of the genes up ( or down ) regulated during MOPV infection were also up ( or down ) regulated during LASV infection , but at much lower levels . Many genes differentially regulated between MOPV- and LASV-infected mDCs were linked to immunity . We also identified genes involved in cell mobility and survival and , more surprisingly , ion and lipid metabolism . The relationship between lipids and VHF has been highlighted in other studies: lipids in serum are associated with febrile illness in LF patients [31] and the LXR/RXR pathway , which integrates lipid metabolism and immune functions , is necessary for dengue virus replication [32] . These differentially regulated genes had an impact on a larger scale: pathways linked to the innate immune response , such as PRR recognition , RIG-I like receptors , PI3K signaling , and DC maturation , were significantly upregulated in MOPV-mDCs relative to LASV-infected mDCs . The regulation of these pathways mostly relied on the same set of genes , including IFN-I , NFκB , CD40 , and CD80 . Such pathway analysis showed that regulated genes had an important effect on the global activation state of the cells . The main observation was that immunity-related genes and pathways were more highly up-regulated during MOPV infection than LASV infection . Altogether , the differences in transcriptomic regulation confirmed the higher immunogenicity of MOPV than LASV . In vivo , the ability of mDCs to induce specific T-cell responses could be critical during LASV infection . Indeed , the control of LF seems to rely mostly on robust T-cell responses [9 , 33] . We developed a mDC/T cell coculture system to measure the immunogenicity of MOPV- and LASV-infected mDCs through their ability to induce specific T-cell responses . When cultured with T cells , neither MOPV nor LASV infection seemed to be productive . However , the presence of T cells affected the amount of Z detected inside mDCs 2 to 8 dpi . Our hypothesis is that interaction with T cells accelerated the recycling of major histocompatibility complexes , and therefore the degradation of viral antigens . In this system , MOPV-infected mDCs produced even higher amounts of IFN-I than alone and induced T cell activation , proliferation , and acquisition of a cytotoxic phenotype . On the contrary , LASV-infected mDCs were not activated , did not produce IFN-I , and poorly activated T cells . These differences in T-cell activation are similar to those observed for moDCs . However , mDCs appeared to be much more immunogenic than moDCs: a single stimulation with mDCs was sufficient to activate T cells , whereas multiple stimulation with moDCs were required [20] . Therefore , moDCs are a relevant model to study T-cell activation in the context of MOPV and LASV infection , but not the innate immune response . In this study , T cells primed with MOPV-infected moDCs acquired the ability to specifically release GrzB in the presence of MOPV-infected cells , controlling viral replication [20] . Although we were unable to test this in our model , T cells expressing GrzB and perforin probably had the ability to lyse MOPV-infected cells . Indeed , naïve T cells do not contain cytotoxic granules . GrzB and perforin expression only occur after TCR stimulation and T-cell proliferation [34 , 35] . T cells primed with MOPV-infected mDCs expressed GrzB and perforin and were thus most likely fully functional cytotoxic T cells . The robust and efficient T-cell responses induced by MOPV-infected mDCs may result from the strong observed transcriptional activation of mDCs . In contrast , LASV-infected mDCs showed lower , but significant , activation and failed to induce the activation and proliferation of T cells . These observations correlate with the high immunogenicity and non-pathogenicity of MOPV , and low immunogenicity and high pathogenicity of LASV . T cells potentiated the activation of MOPV-infected mDCs , resulting in a 10-fold increase in IFNα release relative to that of mDCs alone . On the contrary , T cells had an inhibitory effect on LASV-infected mDCs , abrogating the synthesis of IFN-I mRNA and the expression of co-stimulatory molecules observed in LASV-infected mDCs . This suggests early crosstalk between mDCs and T cells , resulting in bilateral activation during MOPV-infection and mDC suppression during LASV-infection . The crosstalk between mDCs and T cells could occur through direct contact or soluble mediators . IFNγ , IL-15 , TNFβ , FasL , and TRAIL mRNA were overexpressed in the presence of MOPV-infected cells relative to uninfected and LASV-infected cells . These molecules , mainly released by activated T cells , may have been involved in the strengthening of mDC and T cell activation . However , NK cells could have participated in the release of IFNγ and TRAIL , as these cells were still present in the cultures . In the presence of LASV , the level of these T cell-derived mRNAs did not increase . Levels of IL-18 , an APC-derived cytokine , were even significantly lower , confirming the inhibition of the mDC response to LASV . Further studies are needed to identify the mechanism involved in this negative activation loop between mDCs and T cells . We generated MOPV/LASV chimeras to identify viral parameters responsible for the differences in T cell activation by MOPV- and LASV-infected mDCs . MOPVGP LASV , LASVNP ExoN , and LASVGP MOPV showed similar immunogenicity as MOPVWT . LASVZ MOPV and MOPVNP LASV were as poorly immunogenic as LASVWT . MOPVIGRS LASV and LASVIGRS MOPV displayed an “intermediate” phenotype . The segregation of viruses in these three phenotypes did not correlate with their attenuation on VeroE6 , cells , so the activation of mDC-T cocultures by the chimeras cannot be explained by assembly defects . The “LASV-like” viruses were the only ones that showed productive infection . The low activation of mDCs and limited production of IFN-I probably allowed some viral replication , which was still quickly controlled 2 dpi when IFN-I was synthesized . The impact of the IGRS on immunogenicity presumably relied on a difference of structure: the MOPV IGRS has a duplicated hairpin structure , whereas that of LASV has only one [36] . The double hairpin is most likely a better stimulator of intracellular sensors than a single hairpin , and thus more immunogenic . This could explain the intermediate phenotype of IGRS chimeras . Nevertheless , this difference between their IGRS was not sufficient to explain the differences in immunogenicity between MOPVWT and LASVWT , confirming that recognition by intracellular sensors is not the only mechanism involved . We also showed that MOPV Z protein was not a modulator of immunogenicity . We were unable to test LASV Z properties in our model , as MOPVZ LASV was not viable . However , the removal of LASV Z did not affect LASV immunogenicity , indicating that it does not have major immunosuppressive properties . Previous studies suggested that the Z protein of some arenaviruses is able to bind RIG-I and inhibit its signaling , resulting in an inhibition of IFN-I response . One of these studies showed that the Z proteins of New World arenaviruses , but not Old World arenaviruses , inhibit RIG-I signaling , which is coherent with our findings [37] . A more recent study concluded that only the Z proteins of pathogenic arenaviruses , including Old World arenaviruses , present this immunosuppressive function , suggesting that LASV Z ( but not MOPV Z ) is immunosuppressive [38] . However , we did not find in our mDC model any evidence for an immunosuppressive role of either LASV or MOPV Z protein . These discrepancies could be due to the use of different models . Indeed , additional sensors other than RIG-I could contribute for type I IFN synthesis in response to arenavirus infection in human mDC and therefore could conceal the effect of LASV Z protein on RIG-I . MDA5 would be a possible candidate for such an hypothesis . LASV GP did not appear to have major immunosuppressive properties . Indeed , MOPVGP LASV was almost as immunogenic as MOPV and did not replicate . In contrast , LASV became as immunogenic as MOPV when it expressed MOPV GP . It is possible that MOPV and LASV GP address virions to different cellular compartments , influencing their ability to escape or trigger innate immunity and , consequently , viral replication . Their possible use of different fusion receptors , Lamp1 for LASV and unidentified for MOPV , may also be involved [39] . We are currently investigating this possibility . Finally , our results showed that LASV NP was sufficient to abolish the immunogenicity of MOPV-infected mDCs . This shows that LASV NP has major immunosuppressive properties , which require its exonuclease domain . The immunosuppressive properties of arenavirus NP , especially of its exonuclease domain , have already been largely described . This domain is required for LASV NP to inhibit the IRF pathway [40] . LASV detection by RIG-I is inhibited through the digestion of double-stranded RNA by LASV NP , and this mechanism is probably conserved among most arenaviruses [25 , 41 , 42] . However , the discrepancy between the ability of MOPV and LASV to modulate immunogenicity is surprising , as both NPs contain an active exonuclease domain [43] . It is possible that the exonuclease activity of LASV NP is more potent than that of MOPV . Another possibility is that the digestion of dsRNA is not the only function of this domain . The ability of LASV NP to bind IKKε may be involved , as this interaction is associated with the exonuclease domain [44] . LASV NP also interacts with the PACT complex , inhibiting PACT-dependent activation of RIG-I [45] . This interaction is abolished when the exonuclease domain is mutated . NP-PACT interaction has been shown for Junin , Machupo , Tacaribe and Pichinde , but MOPV was not included in this study . Other cellular proteins interacting with LASV NP have been identified , but not yet described , and could be involved in its immunosuppressive properties [46] . Unfortunately , it was not possible to study the impact of MOPV NP on LASV immunogenicity , because LASVNP MOPV was not viable . In conclusion , this ex vivo approach showed that primary human mDCs respond well to both MOPV and LASV . However , only MOPV-infected mDCs were activated and able to induce T-cell responses when cultured with T cells . We demonstrated that NP and GP are viral factors involved in the respective immunogenic and immunosuppressive properties of MOPV and LASV . Duplicated hairpins in the S segment , different cell addressing by GP , and less efficient immunosuppressive factors , such as NP , may result in efficient triggering of RNA sensors by MOPV , explaining its strong immunogenicity . In contrast , multiple and efficient immunosuppressive mechanisms , mediated by NP and Z , and a lower ability of viral RNA to stimulate cell receptors are probably responsible for the poor immunogenicity of LASV . The comparison of LASV and MOPV is used to model the differences between fatal and non-fatal LF in vitro . In vivo , the factors that determine the outcome of LF are still unknown . Factors involved in different mDC responses are good candidates , as mDCs may initiate the entire immune response in humans . Host polymorphisms , initial viral load , and the route of infection may affect LASV replication in mDCs , the detection of LASV by intracellular sensors , and the inhibition of mDC activation by viral proteins . Such factors would be critical for the survival of the host , as modulating the mDC response would influence the initiation of T-cell responses .
Human peripheral blood was obtained from healthy donors with informed consent and was provided by the Etablissement Français du Sang ( Lyon , France , agreement PLER/1‐1820‐05/05/14 ) . Written informed consent was provided by all study participants . VeroE6 cells were grown in DMEM supplemented with 0 . 5% penicillin-streptomycin and 5% fetal bovine serum ( FBS , all from Invitrogen ) . A549 cells were grown in DMEM supplemented with 0 . 5% penicillin-streptomycin , 5% FBS and 1% Hepes ( all from Invitrogen ) . Mopeia ( AN21366 strain [16] ) and Lassa ( AV strain [47] ) viruses were grown in VeroE6 cells at 37°C , with 5% CO2 . Viral supernatants were harvested and used as the virus stock and the absence of mycoplasma was confirmed . LASV and MOPV titers were determined by plaque immunoassays as described below . All experiments with LASV were carried out in biosafety level 4 facilities ( Laboratoire P4 Jean Merieux-Inserm , Lyon ) . The reverse genetics systems of LASV and MOPV rely on a four-plasmid strategy , as described in [24] , with the rescue procedure of recombinant viruses . Lassa virus with a non-functional exonuclease domain ( LASVNP ExoN ) was also obtained by reverse genetics and its characteristics are described in [24 , 25] . We obtained LASV and MOPV chimeras by first generating plasmids coding for the S or L segment and depleted for an ORF ( or intergenic region of the S segment , IGRS ) . Briefly , plasmids encoding the S or L segment were amplified with primers , ( i ) allowing the complete amplification of the plasmid , except the target ORF and ( ii ) flanked with BsmBI restriction sites downstream of the start and upstream of the stop codons of the deleted ORF . PCR products were ligated and depleted segments used for the cloning of the heterologous ORF ( or IGRS ) . LASV-mCh and MOPV-mCh were modified to express the mCherry and LASV/MOPV NP proteins from a single gene . We used plasmids coding for the S segment of LASV or MOPV with a depleted NP ( as described above ) . An insert containing the mCherry and NP ORFs , separated by a P2A self-cleavage site , was generated by overlapping PCR . This insert was cloned into the LASVΔNP or MOPVΔNP plasmid . MOPV and LASV with a FLAG-tagged Z protein were also obtained by reverse genetics . Sequences of all plasmid constructs were verified by sequencing . Autocrine and paracrine effects of the IFN-I response were neutralized ( Fig 2 ) by treating mDCs with cocktails of neutralizing antibodies . The anti-IFN cocktail contained antibodies neutralizing IFNAR ( MMHAR-2 , 5 μg/mL ) , IFNα ( MMHA-2 , 2 . 5 μg/mL ) , and IFNβ ( MMHB-3 , 2 . 5 μg/mL ) , all from PBL Assay Science . In the control condition , the cocktail contained corresponding control isotype antibodies , IgG1 ( 5 μg/mL ) and IgG2a ( 5 μg/mL ) , both from ThermoFisher Scientific . One dose of antibody cocktail was administrated to the cells at the time of infection . Half a dose was added to the cell culture 1 and 3 dpi . Cells were infected at a MOI = 0 . 1 by wild type or mCherry-expressing MOPV or LASV . For MOPVWT- and LASVWT-infected mDCs , small volumes of supernatant were harvested at various times post-infection and titrated . For mCherry-expressing MOPV and LASV , mCherry fluorescence was measured by fluorescence microscopy with Leica DMIRB . VeroE6 cells were infected with sequential dilutions of supernatant and maintained for six days with Carboxy-methyl-cellulose ( 1 . 6% ) ( BDH Laboratory Supplies ) in DMEM supplemented with 2% FBS . Infectious foci were detected by incubation with monoclonal antibodies directed against MOPV and LASV ( mAbs L52-54-6A , L53-237-5 and YQB06-AE05 , generously provided by Dr P . Jahrling , USAM- RIID , Fort Detrick , MD ) , followed by PA-conjugated goat polyclonal anti-mouse IgG ( Sigma-Aldrich ) . PBMCs were isolated by Ficoll ( GE Healthcare ) centrifugation from the blood of consenting healthy donors provided by the Etablissement Français du Sang ( Lyon , France ) . mDCs were isolated by negative selection using the Myeloid Dendritic Cell Isolation kit ( Miltenyi Biotech ) . mDCs were maintained in RPMI 1640 Glutamax I , 0 . 5% penicillin- streptomycin , 10 mM HEPES , 1% nonessential amino acids ( full RPMI ) , and 10% FBS ( all from Invitrogen ) . mDCs were infected at a MOI = 2 or treated with pIC ( Invitrogen ) at 150 μg/mL for a positive control of activation . For mDC-T cell coculture , autologous plasma ( AP ) was heated for 30 min at 56°C and centrifuged for 20 min at 1 , 200 x g before use . mDCs were isolated from 75 to 80% of the PBMCs ( as described above ) and T cells from 20 to 25% . For T cells , peripheral blood lymphocytes ( PBLs ) were isolated by centrifugation on 50% Percoll ( GE Healthcare ) in phosphate-buffered saline ( PBS ) . PBLs were washed three times in full RPMI supplemented with 4% AP . B cells were depleted using CD19 antibodies coupled to immunomagnetic beads ( Dynal ) . T cells and mDCs were maintained overnight in full RPMI supplemented with 1 mM sodium pyruvate ( Invitrogen ) and 10% AP . mDCs were harvested and infected at a MOI = 1 for 1 h . Supernatants of uninfected vero cells were used for the mock condition . Infected mDCs were added to T cells at a ratio of one mDC to 10 T-cells . For intracellular viral staining , purified mDCs , mDC-T coculture and A549 cells were infected with recombinant MOPV and LASV expressing Z-FLAG . mDCs were infected at MOI = 1 and cultures alone or with T cells at a ratio of one mDC to 10 T-cells . A549 cells were infected at MOI = 0 . 1 . Replicates of all experiments were performed with blood from different donors . To quantify cellular mRNA , cells were harvested and centrifuged . RNA from the cell pellets were extracted using the RNeasy kit and DNAse I digestion ( both from Qiagen ) , followed by a second DNAse digestion ( Ambion ) . cDNAs were randomly reverse transcribed using SuperScript III reverse transcriptase , Oligo ( dT ) 12–18 primers , and RNAse OUT ribonuclease inhibitors ( all from Invitrogen ) . Cellular genes expression was assessed using the Taqman Universal master mix and Taqman commercial primers and probes for FasL , IFNα6 , IFNα8 , IFNγ , IL-15 , IL-18 , TNFα , TNFβ , TRAIL , and CXCL10 ( Life Technologies ) . For the IFN-I genes , we used the following primers and probes: 5’-GTGGTGCTCAGCTGCAAGTC-3’ ( sense ) , 5’-TGTGGGTCTCAGGGAGATCAC-3’ ( antisense ) and 5’-AGCTGCTCTCTGGGC-3’ ( probe ) for IFNα1; 5’-CAGTCTAGCAGCATCTGCAACAT-3’ ( sense ) , 5’-GGAGGGCCACCAGTAAAGC-3’ ( antisense ) and 5’-ACAATGGCCTTGACCTT-3’ ( probe ) for IFNα2; 5’-TCTCCACGACAGCTCTTTCCA-3’ ( sense ) , 5’-ACACTGACAATTGCTGCTTCTTTG-3’ ( antisense ) and 5’-AACTTGCTTGGATTCCT-3’ ( probe ) for IFNβ . GAPDH mRNAs were amplified using commercial primers and probes ( Applied Biosystems ) to normalize the results . Relative mRNA levels were calculated as 2−ΔCt , with Ct the cycle threshold and ΔCt = [gene Ct]–[GAPDH Ct] . For viral genome quantification , viral RNAs were extracted from culture supernatants using the QIAamp Viral RNA Mini Kit ( Qiagen ) . Viral genomes were quantified using the EurobioGreen qPCR Mix Lo-ROX ( Eurobio ) . Primers targeted MOPV NP ( 5’-CTTTCCCCTGGCGTGTCA-3’ and 5’-GAATTTTGAAGGCTGCCTTGA-3’ ) or LASV NP ( 5’-CTCTCACCCGGAGTATCT-3’ and 5’-CCTCAATCAATGGATGGC-3’ ) . We transcribed the pGEM plasmid , coding for fragments of the MOPV and LASV NP gene ( including the sequence amplified by PCR ) , in vitro , using the Riboprobe in vitro Transcription System ( Promega ) , to obtain RNA standards . Viral genomes were quantified ( by copy number ) by comparing our samples with sequential dilutions of these standards . All runs were performed in duplicate using a LightCycler480 ( Roche ) . mDCs were infected at a MOI = 1 by LASV or MOPV or were uninfected and incubated for 24 h at 37°C , 5% CO2 . Cells were centrifuged and RNA from the cell pellets extracted using the RNeasy kit and DNAse I digestion ( both from Qiagen ) , followed by a second DNAse digestion ( Ambion ) . Experiments were replicated three times with cells from different donors . RNA quantity and quality were assed using the Agilent RNA 6000 nano kit and 2100 Bioanalyzer ( Agilent Technologies ) . Transcriptome sequencing and bioinformatic analyses were performed by GATC ( Constance , Germany ) . cDNAs were synthesized by random priming of poly-A RNAs . Pair-end , 2x50 bp read-length illumina sequencing was performed on cDNA libraries , with a minimum of 30 million read pairs per sample . Pathway analysis was performed with Ingenuity Pathway Analysis software ( Qiagen ) and heatmaps were made using R . For mDC activation , mDCs were harvested 24 hpi , washed , and the pellets suspended in PBS complemented with 5% pooled human plasma . We incubated cells for 30 min at 4°C with Lin1-FITC , CD83 ( HB15e ) -PE , CD40 ( 5C3 ) -APC-H7 ( BD Biosciences ) , CD11c ( BU15 ) -PeCy5 , CD86 ( HA5 . 2B7 ) -PeCy7 , CD80 ( MAB104 ) -APC-AlexaFluor750 , HLADR ( Immu-357 ) -KromeOrange ( Beckman Coulter ) , and/or CD253 ( RIK2 . 1 ) -APC ( Miltenyi Biotech ) . For mDC-T coculture , mDC activation was assessed at 48 hpi ( using the same protocol as for mDCs alone ) . T-cell activation was assessed 2 , 12 , and 15 dpi . Cells were harvested , washed , and the pellets suspended in PBS complemented with 5% pooled human plasma . We incubated cells for 30 min a 4°C with CD4 ( SFCI12T4D11 ) -PECy7 , CD3 ( UCHT1 ) -KromeOrange ( Beckman Coulter ) , CD69 ( FN50 ) -PeCy7 , CD4 ( RPA-T4 ) - AlexaFluor647 , and/or CD8 ( RPA-T8 ) -BV421 ( BD Biosciences ) . The expression of intracellular proteins was analyzed by treating the cells with the FoxP3 Staining Buffer Set and FcR Blocking Reagent , human ( Miltenyi Biotech ) , according to the manufacturer’s instructions , and incubating them with Perforin ( δG9 ) -FITC , GrzB ( GB11 ) -PE , and/or Ki67 ( B56 ) -PerCPCy5 . 5 ( BD Biosciences ) . For staining of intracellular virus , purified mDCs and mDC-T cocultures were harvested at 2 , 5 , and 8 dpi and centrifuged . The cell pellets were suspended in PBS complemented with 5% pooled human plasma and incubated for 30 min a 4°C with Lin1-FITC , CD8 ( RPA-T8 ) -BV421 ( BD Biosciences ) , CD11c ( BU15 ) -PeCy5 , CD4 ( SFCI12T4D11 ) -PECy7 , CD3 ( UCHT1 ) -KromeOrange and/or HLADR ( Immu-357 ) -KromeOrange ( Beckman Coulter ) . Cells were treated with the FoxP3 Staining Buffer Set and FcR Blocking Reagent , human ( Miltenyi Biotech ) , according to the manufacturer’s instructions , and incubated with anti-DYKDDDDK-APC antibody ( Miltenyi Biotech ) . The fluorescence of paraformaldehyde-fixed cells was measured using a Gallios flow cytometer ( Beckman Coulter ) . Data were analyzed using Kaluza software ( Beckman Coulter , version 1 . 2 ) . DCs were gated as Lin1-/HLA-DR+ cells , and CD11c was used to confirm their phenotype . After phenotypic selection , based on FSC/SSC , CD4 and CD8 T cells were gated as CD3+/CD4+ and CD3+/CD8+ cells , respectively . See S7 Fig for more details . mDCs were harvested 24 hpi and the culture medium collected . Fifty cytokines were quantified in the samples using the Milliplex map kit Human Cytokine/Chemokine Magnetic Bead Panel ( PX38 ) and Human Cytokine/Chemokine Magnetic Bead Panel IV ( Merck Millipore ) . Runs were performed with a Magpix Luminex ( Merck Millipore ) . VeroE6 cells infected with MOPV and LASV chimeras ( MOI = 0 . 01 ) were lysed in Laemmli buffer ( Bio-Rad ) at 4 dpi . Heat-denaturated proteins were loaded and separated on 4–15% gradient precast gels and transferred onto PVDF membranes before staining . Samples were immunoblotted with primary antibodies against LASV GP1 , LASV NP or LASV/MOPV Z , anti-mouse or anti-rabbit antibody conjugated to peroxydase ( Jackson ImmunoResearch ) and SuperSignal West Dura Extended Duration Substrate ( ThermoFisher Scientific ) . Actin was used as a control for protein extraction and staining . The mean and standard error of the mean ( SEM ) for each set of data were calculated using R . Graphs were generated using SigmaPlot ( SyStat Software Inc ) . Microscopy images were analyzed with AxioVision software ( Zeiss , version 4 . 9 ) . Results between the various infection conditions ( mock , pIC , MOPV , LASV … ) were compared by performing Wilcoxon tests with R . Replication kinetics ( in Fig 2A ) were compared performing Repeated-Measures ANOVA followed by All Pairwise Multiple Comparison Procedures ( Holm-Sidak method ) using SigmaPlot . Differences between two sets of data were considered to be significant for p < 0 . 05 .
|
Lassa fever is a viral hemorrhagic fever and a major public health issue in West Africa . Lassa virus , the causative agent of Lassa fever , is listed by the World Health Organization as one of the emerging pathogens likely to cause severe outbreaks in the near future . Indeed , there is currently no vaccine and no treatment against Lassa virus . Determinants of Lassa virus high pathogenicity are not completely understood . However , it has been shown that rapid type I interferon response and efficient T cell response were critical to survive Lassa fever . Dendritic cells are at the crossroads of innate and adaptive immunity . Their direct response to viral infection includes type I interferon production . They can also present viral antigens , initiating the T cell responses . We decided to investigate how dendritic cells respond to Lassa virus to evaluate their importance in the global immune response . We showed that primary human myeloid dendritic cells are activated by Lassa virus infection , and produce type I interferon . However , Lassa virus-infected dendritic cells were not able to activate T cells . We also elucidated the roles of viral proteins in the modulation of dendritic cell responses .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"blood",
"cells",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"immune",
"physiology",
"gene",
"regulation",
"antigen-presenting",
"cells",
"drugs",
"immunology",
"microbiology",
"immunosuppressives",
"dendritic",
"cells",
"cytotoxic",
"t",
"cells",
"pharmacology",
"immune",
"system",
"proteins",
"white",
"blood",
"cells",
"animal",
"cells",
"proteins",
"antigens",
"gene",
"expression",
"t",
"cells",
"viral",
"replication",
"immune",
"response",
"biochemistry",
"cell",
"biology",
"virology",
"physiology",
"genetics",
"biology",
"and",
"life",
"sciences",
"cellular",
"types"
] |
2018
|
Lassa virus activates myeloid dendritic cells but suppresses their ability to stimulate T cells
|
While it is known that a large fraction of vertebrate genes are under the control of a gene regulatory network ( GRN ) forming a clock with circadian periodicity , shorter period oscillatory genes like the Hairy-enhancer-of split ( Hes ) genes are discussed mostly in connection with the embryonic process of somitogenesis . They form the core of the somitogenesis-clock , which orchestrates the periodic separation of somites from the presomitic mesoderm ( PSM ) . The formation of sharp boundaries between the blocks of many cells works only when the oscillators in the cells forming the boundary are synchronized . It has been shown experimentally that Delta-Notch ( D/N ) signaling is responsible for this synchronization . This process has to happen rather fast as a cell experiences at most five oscillations from its ‘birth’ to its incorporation into a somite . Computer simulations describing synchronized oscillators with classical modes of D/N-interaction have difficulties to achieve synchronization in an appropriate time . One approach to solving this problem of modeling fast synchronization in the PSM was the consideration of cell movements . Here we show that fast synchronization of Hes-type oscillators can be achieved without cell movements by including D/N cis-inhibition , wherein the mutual interaction of DELTA and NOTCH in the same cell leads to a titration of ligand against receptor so that only one sort of molecule prevails . Consequently , the symmetry between sender and receiver is partially broken and one cell becomes preferentially sender or receiver at a given moment , which leads to faster entrainment of oscillators . Although not yet confirmed by experiment , the proposed mechanism of enhanced synchronization of mesenchymal cells in the PSM would be a new distinct developmental mechanism employing D/N cis-inhibition . Consequently , the way in which Delta-Notch signaling was modeled so far should be carefully reconsidered .
Adaption to the day-and-night-cycle on earth provides an evolutionary advantage for organisms that can adjust their gene activity to this 24-hour rhythm . Therefore many metabolic processes show a circadian periodicity because they are all controlled by a GRN forming the so-called circadian clock [1] . Shorter period oscillators are called ultradian [2] . Some play an important role in the embryonic process of somitogenesis , where the vertebrae-precursors , the somites , bud off with a species-specific periodicity at the anterior end from a mesenchymal tissue on both sides of the notochord , the presomitic mesoderm . For mice this period is with two hours much shorter than circadian . The core of the somitogenesis clock , first simulated in a computer model by Meinhardt [3] , is set up in probably all vertebrate species by the Hes/Hairy/her gene families [4] , which are negative feedback oscillators . A short decay time for the gene products together with a long enough time delay between gene expression and binding of the protein on its own gene promoter results in oscillatory gene expression . In mice the Hes1 , Hes5 and Hes7 genes ( and many others connected to them in an intricate network ) were found to oscillate in the PSM [5] . Hes1 , which also oscillates in neural progenitors [6] , could be stimulated to oscillate with a two-hour period in vitro in fibroblasts , neuroblasts , myoblasts and other cell types [7] . In the anterior unsegmented PSM of mice , also called wave zone , Hes7 needs additional activation by D/N signaling to maintain oscillatory gene expression [8] . The D/N pathway works by juxtacrine signaling: Membrane-anchored DELTA or JAGGED ligands of a signal-sending cell bind to NOTCH receptors embedded in the cell membrane of an adjacent cell . This induces a proteolytic cleavage of the NOTCH receptor and releases the intracellular domain of NOTCH ( NICD ) into the cytoplasm , which then moves into the nucleus where it serves together with various co-factors as transcription regulator and activates , among others , the Hes1/7 genes [9] . These events finally lead to a moving wave of NICD from posterior to anterior in the PSM . We proposed in our 2012 model that this wave is generated by the action of the posterior-to-anterior gradients of FGF8 and WNT3a on decay rates of the core oscillator consisting of D/N and Hes7 [10] . When the NICD wave comes to a halt in the anterior PSM , NICD determines together with TBX6 the expression of Mesp2 that induces the formation of a border between a forming somite and the remaining PSM [11] . Another important function of D/N signaling in somitogenesis is synchronization of the cellular oscillators in the PSM [12] , [13] , which requires cell-cell contact [14] . Without this synchronization somite formation is severely disturbed [14] . The synchronization of cellular oscillators was also examined theoretically , mostly for the zebrafish her1/7 system . Using delay differential equations , D/N signaling was able to synchronize two cells [15] or a row of cells [16] . However , if this system is expanded to 2-dimensional arrays of cells the short-range interaction of D/N causes different domains to be synchronized to different phases and no domain is able to conquer the others [17] , [18] . It was shown for zebrafish and chicken that cell movements in the posterior part of the PSM occur depending on the concentration of FGF8 [19] , [20] . Uriu et al . included these movements into simulations of the zebrafish PSM and could thereby demonstrate a much better synchronization of the her oscillators [18] . Later , this theory was extended to find an optimal rate for cell movements and to describe the effect of gradual recovery of intercellular coupling experienced by a cell after movement [21] . All these models assumed direct interaction between DLL1 and NOTCH1 when they are positioned in membranes of adjacent cells . However , Delta-ligand and Notch-receptor molecules can also interact within the endoplasmatic reticulum ( ER ) or cell membrane of the same cell [22] , [23] , which would lead to a fast clearance of the intracellular dimer . This mechanism , where Delta and Notch inhibit each other in the same cell , was therefore termed D/N-cis-inhibition . For example , D/N cis-inhibition is able to generate mutually exclusive signaling states in a mammalian cell-culture system [24] . Applied to computer simulations , D/N cis-inhibition leads to sharper and faster boundary formation during development of the Drosophila wing vein system and improves the equidistant distribution of bristle precursor cells by lateral inhibition [25] . Here , we propose another beneficial effect of D/N cis-inhibition: It accelerates in computer simulations the synchronization of D/N coupled ultradian oscillators and extends the parameter range wherein synchronization is possible without taking cell movements into account . Although experimentally not yet confirmed , the proposed mechanism of enhanced synchronization of mesenchymal cells in the PSM would be the third distinct developmental mechanism employing D/N cis-inhibition . Consequently , the way in which Delta-Notch signaling was modeled so far should be carefully reconsidered .
We employ the same cell- and gene-based simulation tool as described in [10] . The GRN incorporated in each virtual cell consisting of Hes7 , Delta1 , Notch1 , is shown in Fig . 1 A and in an advanced version including also Lfng in Fig . S1 . Oscillations are generated by a negative feedback of HES7 onto the Hes7 promoter with delay , which is simulated by the transport of proteins and mRNAs between the nucleus and cytoplasm similar to the transport-model by Uriu et al . [18] . Furthermore , the Hes7 oscillators are coupled by D/N signaling and we assume that HES7 acts on the Dll1 promoter as it was shown for HES1 [26] . The DLL1 ligand and NOTCH1 receptor are modeled with two compartments for the proteins ( cytoplasm and membrane ) and for their mRNAs ( cytoplasm and nucleus ) : Since we assume that Notch1 expression does not oscillate we do not differentiate between nucleus and cytoplasm in this case , because a mathematical description without delay for the mRNA is sufficient . Our model is designed for the simulation of mouse development , therefore the reaction rates are taken from literature or if not available adjusted to reproduce a mouse specific oscillation period of around 2–3 hours . However , our program allows other oscillation periods by simply rescaling all reaction rates in the differential equations – except those in the denominators – via its graphical user interface , which is equivalent to a rescaling of time . In addition to the reaction of DLL1 and NOTCH1 between neighboring cells leading to the release of NICD as transcription co-factor ( trans-activation ) , this work also considers the reaction of NOTCH1 and DLL1 in the membrane and cytoplasm of the same cell ( cis-interaction ) , which leads to their immediate decay – shown graphically in Fig . 1 B . So , the titration of one membrane protein against the other in each cell leads to an excess of either the ligand or the receptor and consequently determines whether the cell acts as a sender or receiver . Contrary to our previous work [10] , where every cell started with the same initial concentration values and received after mitosis the concentration values of its mother cell at their respective oscillation phases , here , all cells start with random initial values . To avoid that the cells start too far away from their limit cycle we add random values between zero and one multiplied to each of the initial concentration values used in [10] and scaled with a percentage value that gives a simple measure for the initial noise . For instance , 200% noise means: to each concentration its doubled value is added multiplied by a random number taken from the interval between zero and one . Our program allows for real time observation of the simulation , so synchronization can be easily observed by visual inspection . However , to get a quantitative measure for synchronization we introduced a simple correlation function that falls to zero when perfect synchronization is achieved and shows oscillatory behavior otherwise . In the case of anti-synchronization , the values of the correlation function display negative oscillations . Here stands for any concentration value of a gene product in cell k ( or i or j ) at time t and is the average concentration value . For each cell with index i its concentration is multiplied with the average concentration of its neighboring cells with index j , where N is the number of neighboring cells . A rectangular arrangement of cells results in N = 2 for 1 dimension , N = 4 for 2 dimensions , and N = 6 for 3 dimensions . For cells situated on an edge or corner the number of neighbors is reduced , i . e . we use not periodic boundary conditions in our simulations . So N in the formula above depends on cell index i , but we suppress this dependence to simplify the notation . Furthermore , the user can define an extended neighborhood , which means that e . g . in 2 dimensions the diagonal adjacent cells are counted as neighbors . If all cells are synchronized , ci ( t ) and cj ( t ) have the same value , which is equal to the average value . So , the difference in the first formula will become zero . For the evaluation of the correlation function we used Hes7 mRNA concentration in the cytoplasm if not stated otherwise . Although the correlation function uses only information about neighboring cells , it shows us synchronization by dropping to zero , because if each cell is synchronous to its neighbor , all cells are synchronized . Compared to the R-synchronization measure ( defined in the supplementary material Text S1 ) , which goes to one for perfect synchronization , the advantage of the correlation function C ( t ) is the observation that it becomes negative , if the configuration becomes anti-synchronized , i . e . one observes a salt-and-pepper pattern , which can be oscillating or not . See Fig . 2 first and last row for an example for each case . In our search for parameter values resulting in fast synchronization we observed in the case without D/N cis-inhibition that parameters that allowed for fast synchronization made the system unstable against anti-synchronization . After a period of almost perfect synchronization with C ( t ) almost exactly zero the system drifts slowly into an oscillating salt-and-pepper pattern with the difference between neighboring cells becoming ever larger . Unfortunately , the faster the synchronization , the shorter the duration of synchronized behavior before reverting into the anti-synchronized state . Because the correlation function allows us to see this behavior before it becomes visible by eye , it is very useful for interactively searching for parameters providing for fast synchronization . The effect of D/N cis-inhibition on synchronization of a 7×7×7 cell cube with 100% noise added is shown in Fig . 2 , where simulation snapshots are displayed for increasing strengths of D/N cis-inhibition . Clearly , D/N cis-inhibition accelerates synchronization , whereas without ( see movie S1 ) or small cis-inhibition the oscillator-system synchronizes badly and turns after some time into an anti-synchronized state , which was already described for a 2-cell [15] and a 2-dimensional system [17] ( see also supplementary movie S2 for the case of rDNcis = 0 . 01 ) . For intermediate ( 0 . 005 ) values of D/N cis-inhibition one observes incomplete synchronization . Large parts of the cube are synchronous but in different phases to each other so that ‘waves’ of expression moving over the cube volume can be observed . Increasing the D/N cis-inhibition strength leads to complete and ever faster synchronization with the best result achieved for 0 . 0115 . However , increasing D/N cis-inhibition further leads to a progressive damping of the oscillations . This non-oscillating state then turns slowly into a static salt-and-pepper pattern . So in this case we get the classic lateral inhibition case without oscillations . Simulation snapshots and the time course of our correlation function for systems with different dimensions are shown in Fig . 3 . Compared to the 3-dimensional simulation with a 7×7×7 cube of interacting cells , the synchronization of a 2-dimensional array of cells is slower and deviations from perfect synchronization are larger . Only if one reduces the noise amplitude to 60% , the initial deviations in the correlation function are comparable , but synchronization is still slower . A similar effect is observed for a 1-dimensional chain of cells . This can be explained by the nature of our model , where the effect of D/N signaling in the receiving cell is averaged over the number of its neighbors due to practical reasons . This has the advantage that one does not have to change all parameters in the network when dimensionality of the system is changed . Consequently , the noise one cell receives in D/N signaling reduces with the number of its neighbors because fluctuations are cancelled out better in summation with more neighboring cells sending noisy signals . This effect is also demonstrated in Fig . S2 , where a 3-dimensional array with 6 neighbors per cell gives comparable results to a 2 dimensional array with 8 neighbors per cell . Likewise , we analyzed the influence of cell number , i . e . the volume of a cell array , on synchronization and compared cubes with a length of 5 , 7 , 9 , 11 , and 14 cells ( Fig . S3 ) . While at the beginning the correlation functions vary due to the randomly chosen initial values , they decay in the further course of the simulation to very small values with a similar behavior . The same behavior can be observed also for the R-synchronization-measure , which quickly reaches values very near 1 , indicating very good synchronization , independently of the size of the cell cube . To explore the robustness of the system and the speed of D/N-mediated synchronization , with and without cis-inhibition , we performed an extensive scan over all parameters in a simple two-cell system . As expected , D/N cis-inhibition provides for faster synchronization of cells over a wide parameter range , independent from the chosen initial concentration values ( for details see supplemental text S1 ) . There are also parameter ranges where synchronization is not achieved with D/N cis-inhibition , if one looks at the R-synchronization measure . However , if one looks at the concentration time course behavior one sees that this downward trend of the R-function results from a progressive damping of the oscillations if one increases the Hes7 mRNA or protein decay rates more than ten percent , for instance . The influence of the system parameters on the amplitude ( minimal and maximal cytoplasmic HES7 expression ) of the cellular oscillator is shown in Fig . 3 of the supplemental text S1 . One can observe the strong dependence of the oscillator amplitude on Hes7 mRNA and protein decay rate , for instance , and that the cis-inhibition strength rDNcis abolishes the oscillation if it increases beyond 0 . 014 , as already seen in Fig . 2 . To examine the robustness of the system further we generated 40 parameter sets by randomly varying all production , transport , and decay rates within a range of plus-minus ten percent around our standard parameters and tested these parameters sets in a cube with an edge length of 7 cells with 100% initial noise added . 16 of the random parameter sets resulted in damped oscillation and of the 24 undamped oscillating systems 21 showed complete synchronization . Only for three parameter sets synchronization was not complete . Instead , expression waves were generated . Results for all oscillating parameter sets are shown Fig . S4 . The input files for running simulations with the different parameter sets are supplied in the supplemental material as file S1 ( Config-files . tar . gz ) . In our previous work on boundary formation in the PSM of mouse [10] we postulated a positive action of LFNG on D/N signaling . Likewise , we have extended our minimal model by Lfng , which is controlled by HES7 ( Fig . S1 ) . Here , the parameters chosen for the relative contributions of unaided D/N signaling and D/N interaction with LFNG-action have to meet two demands: ( i ) they should allow fast synchronization with D/N cis-inhibition , and ( ii ) they should reproduce the diminished oscillation amplitude observed experimentally in the mouse PSM when Lfng is non-functional [27] . These demands are fulfilled when we set the ratio of unaided to LFNG-promoted D/N reaction to about 1∶4 ( Fig . S5 ) . So far , all discussions on synchronization of ultradian oscillators by D/N signaling examined the static case , i . e . a non-growing tissue . However , a real test for synchronization would be a growing tissue , for example , the tail bud during somitogenesis ( Fig . 1 C ) . Therefore , we implemented D/N cis-inhibition in one of our models of somitogenesis , which is characterized by a growing tissue and a posterior-to-anterior FGF8 gradient controlling HES7 degradation [10] . When daughter cells inherit the concentration values of their mother cells and a 100 percent noise is added , we observed a clear difference between simulations without ( movie S3 ) and with ( movie S4 ) D/N cis-inhibition ( Fig . 4 ) . However , even with cis-inhibition instabilities have arisen after the fourth oscillation . To allow for more realistic noise-affected gene expression , we simulated mitosis by developing a model in which the dividing cells in the growth zone of the PSM shut off transcription , which consequently disturbs Hes7 expression waves after two oscillations even when the cells started synchronized at the beginning of the simulation . Furthermore , we allowed diagonal neighbors to signal via D/N . For a mitosis phase of 20 min , D/N cis-inhibition was able to maintain phase coherence reasonably well ( movie S5 ) , whereas without D/N cis-inhibition ( movie S6 ) the initial order was lost after two oscillation periods ( Fig . 4 ) . In summary , our results demonstrate that the inclusion of D/N cis-inhibition in the formulation of the model brings about a decisive improvement in the ability of D/N signaling to synchronize cellular oscillators . This is achieved not only for a specially chosen set of parameters , but a wide range of model parameters .
The aim of our modeling work in somitogenesis is to explain how the various expression waves in the mouse PSM are generated , why they slow down when they are nearing the anterior end of the unsegmented PSM , and how the boundary between the PSM and the next forming somite is formed . In our previous paper [10] we were concerned with the generation of the NICD wave and why it stops , because together with the TBX6 and FGF8 gradients NICD induces Mesp2 , which is critically important for boundary formation . Our hypothesis for the generation of the NICD wave was that the WNT3A and/or FGF8 gradients in the PSM influence an intracellular process of the core oscillator consisting of Hes7 and D/N thereby slowing the oscillator down when it gets out of the range of the gradients . Therefore , we modeled the core oscillator as a transport model with the most important cellular compartments ( nucleus , cytoplasm , and membrane ) and processes like transcription , translation and transport and allowed a possible coupling of each gradient to each cellular process . Furthermore , we included as many measurable parameters and especially promoter information as we could find in the literature ( which is unfortunately rather sparse ) . However , with plausible assumption one can generate at least the qualitative behavior with its characteristic expression pattern rather well . The drawback of our method is that one cannot sample the multidimensional parameter space . However , if new information becomes available , one can feed it directly into our model . In our 2012 paper [10] we had excluded the synchronization problem . Cells started synchronized and stayed so , because during proliferation daughter cells inherited the oscillatory phase of their mother cells . However , as NICD and D/N-signaling are widely held to be responsible for the maintenance of oscillations and synchronization of wave formation and in creating boundaries in space as the waves come to rest , one should work towards a comprehensive model including synchronization . In somitogenesis the formation of sharp boundaries between the block of cells forming a new pair of somites and the remaining PSM works only when gene expression in the cells forming the boundary is synchronized . It has been shown experimentally that D/N signaling is responsible for this synchronization . The species-specific periodicity of somitogenesis is controlled by cellular oscillators , in mouse most probably by the negative feedback oscillator Hes7 . The synchronization has to happen rather fast as a cell experiences about five oscillations from its birth to its incorporation into a somite [28] . Computer simulations describing oscillators coupled by classical modes of D/N-interaction failed so far to achieve synchronization in an appropriate time approach except by introducing cell movements in simulations . Here we show that fast synchronization of Hes-type oscillators can be achieved without cell movements by including the process of D/N cis-inhibition . While in conventional models of D/N synchronized oscillations each cell is sender as well as receiver of D/N-signaling because DELTA ligands as well as NOTCH receptors are active in the membrane of the cell , in a system with perfect cis-inhibition i . e . perfect titration of DELTA against NOTCH or vice versa , a cell is either sender or receiver . That means that a cell with DELTA excess – an information sender - can enforce a change in NICD controlled gene expression in a neighboring receiver cell , i . e . with NOTCH excess , as fast as intrinsic NICD processes allow in the receiver cell . If Delta expression is oscillatory – as in our model - the sender cell could go into receiver mode if Delta expression is low . So other cells could influence/synchronize this cell . In this manner , fast synchronization could be achieved despite the fact that the cell-interaction is still local ( even if one considers communication by cytonemes as observed in zebrafish [29] ) . This does not exclude the possibility that for very large volumes the locality of cell-cell-communication leads to domains synchronized to different phases , but for realistic numbers of cells the above acceleration of synchronization could be sufficient i . e . fast enough . However , for D/N synchronization of Hes7-oscillators the considerations shown above are too simplified , as a cell cannot be only sender , i . e . have any active NOTCH in its membrane . This is so because Hes7 activation relies on NICD and in our model of the core oscillator HES7 suppresses Dll1 expression leading to the oscillatory DELTA expression mentioned above . Consequently , a sender-only cell would have no interesting message to send . So a perfect titration of NOTCH against DELTA is not desirable . There has to be an optimum value of cis-inhibition . If this value is surpassed oscillations are damped and die out . This was shown in Fig . 2 . At least for mouse , there is strong evidence that the Hes7 gene oscillates by negative feedback of its protein on its own promoter , thereby serving as the core oscillator of the somitogenesis clock [30]–[32] . Furthermore , promoter analysis revealed that Hes7 is induced by D/N signaling [33] . The NOTCH modifying gene Lfng is also induced by D/N-signaling and oscillates in the PSM because its expression is inhibited by HES7 [33] . The fact that D/N-signaling is required for the synchronization of ultradian oscillators in the PSM was shown for zebrafish [16] , [34] in experiments with single cell resolution . Because it is not easy to separate the induction of oscillation and synchronization in mouse on the cellular level , Okubo et al . used chimeric embryos composed of wild-type and Dll1-null cells to demonstrate that D/N-signaling is responsible for the synchronization of oscillations in the PSM also in mouse [13] . To clarify the role of Lfng in the somitogenesis clock , Okubo et al . also analyzed Lfng chimeric embryos and used Notch signal reporter assays in a co-culture system [13] . As interpretation of the results they proposed a novel , in this form not yet described action of LFNG on DLL1 . The knockout of Lfng resulted in an enhanced activity of NICD in the PSM , which indicates that LFNG might affect NOTCH1 and DLL1 negatively . Okubo et al . also demonstrated that the synchronization of cellular oscillators was proportional to the number of Dll1 expressing ( wild-type ) cells in chimeric embryos , which confirmed that D/N synchronizes Hes7 oscillations in the PSM . Similarly , using Lfng chimeric embryos , they showed that LFNG seems to be required for this synchronization . Interestingly , computer simulations that integrated the proposed effect of LFNG on NOTCH1 and DLL1 showed fast oscillator synchronization and were able to reproduce their experimental findings [13] . In their model the Hes7 oscillator in every cell is coupled to neighboring cells via LFNG , which is itself driven by HES7 oscillations and regulates the intracellular coupling by inhibition of both NOTCH1 and DLL1 activity in the same cell . Thus , LFNG not only represses D/N signaling inside the LFNG expressing cell by modifying NOTCH1 cell-autonomously , but also represses D/N signaling between neighboring cells by also modifying the DLL1 ligand . In short , in their model the output of the Hes7 oscillator is coupled to D/N signaling exclusively by the way of LFNG action . In contrast , in our model we assume that HES7 inhibits Dll1 expression like Her1/7 inhibits deltaC in zebrafish . We will not repeat the extensive discussion provided in our previous publication [10] , but strengthen the main arguments , which are that expression of Dll1 is dynamic in the PSM [35] and that only the orthologs of Hes7 and Dll1 are dynamic in the PSM of all vertebrate systems examined so far [36] . For example , Lfng expression is constant in zebrafish as well as in medaka [36] . Therefore , we argue for an evolutionary mechanism with a zebrafish-like core oscillator in which LFNG acts only in a modulatory role . Consistent with this notion , NICD expression is still dynamic in Lfng deficient mice [37] and Lfng is not required for somite formation in the tail bud phase [38] . In this work , we therefore examined the effect of D/N cis-inhibition primarily in a model without modulation of D/N signaling by LFNG . Quantitative data regarding cell cycle parameters in mouse embryogenesis are sparse . Power and Tam give a value of ca . 30 min for 7 . 0-day embryos [39] . When judging about the success or failure of our model with respect to the real facts one should not forget that there may be biological mechanism that are not covered by the model , but could be crucial for the functioning of the synchronization . For example , it was found that Dll1 mRNA is stabilized during mitosis , by Elavl1/HuR in neuroepithelial cells [40] . If similar mechanisms are operative in the growth zone of the PSM , our assumption that mRNA decay rates are constant in time could be too pessimistic . A smaller decay rate during mitosis would very probably diminish the perturbation to oscillations and thereby improve synchronization . Interestingly , a study observing oscillatory expression of a Her1-Venus reporter at single cell resolution in the zebrafish PSM found that her1 oscillations are linked to mitosis [34] . Therefore , it is possible that cell divisions introduce less noise than our model assumes . In the hypothalamus of the mammalian brain , 20000 nerve cells function as circadian oscillators and have to be synchronized to function as the master circadian clock of the body [41] . Like ultradian oscillators , these circadian oscillators function by a negative transcription-translation feedback loop and are often also modeled by Goodwin-models ( see for example [42] , [43] and references therein ) , but also by delay differential equations or very simple toy models [44] . However , compared to the somitogenesis clock , in the circadian clock there are more interlocking feedback loops [41] and the communication between cells works either by secretion of neuropeptides and/or by direct innervation . So , coupling in the circadian clock is not mediated by communication between directly adjacent cells but by non-local interactions , which probably favors tissue-wide synchronization and prevents the phenomenon of cell territories synchronized to different phases ‘fighting’ for dominance . Furthermore , in circadian clock models the synchronization signal acts positively on the transcription of the clock genes . This is also the case in our model of the ultradian oscillator , where NICD acts as an activator on Hes7 transcription . However , HES7 represses Dll1 in the same cell and therefore NICD generation in the adjacent cell . This is the reason why lateral inhibition occurs in the static case or leads to anti-synchrony in the dynamic setting . Another difference concerns the coupling of the synchronization signal to the promoter of the clock feedback loop . In circadian models , this is mostly assumed to be additive , whereas we do not assume an additive but a multiplicative coupling of D/N signaling to the Hes7 promoter because it was shown that in most of the PSM Hes7 ceases to oscillate without D/N input . We disregard in our model the fact , that Hes7 is induced by FGF8 in the tailbud [8] , which would be an additive coupling to FGF8 . It was found in circadian oscillator models that weak oscillators , which are damped without a synchronization signal , synchronize faster [42] , [43] , [45] . As our Hes7 oscillator is coupled in ‘AND’ modus to the synchronization signal ( NICD ) , this could possibly be seen as an example of this principle . ( It was also found for the circadian clock that the oscillator's radial relaxation time scale and the ratio of synchronization signal to the oscillator amplitude are important for synchronization and oscillator entrainment [44] . ) Contrary to Wang et al . [46] who simulate neural fate decisions in the developing nervous system and proposed that D/N cis-inhibition causes asynchrony between adjacent cells , adding D/N cis-inhibition terms to our model of ultradian oscillators of the Hes/Hairy/her type clearly leads to a faster synchronization . Furthermore , the phenomenon of different regions that are synchronized to different oscillation-phase values , and that one region cannot overwhelm another , can be overcome without cell movements , at least for the non-growing case , by introducing D/N cis-inhibition . Since cis-inhibition allows faster reaction of cells on changes in their neighborhood , cell movement may not be required for all situations in which synchronization is mediated by D/N signaling . We also show that D/N cis-inhibition does not interfere with a proposed mechanism for wave generation in the PSM , in which the control of HES7 degradation by the posterior-to-anterior FGF8 gradient slows down the oscillators as they get out of the range of the gradients by the continuous growth of the PSM . That D/N cis-inhibition does not lead to complete synchronization in the whole PSM , which would resist slowing down , is probably caused by the fact that the slowing down gets appreciable only in the last oscillation a cellular oscillator experiences before being incorporated into a somite [10] . However , ultimately , only experiments can clarify whether D/N cis-inhibition [22] , [23] is functional also during somitogenesis .
To model gene expressions we use essentially the same methodology as described in [17] , i . e . a gene- and cell-based simulation program that numerically solves differential equations describing a gene regulatory network and displays the concentration of a selected gene product by color intensity ( virtual in situ staining ) in each cell . For showing the consequences of the gene regulatory network ( Fig . 1 ) we use the same cell- and gene based simulation program as in [10] except that cis-inhibitory interaction-terms in the membrane and cytoplasmic compartment were added . Specifically , we use the same formulas and rate constants as in our previous publication , except the addition of the D/N cis-inhibition terms , different values for Hill coefficient and Hill threshold describing the action of NICD at the Hes7 promoter , and the LFNG coupling . Furthermore , it is now possible to enlarge the neighborhood of a cell so that also diagonally adjacent cells are treated as interacting neighboring cells . In addition , we take into account that the Hill-coefficient for the action of the NICD complex on the Hes7 promoter could be higher than 2 because of cooperative effects between the dimer formed of a NICD-Maml1-Rbpj-kappa complex and additional chromatin modifying co-factors . As discussed in [10] , we introduce distinct variables for cytoplasmic and nuclear concentrations of proteins and the respective mRNAs . This distinction is made for the oscillatory factors HES1/7 , NICD and LFNG , but not for the slow-changing concentrations of protein and mRNA of Notch1 . The DLL1 ligand and the NOTCH receptor are modeled with independent variables in the cytoplasm and membrane compartments . In the somitogenesis model we included only the genes from our previous model [10] that are needed to generate the ‘wave’-pattern i . e . Dll1 , Notch1 , Hes7 , Fgf8 , Wnt3a , and Tbx6 , because the downstream genes like Mesp2 , Ripply2 and Epha4 would function similar as in our 2012 publication [10] except for possible Hill-threshold adjustments . A schematic view of the GRN used in our simulations is depicted in Fig . 1 . Its central element is the negative feedback oscillator Hes7 . By binding to the promoter it inhibits its own production . The Hes7 promoter also receives input from D/N signaling while we disregard here the contribution of Fgf signaling in the tailbud [8] . In an extended model , HES7 inhibits Lfng , which is induced by NICD , and in turn modulates D/N interaction . NICD acts as an activator of Hes7 . Here , we assume that HES7 inhibits Dll1 expression . For the mathematical description of the model we use ordinary differential equations . To describe negative feedback oscillators one has to introduce a function describing the repressive action of the gene product on the promoter of its gene . We use Hill functions of the form to describe this negative feedback , wherein the Hill-coefficient h is a measure for the cooperativity of the repressor binding to the promoter and HR as well as HA are the thresholds determining half-inhibition or activation , respectively ( see below ) . For transcription factors binding as homo-dimers we set the Hill coefficient to the value of 2 . To describe activating gene action we use analogously Hill functions of the form Oscillations start only when there is a delay between gene expression and negative feedback . This is often modeled with direct introduction of delayed arguments into the differential equations specifying the time used for transcribing a gene into mRNA and translating a mRNA into protein , resulting in a so-called delay differential equation system ( for an example see [15] , [47] ) . In the following , we specify the differential equations of our gene regulatory network . In all cases the gene indices on the variables written on the right side of the equations are not shown except when the variables refer to other genes . Decay rates are always given in min−1 and concentration values are given in arbitrary units . The equations below describe the negative feedback oscillator at the core of our GRN:Here pC ( t ) , pN ( t ) , mC ( t ) , and mN ( t ) designate concentrations of cytoplasmic protein , nuclear protein , cytoplasmic mRNA , and nuclear mRNA , respectively . The export rates of the protein from cytoplasm to nucleus , from nucleus to cytoplasm , and for the transport of mRNA from nucleus to cytoplasm are chosen as: epC = 0 . 007 , epN = 0 . 001 , and emN = 0 . 038 . Furthermore , dmC = 0 . 067 , dmN = 0 . 001 , and dpC = 0 . 031 describe the degradation rates for cytoplasmic and nuclear mRNA , and cytoplasmic protein , respectively . Based on experimental evidence , we assume a rather low rate of mRNA degradation in the nucleus for all genes [48] . We suppose saturated protein decay in the nucleus characterized by threshold value F = 0 . 2 and maximum rate G = 0 . 96 . The translation rate and the maximal transcription rate are given by K = 1 . 5 and k = 0 . 5 , respectively . The Hill function with HR = 1 . 0 and HA = 4 . 5 describes the negative feedback of HES7 on its own promoter and the control of Hes7 transcription by the Notch intracellular domain ( NICD ) . The bHLH-transcription factor HES7 binds as dimer to its own promoters thereby inhibiting transcription . The Hes7 gene contains only one N-box in its promoter [49] . If HES7 would bind also to the so-called E-boxes in the Hes7 promoter the Hill-coefficient could also be higher [50] . However , Chen et al . have shown that HES7 only binds to the N-box [33] , so only one HES7 dimer binds . Therefore we chose a Hill-coefficient of 2 . Furthermore , we subsume all interactions with co-factors of HES7 like Groucho/Tle1 in the basal transcription rate . HES7 is a target of D/N signaling . This entails NICD acting as transcriptional co-factor on the Hes7 promoter . As it was shown that two complexes comprising NICD , MAML1 and CSL bind as a dimer to the Hes1 promoter [51] and we assume a similar Hes7 promoter structure regarding activation by NICD , we also use a Hill-coefficient of 2 or higher for the Hill-function describing the effect of NICD in our simulations . NICD is a fragment of the Notch receptor , which is generated after binding of the DLL1 ligand to the NOTCH1 receptor . Ligand binding enables access of proteases to cleavage sites in the intracellular part of NOTCH1 and subsequent transport of NICD from the cytoplasm to the nucleus [52] . Here , rDN = 0 . 05 is the reaction rate between NOTCH1 receptors and the DLL1 ligands on the n neighboring cells , while raLfng describes the activation of D/N signaling by LFNG , and r0 is the reaction rate of DLL1 and NOTCH1 without LFNG action . For the simulations shown here the default value is 0 . 256 . pMNotch1 designates NOTCH membrane protein concentration , pMDll1 DLL1 protein concentration in the membrane . epC = 0 . 12 and epN = 0 . 6 are the export rates for NICD from the cytoplasm to the nucleus and vice versa , and dpC = 0 . 2 is the NICD decay rate in the cytoplasm . As NICD acts as a co-transcription factor in the nucleus its import rate to the nucleus is chosen larger as the export rate . In the simulations without Lfng in the GRN raLfng is set to 1 . At least in the presomitic mesoderm it was demonstrated that Dll1 expression is dynamic [35] . So the mathematics of negative feedback systems necessitates the use of a transport equation system with at least three equations for oscillatory behavior to be possible [53] . We use two equations for Dll1 mRNA and protein , each in nucleus and cytoplasm , making four differential equations:In the PSM Dll1 is activated directly and indirectly via TBX6 by Wnt signaling [54] . Based on experimental evidence , we assume an additional control by HES7 ( see [10] for an extensive discussion ) . In the spatially constant model system we disregard the control by TBX6 and WNT3A . Therefore , we chose a Hill function of the form , with HR = 1 . 0 . We chose the rate constants as in [10]: K = 1 . 5 , dpC = 0 . 09 , epC = 0 . 1 , epM = 0 . 1 , dpM = 0 , dmC = 0 . 12 , emN = 0 . 09 , dmN = 0 . 001 and k = 1 . 25 . The rate constant rDNcis = 0 . 01 describing D/N cis-inhibition results in fast synchronization . After binding of one DLL1 molecule in the membrane of one cell to a NOTCH1 receptor in the membrane of a neighboring cell , the intracellular part of NOTCH1 is cleaved off to release NICD . This results in the destruction of the NOTCH1 molecule in this reaction . Therefore , the reaction term is subtracted in the equation describing NOTCH1 in the membrane , while it is added to the NICD equation . Because the DLL1 ligand bound to the extracellular domain of NOTCH1 is endocytosed and probably degraded [55] , the same reaction term is subtracted in the equation describing DLL1 in the membrane . We assume that the same applies to the intracellular complex formed by a DELTA and NOTCH molecule . Since we assume Notch1 expression to be static it suffices to describe its mRNA concentration by one simple equation with a production and decay term i . e . without differentiating between nucleus and cytoplasm . We chose K = 1 . 5 , dpC = 0 . 2 , epC = 0 . 1 , epM = 0 . 0 , dpM = 0 . 1 , dm = 0 . 02 , and k = 0 . 5 for the rate constants . The differential equation system for Lfng has essentially the same structure as the one for Hes7 , except that HES7 exerts a repressive influence on the Lfng promoter while NICD activates it . This is described by the Hill function with HR = 1 . 0 and HA = 4 . 5 . Here pC ( t ) , pN ( t ) , mC ( t ) , and mN ( t ) designate concentrations of cytoplasmic protein , nuclear protein , cytoplasmic mRNA , and nuclear mRNA , respectively . The export rates of the protein from cytoplasm to nucleus , from nucleus to cytoplasm , and for the transport of mRNA from nucleus to cytoplasm are chosen as: epC = 0 . 007 , epN = 0 . 001 , and emN = 0 . 038 . Furthermore , dmC = 0 . 067 , dmN = 0 . 001 , and dpC = 0 . 031 describe the degradation rates for cytoplasmic and nuclear mRNA , and cytoplasmic protein , respectively . Again we suppose saturated protein decay in the nucleus characterized by threshold value F = 0 . 2 and maximum rate G = 0 . 96 . The translation rate and the maximal transcription rate are given by K = 1 . 5 and k = 0 . 5 , respectively . For the modeling of growth and geometry in the growing PSM we refer to [10] . We also use the same parameters and equations for the Wnt3a , Tbx6 , and Fgf8 genes described therein . Dll1 and Notch1 induction by WNT3A and TBX6 , i . e . , their corresponding Hill functions , are also chosen as in [10] . Noise is introduced by shutting off transcription only for Hes7 , Dll1 , Notch1 , and Lfng i . e . , not for the gradient generating genes , because in our model there is no way this noise could be corrected by D/N signaling . To include this , one would have to simulate a full model of the Wnt3a and Fgf8 pathway with genes like Nkd1 or Dusp4 and others , which exert a negative feedback on their respective pathways and are known to be controlled by D/N signaling [56] .
|
During vertebrate embryonic development , the segmented structure of the axial skeleton is laid down by the process of somitogenesis . Periodically , blocks of cells separate at the anterior end of a mesenchymal tissue ( PSM ) on either side of the neural tube and develop later into spinal vertebrae . Cellular oscillators operating in each cell of the PSM control this process . Their synchronization is essential , and is effected by direct cell-to-cell signaling of the Delta/Notch ( D/N ) pathway . To better understand the regulation of the genes involved , we employ computer modeling . In this case , the fast synchronization of the oscillators represents a special challenging and worked so far only by the integration of cell movements . Now , we have succeeded in accelerating the synchronization for the first time without cell movements by the interposition of the novel mechanism of intracellular reciprocal inhibition termed D/N cis-inhibition into our computer simulations .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"theoretical",
"biology",
"biology",
"and",
"life",
"sciences",
"developmental",
"biology",
"computational",
"biology"
] |
2014
|
Fast Synchronization of Ultradian Oscillators Controlled by Delta-Notch Signaling with Cis-Inhibition
|
Genes with male- and testis-enriched expression are under-represented on the Drosophila melanogaster X chromosome . There is also an excess of retrotransposed genes , many of which are expressed in testis , that have “escaped” the X chromosome and moved to the autosomes . It has been proposed that inactivation of the X chromosome during spermatogenesis contributes to these patterns: genes with a beneficial function late in spermatogenesis should be selectively favored to be autosomal in order to avoid inactivation . However , conclusive evidence for X inactivation in the male germline has been lacking . To test for such inactivation , we used a transgenic construct in which expression of a lacZ reporter gene was driven by the promoter sequence of the autosomal , testis-specific ocnus gene . Autosomal insertions of this transgene showed the expected pattern of male- and testis-specific expression . X-linked insertions , in contrast , showed only very low levels of reporter gene expression . Thus , we find that X linkage inhibits the activity of a testis-specific promoter . We obtained the same result using a vector in which the transgene was flanked by chromosomal insulator sequences . These results are consistent with global inactivation of the X chromosome in the male germline and support a selective explanation for X chromosome avoidance of genes with beneficial effects late in spermatogenesis .
Sex chromosomes , such as the X and Y chromosomes of Drosophila , are thought to have evolved from a pair of homologous autosomes that lost their ability to recombine with each other [1 , 2] . Over evolutionary time , the sex chromosome that is present only in the heterogametic sex ( the Y in Drosophila and mammals ) tends to degenerate , losing most of its gene complement and accumulating transposable elements [3–6] . The X chromosome , which is still able to recombine within the homogametic sex , maintains a fully functional complement of genes and resembles an autosome in its size , cytogenetic appearance , repetitive element content , and gene density . Recent genomic studies , however , have revealed a number of more subtle differences in gene content , expression pattern , and molecular evolution between the X chromosome and the autosomes [7] . One pattern that has emerged from the genomic analysis of Drosophila melanogaster is that there is a significant excess of gene duplications in which a new autosomal gene has arisen from an X-linked parental gene through retrotransposition [8] . Most of these new autosomal genes appear to be functional and are expressed in testis [8] . Several of these genes that have been studied in detail show evidence of adaptive evolution and/or functional diversification [8–11] . Another pattern that has emerged from functional genomic studies is that genes with male-enriched expression are under-represented on the X chromosome [12 , 13] . For example , about 19% of all D . melanogaster genes reside on the X chromosome , but only 11% of the genes with a 2-fold or greater male bias in expression are X-linked [14] . Furthermore , the male-biased genes that are X-linked tend to show less sex bias in their expression than those that are autosomal [15] . A number of hypotheses have been put forth to explain the above observations [16–18] . To explain the large excess of retrotransposed genes that have “escaped” the X chromosome , Betrán et al . [8] proposed the X inactivation hypothesis , which posits that genes with a beneficial effect late in spermatogenesis are selectively favored to be autosomally located . Otherwise , their expression would be prevented by male germline X inactivation , which is supposed to occur early in spermatogenesis when autosomal genes are still actively transcribed . Early X inactivation could also explain the paucity of genes with male-biased expression on the X chromosome: if X-linked genes cannot be expressed in the later stages of spermatogenesis , then one would expect to see fewer X-linked genes with enriched expression in adult males . In particular , this should be true for genes expressed in the male germline and those encoding sperm proteins , which has been observed [12 , 19] . Male germline X inactivation , however , cannot completely explain the observations . For instance , male-biased genes that are expressed only in somatic cells , where X inactivation does not occur , are also significantly under-represented on the X chromosome [12 , 20] . An alternative explanation that accommodates this observation invokes sexual antagonism , that is , evolutionary conflict between males and females . The fixation probability of an X-linked , sexually antagonistic mutation is expected to differ from that of an autosomal one , with the direction of this difference depending on the dominance coefficient [21 , 22] . If the antagonistic effects are ( at least partly ) dominant , then female-beneficial/male-harmful mutations will accumulate on the X chromosome , while male-beneficial/female-harmful mutations will be removed from the X . This is because the X chromosome spends two-thirds of its evolutionary history in females and , thus , is more often under selection in the background of this sex . Since genes with sex-biased expression may be prime targets for sexually antagonistic mutations , the above scenario could lead to an excess of female-biased genes and a paucity of male-biased genes on the X [13] , resulting in “feminization” or “demasculinization” of this chromosome [12] . A hypothesis that combines the concepts of sexual antagonism and X inactivation was proposed by Wu and Xu [23] . This hypothesis , termed SAXI ( sexually antagonistic X inactivation ) , suggests that natural selection has favored the movement of sexually antagonistic X-linked genes whose expression is beneficial to males to the autosomes , leaving those beneficial to females on the X . Over evolutionary time , the accumulation of female-beneficial/male-harmful genes on the X leads to selection for X inactivation in the male germline , particularly during the later stages of spermatogenesis where the effects of sexual antagonism are expected to be greatest [23] . The hypotheses of Betrán et al . [8] and Wu and Xu [23] assume that the X chromosome becomes inactive before the autosomes during spermatogenesis . This phenomenon has been established in mammals and nematodes [24–26] . However , the evidence for male germline X inactivation in Drosophila has been equivocal . Lifschytz and Lindsley [27] cited cytological observations and genetic experiments to argue that X inactivation during spermatogenesis was common to most animal species with heterogametic males , including D . melanogaster . However , similar evidence was used to argue against X inactivation in Drosophila [28] . A later study of the expression of sperm-specific proteins in transgenic Drosophila provided experimental support for X inactivation [29] . Here the authors used a testis-specific promoter to drive the expression of altered forms of β-tubulins in the male germline and noted that X-linked inserts of the constructs showed reduced expression relative to autosomal inserts . Although this result was consistent with X inactivation , there were some limitations . For instance , the sample sizes were small for each of the expression constructs , with only one or two X-linked inserts per construct . Furthermore , the expression level of the genes was only roughly estimated from protein abundance on electrophoresis gels . A more recent experimental study failed to find support for male germline X inactivation in Drosophila [30] . These authors examined the expression and intracellular location of the MLE protein ( encoded by maleless ) , as well as the acetylation pattern of histone H4 , in male germline cells . Although MLE is known to be involved in X chromosome hypertranscription in somatic cells , presumably through the recruitment of histone acetylation factors [31 , 32] , it does not associate specifically with the X chromosome in male germ cells . Furthermore , H4 acetylation at lysine 16 , which is thought to be a reliable marker for active transcription , was observed equally on the X chromosome and the autosomes . Thus , there was no evidence for dosage compensation or X inactivation in the male germline . However , it is not necessary that these two processes occur through the same mechanism , or that they rely on the same proteins required for somatic cell dosage compensation . Indeed , a microarray analysis of germline gene expression indicated that dosage compensation does occur in the male germline [33] . Because these microarray experiments used reproductive tissues that contained somatic cells and germline cells from all stages of gametogenesis , they could not directly address the issue of early X inactivation . However , the fact that most X-linked genes showed similar levels of expression in both male and female reproductive tissues suggests that , if X chromosome inactivation does occur in the male germline , it does not have a large effect on the global pattern of sex-biased gene expression . In this study , we perform a more rigorous experimental test for X inactivation in the male germline . Using a transgenic construct in which the expression of a reporter gene is driven by the promoter of the autosomal , testis-specific ocnus ( ocn ) gene , we show that autosomal inserts are expressed specifically in males and in testis . X-linked inserts , in contrast , show greatly reduced levels of expression . These results hold for a large sample of independent insertions and for two different transformation vectors and , thus , provide strong support for inactivation of the X chromosome during Drosophila spermatogenesis .
The ocn gene is expressed specifically in testis and encodes a protein abundant in mature sperm [19 , 34] . It is part of a cluster of three tandemly duplicated genes on chromosome arm 3R that are present in all species of the D . melanogaster species subgroup and shares greatest homology to the neighboring janusB ( janB ) gene , which is also expressed in testis . Although ocn lies only 250 bp distal to janB , it produces a unique transcript that does not overlap with that of janB [34] . The first half of the janB-ocn intergenic region is highly diverged among species of the D . melanogaster subgroup and cannot be aligned unambiguously . However , the portion just upstream of the ocn start codon is well conserved , suggesting that it has regulatory function ( Figure S1 ) . We refer to this region as the ocn promoter . To test its ability to drive tissue-specific gene expression , we fused it to the open reading frame of the Escherichia coli lacZ gene , which encodes the enzyme β-galactosidase ( Figure 1A ) . Transgenic flies with autosomal insertions of P[wFl-ocn-lacZ] showed reporter gene expression specifically in testis , as expected ( Figure 2 ) . Overall , we obtained 15 independent autosomal insertions of P[wFl-ocn-lacZ] . The mean β-galactosidase activity in adult males was 8 . 67 units , while that in adult females was 0 . 34 units . The difference between the sexes was highly significant ( Mann-Whitney U test , p < 0 . 001 ) . The mean β-galactosidase activity of gonadectomized males was 0 . 24 units , which was significantly less than whole males ( Mann-Whitney U test , p < 0 . 01 ) . If the X chromosome is inactivated before the autosomes during spermatogenesis , then one would expect transgenic lines with X-linked insertions of P[wFl-ocn-lacZ] to show lower levels of reporter gene expression than those with autosomal insertions . This is indeed what we observed . In total , we obtained ten independent X-linked insertions of P[wFl-ocn-lacZ] . All of these lines showed reduced β-galactosidase activity in adult males relative to the autosomal-insertion lines ( Figures 2 and 3 ) . On average , the activity difference between autosomal and X-linked insertions was 7-fold ( 8 . 67 versus 1 . 19 units ) , and the difference between the two groups was highly significant ( Mann-Whitney U test , p < 0 . 001 ) . Although β-galactosidase activity was very low for the X-linked insertions , it was significantly greater than zero . Assuming a normal distribution of activity among the X-insertion lines , the 95% confidence interval was 0 . 82–1 . 56 units . Five of the autosomal insertion lines ( the last five in Figure 3 ) were obtained through the re-mobilization of X-linked inserts ( see Materials and Methods ) , demonstrating that the reduction in expression was not caused by undesired sequence changes in the ocn promoter or lacZ coding sequence , but instead was a direct result of X linkage . Because the assays of β-galactosidase activity measure expression at the level of protein abundance , it is possible that they do not reflect underlying levels of transcription . To test this possibility , we performed quantitative reverse-transcription PCR ( qRT-PCR ) to estimate the relative transcript abundance of a subset of eight transformed lines , including four with autosomal and four with X-linked inserts . The autosomal inserts had significantly higher transgene expression at the level of mRNA ( Mann-Whitney U test , p = 0 . 02 ) , with the relative expression difference being 5-fold ( Figure 4B ) , which corresponds well to the observed difference in β-galactosidase activity and suggests that the enzymatic assays provide a reliable estimate of expression . To test if the reduced expression of the X-linked ocn-lacZ transgenes could be attributed to the presence of localized transcriptional repressors bound to the X chromosome , we performed additional experiments using the P[YEStes-lacZ] transformation vector ( Figure 1B ) , which contains binding sites for the protein encoded by suppressor of Hairy-wing . These binding sites flank the inserted transgene and serve to insulate it from the effects of external transcriptional regulators [35] . We obtained 12 independent autosomal insertions of P[YEStes-lacZ] , and these lines showed male- and testis-specific expression of the lacZ reporter gene . The mean β-galactosidase activity in adult males was 1 . 84 units , which was significantly greater than that of adult females ( mean = 0 . 42; Mann-Whitney U test , p < 0 . 001 ) or gonadectomized males ( mean = 0 . 22; Mann-Whitney U test , p < 0 . 001 ) . We also obtained ten independent insertions of P[YEStes-lacZ] on the X chromosome . Adult males of these lines had a mean β-galactosidase activity of 0 . 17 units , which differed significantly from the autosomal-insert lines ( Mann-Whitney U test , p < 0 . 001 ) , but did not differ significantly from zero ( 95% confidence interval = −0 . 09–0 . 43 ) . The reduction in reporter β-galactosidase activity caused by X linkage was more than 10-fold ( Figure 4A ) . We also assayed expression at the level of transcript abundance by performing qRT-PCR on a subset of eight transformed lines ( four with autosomal and four with X-linked inserts ) . Again , the X chromosome insertion lines showed significantly less transgene expression than the autosomal insertion lines ( Mann-Whitney U test , p = 0 . 02 ) . The reduction in reporter gene expression measured by qRT-PCR was 3 . 4-fold ( Figure 4B ) . Thus , the presence of the chromosomal insulator sequences did not alleviate transcriptional repression of the X-linked transgenes . For adult males with autosomal insertions , the coefficient of variation ( CV ) for β-galactosidase activity was lower among the P[YEStes-lacZ] transformed lines ( CV = 0 . 16 ) than among the P[wFl-ocn-lacZ] transformed lines ( CV = 0 . 28 ) . A more pronounced difference was seen at the level of mRNA abundance , where the CVs for P[YEStes-lacZ] and P[wFl-ocn-lacZ] transformants were 0 . 07 and 0 . 44 , respectively . This suggests that the insulator sequences successfully reduced position effect variation caused by the chromosomal context of the insertion . The P[YEStes-lacZ] transformants , however , showed significantly less β-galactosidase activity than the P[wFl-ocn-lacZ] transformants ( Mann-Whitney U test , p < 0 . 001; Figure 4A ) . Interestingly , this difference was not detectable at the level of mRNA abundance ( Figure 4B ) , which suggests additional , post-transcriptional regulation of the P[YEStes-lacZ] transgenes .
Although a number of hypotheses regarding genome and sex chromosome evolution assume that the Drosophila X chromosome becomes transcriptionally inactive before the autosomes during spermatogenesis , little direct evidence for this scenario has been reported . Our experimental results indicate that X chromosome inactivation does occur in Drosophila and that it can have a considerable effect on gene expression in the male germline . In total , we examined 27 autosomal and 20 X-linked insertions of a testis-specific reporter gene in two different transformation vectors . In all cases , transformed lines with autosomal insertions showed significantly greater transgene expression than their X-linked counterparts , with the differences in expression ranging from 3 . 4- to 10-fold . The consistency of these results across a large number of independent insertions suggests that this transcriptional inactivity is a global property of the X chromosome . The fact that we observe the same pattern when using a vector that insulates the transgene from external transcriptional regulators further suggests that inactivation of the X chromosome in the male germline occurs through a major structural change , rather than by the binding of localized transcriptional repressors . Could our results be explained by something other than male germline X inactivation ? One possibility is that there is an insertional bias of our transgenes that differs between the X chromosome and the autosomes . For example , X-linked inserts could preferentially target inactive or heterochromatic regions . To investigate this possibility , we used inverse PCR to map the insertion sites ( Figure 5 ) . We found that the insertions span the euchromatic regions of the X and autosomes , with many being in or near genes ( Table S1 ) . Thus , our mapping results run counter to the expectations of insertional bias as a cause of the observed differences in expression . Another possibility is that insertion of the transgenes onto the X chromosome may cause rearrangements or other disruptions to the gene or promoter that prevent proper expression . However , by remobilizing multiple , independent X inserts to new autosomal locations , we have shown that their expression can be restored . Thus , the X-linked insertions must have been intact . Finally , a lack of proper dosage compensation of transgenes inserted onto the X chromosome could possibly lead to reduced expression . We consider this mechanism unlikely for two reasons . First , X chromosome dosage compensation has been shown to occur on a global level in the Drosophila germline [33] . Second , the expression assays for the autosomal-insert lines were performed on flies heterozygous for the insertion . Thus , even if dosage compensation did not occur , we would expect to observe equal expression of X-linked and autosomal transgenes . Any degree of dosage compensation would result in higher activity in the X-insertion lines , which makes our test conservative . The use of the ocn promoter may make our experimental system especially sensitive to the effects of male germline X inactivation for two reasons . First , the promoter fragment used here is rather short ( 150 bp ) and , thus , may be abnormally influenced by differences in chromatin environment between the autosomes and the X chromosome . It should be noted , however , that other known testis-specific promoters are also relatively short , in the range of 76–390 bp [36–38] . Second , ocn is likely to be expressed relatively late in spermatogenesis , when the effects of X inactivation should be pronounced . The ocn gene was originally identified as one encoding a protein abundant in the testes of mature males , but absent from those of immature males [34] . Our observation that β-galactosidase activity imparted by the ocn-lacZ transgenes is greatest in proximal regions of the testis ( Figure 2 ) also supports its relatively late expression . Furthermore , levels of β-galactosidase activity , as well as transgene transcript abundance as measured by qRT-PCR , are at least 50-fold lower in the third larval instar stage , when spermatogenesis is not yet complete , than in adult males ( unpublished data ) . Thus , it may be that a large proportion of ocn expression occurs after the X chromosome is inactivated . Indeed , if X-linked genes expressed early in spermatogenesis are hypertranscribed through a dosage compensation mechanism [33] , the effects of later X inactivation may be masked . Finally , we wish to point out that , although testis-expressed genes are under-represented on the X chromosome , they are not absent . Thus , many X-linked genes involved in spermatogenesis must be expressed at levels sufficient for proper function . This may be a result of their ( hyper ) transcription early in spermatogenesis . Recently , it has been noted that a region of the X chromosome is enriched for newly evolved , testis-expressed genes [39–41] , which suggests that this region may escape germline X inactivation . One of our transgene inserts falls within ∼500 kb of this interval , but does not differ in expression from other X-linked insertions . A higher density of X-linked transgene insertions may reveal specific regions that escape inactivation . Overall , P[YEStes-lacZ] transformants had much lower β-galactosidase activity than P[wFl-ocn-lacZ] transformants ( Figure 4A ) . This difference was not observable at the level of mRNA ( Figure 4B ) , suggesting additional regulation at the level of translation . Two major differences between the vectors could account for the discrepancy between the enzymatic assays and the qRT-PCR measurements . The first is the suppressor of Hairy-wing chromosomal insulator sequences in P[YEStes-lacZ] ( Figure 1 ) . However , it seems unlikely that these insulator sequences , which lie far outside of the transcriptional unit , would be involved in post-transcriptional regulation . Furthermore , putting the transgenes into a genetic background homozygous for a mutant suppressor of Hairy-wing allele had no effect on levels of β-galactosidase activity ( Figure S2 ) . The second difference is that P[YEStes-lacZ] contains the ocn 3′ untranslated region ( UTR ) ( Figure 1 ) . Although functional information for this 3′ UTR is lacking , the presence of two conserved sequence blocks suggests that it may play a role in the regulation of expression ( Figure S1 ) . Our finding that a testis-specific gene is not properly expressed when located on the X chromosome provides compelling experimental evidence for male germline X inactivation in Drosophila , something that was first proposed over thirty years ago [27] . It is also consistent with a selective explanation for the overabundance of retrotransposed genes that have moved from the X to the autosomes [8] . If such genes have a beneficial effect when expressed in testis ( especially in later stages of spermatogenesis ) , then selection would favor the maintenance of autosomal copies . The acquisition of expression late in spermatogenesis may even predispose a gene to adaptive evolution , as testis-expressed genes appear to be targets of positive selection more often than genes of other expression classes [42] . Our results also have relevance to the SAXI hypothesis [23] , which proposes that sexual antagonism leads to the selective relocation of male-beneficial genes expressed late in spermatogenesis to the autosomes . After all such genes have been relocated , selection could favor global inactivation of the X chromosome during spermatogenesis to prevent the expression of female-beneficial genes that have a harmful effect when expressed in males . Alternatively , the X may be inactive at this stage simply because it no longer contains genes with the proper regulatory sequences required for male germline expression . Our results are consistent with the former scenario , as the ocn promoter , which drives testis-specific expression on autosomes , does not function properly when relocated to the X chromosome .
Two different expression vectors that combined the ocn promoter of D . melanogaster with the lacZ coding region of E . coli were generated using standard techniques [43] . For the first , we PCR-amplified a 150-bp fragment of D . melanogaster genomic DNA that spanned bases 25 , 863 , 383–25 , 863 , 532 of Chromosome 3R in genome release 5 ( http://www . fruitfly . org/sequence/release5genomic . shtml ) . The amplified region includes 80 bases of 5′ flanking sequence and 70 bases of 5′ UTR of the ocn gene , corresponding to bases −165 to −16 relative to the A in the ATG start codon . We chose to end the promoter fragment at −16 because the preceeding sequence presented a good target for PCR-primer design; we know of no functional reason to include or exclude the final 15 bp before the start codon . The PCR product was cloned directly into the pCR2 . 1-TOPO vector ( Invitrogen , http://www . invitrogen . com/ ) . The identity and orientation of the cloned fragment were confirmed by restriction analysis . A 3 . 5-kb NotI fragment containing the complete E . coli lacZ open reading frame was excised from the plasmid pCMV-SPORT-βgal ( Invitrogen ) and inserted into the NotI site of the above plasmid , just downstream of the ocn promoter and in the same orientation . A 3 . 6-kb fragment containing the ocn promoter and the lacZ coding region was then excised as an SpeI/XbaI fragment and cloned into the SpeI site of the pP[wFl] transformation vector . This vector is based on the P transposable element and contains the D . melanogaster white ( w ) gene as a selectable marker [44] . The final construct was designated pP[wFl-ocn-lacZ] ( Figure 1A ) . The second expression vector contained the ocn promoter described above as well as the ocn 3′ UTR sequence ( Figure S1 ) . The ocn promoter was excised from the pCR2 . 1-TOPO vector as a BamHI/XbaI fragment and inserted into the BamHI/XbaI sites of the plasmid pUC18 ( Invitrogen ) . The ocn 3′ UTR sequence was PCR-amplified from genomic DNA corresponding to bases 25 , 862 , 721–25 , 862 , 830 of chromosome arm 3R ( bases −16 to +93 relative to the T in the TGA stop codon ) and cloned into the pCR2 . 1-TOPO vector . After confirming the identity and orientation of the cloned fragment by restriction analysis , a HindIII fragment ( where one HindIII site was internal to the 3′ UTR fragment , occurring 30 bp from the 5′ end ) was extracted and inserted into the HindIII site of the pUC18 plasmid containing the ocn promoter , such that the promoter and 3′ UTR were in the same orientation . An SpeI fragment containing both the promoter and the 3′ UTR was then excised and cloned into the XbaI site of the YES vector [35] . This vector is also based on the P transposable element and contains the yellow ( y ) gene of D . melanogaster as a selectable marker . Additionally , it contains binding sites for the Suppressor of Hairy-wing protein that flank the inserted DNA and serve to insulate it from position effects caused by random insertion of the vector into the genome [35] . The resulting transformation vector was designated as YEStes ( YES vector for testes-specific expression ) and contains the ocn promoter and 3′ UTR separated by unique XbaI and NotI restriction sites . To complete the expression construct , a 3 . 5-kb NotI fragment of the plasmid pCMV-SPORT-βgal containing the complete lacZ open reading frame was cloned into the NotI site of the YEStes vector in the appropriate orientation . This final construct was designated pP[YEStes-lacZ] ( Figure 1B ) . Plasmid DNA of the above expression constructs was purified using the QIAprep Spin kit ( QIAGEN , http://www . qiagen . com/ ) and used for microinjection of early stage embryos of the y w; Δ2–3 , Sb/TM6 strain of D . melanogaster following standard procedures [45 , 46] . Because it carries both the y and w mutations , this strain could be used for both transformation vectors . The Δ2–3 P element on the third chromosome served as source of transposase [47] . Following transformation , all lines were crossed to a y w stock to remove the transposase source . In cases where the transgene insertion was linked to the Δ2–3 source of transposase , the inserts were immediately remobilized by crossing transformed males to y w females and selecting offspring carrying the transgene , but not the Δ2–3 element . These flies were then mated to y w flies of the opposite sex to establish stable transgenic lines . X-linked insertions were identified by crossing transformed males to y w females and following inheritance of the phenotypic marker ( y+ or w+ ) ; crosses in which all daughters , but no sons , showed the marker phenotype revealed X linkage . Some X-linked insertions were mobilized to the autosomes by the following procedure . Transformed females were mated to y w; Δ2–3 , Sb/TM6 males and the male offspring carrying both the transgene and the Δ2–3 source of transposase were mated to y w females . From this cross , we selected male offspring carrying the transgene ( which could not be on the X chromosome inherited from the mother ) . These males were mated to y w females to establish stable transformed lines with new autosomal insertions of the transgene . To map the intrachromosomal location of the transgene insertions , the genomic sequence flanking the P-element vector was determined by sequencing the products of inverse PCR [48] . Briefly , genomic DNA was extracted from insertion-bearing flies and digested with either HpaII or HinP1I . The digestion products were self-ligated and used as a template for PCR with primer pairs Pry1 ( 5′-CCTTAGCATGTCCGTGGGGTTTGAAT-3′ ) and Pry2 ( 5′-CTTGCCGACGGGACCACCTTATGTTATT-3′ ) ; Plac1 ( 5′-CACCCAAGGCTCTGCTCCCACAAT-3′ ) and Plac4 ( 5′-ACTGTGCGTTAGGTCCTGTTCATTGTT-3′ ) to determine 3′ or 5′ flanking sequences , respectively . PCR products were sequenced with BigDye v1 . 1 chemistry on a 3730 automated sequencer ( Applied Biosystems , http://www . appliedbiosystems . com/ ) using the PCR primers as sequencing primers . In all cases , the chromosomal locations assigned by inverse PCR were consistent with those determined by genetic crosses . To determine in vivo expression levels of our transgenic constructs , we measured the level of β-galactosidase activity in transformed flies . For all autosomal insert lines , transformed males were mated to y w females and offspring heterozygous for the transgene insertion were used for assays . For transformants with X-linked inserts , females were mated to y w males and offspring heterozygous ( female ) or hemizygous ( male ) for the transgene insertion were used for assays . In all cases , the offspring were collected shortly after eclosion and separated by sex until they were assayed at age 5–7 d . All flies were raised on cornmeal-molasses medium at 25 °C . For assays of β-galactosidase activity , five adult flies of the same sex were homogenized in 150 μl of a buffer containing 0 . 1 M Tris-HCl , 1 mM EDTA , and 7 mM 2-mercaptoethanol at pH 7 . 5 . After incubation on ice for 15 min , the homogenates were centrifuged at 12 , 000 g for 15 min at 4 °C and the supernatant containing soluble proteins was retained . For each assay , 50 μl of this supernatant were combined with 50 μl of assay buffer ( 200 mM sodium phosphate [pH 7 . 3] , 2 mM MgCl2 , 100 mM 2-mecaptoethanol ) containing 1 . 33 mg/ml o-nitrophenyl-β-D-galactopyranoside . β-Galactosidase activity was measured by following the change in absorbance at 420 nm over 30 min at 25 °C . β-Galactosidase activity units were quantified as the change in absorbance per minute multiplied by 1 , 000 ( mOD/min ) . For all transformed lines , we performed at least two technical and two biological replicates ( always in equal numbers ) , where the former used the same soluble protein extraction and the latter used extractions from independent cohorts of flies . The activity of each line was calculated as the mean over all replicates , with the variance and standard error calculated among replicates . For comparisons between chromosomes or vectors , we averaged over the means of the individual lines and used the among-line variation to calculate variance , standard error , and CV . This approach is conservative , as the among-line differences ( position effects ) tended to be the largest source of variation . Statistical tests for differences between groups were performed using nonparametric methods , such as the Mann-Whitney U test , that do not rely on estimates of variance . For our purposes this approach is conservative . For lines that showed β-galactosidase activity in adult males , we also performed assays on gonadectomized males . This was done following the above protocol , after removal of the testes by manual dissection . For visualizing β-galactosidase activity in whole tissues , we incubated dissected testes in the above assay buffer containing 1 mg/ml ferric ammonium citrate and 1 . 8 mg/ml of S-GAL sodium salt ( Sigma-Aldrich , http://www . sigmaaldrich . com/ ) for 6 h at 37 °C . To measure expression at the level of transcription ( mRNA abundance ) , we performed qRT-PCR using a TaqMan assay ( Applied Biosystems ) designed specifically to our transgene ( i . e . , spanning the junction between the ocn 5′ UTR and the lacZ coding region ) . For this , 1 μg of DNase I-treated total RNA isolated from heterozygous ( autosomal insertions ) or hemizygous ( X insertions ) males was reverse transcribed using Superscript II reverse transcriptase and random hexamer primers ( Invitrogen ) according to the manufacturer's protocol . A 1:10 dilution of the resulting cDNA was used as template for PCR on a 7500 Fast Real-Time PCR System ( Applied Biosystems ) . The average threshold cycle value ( Ct ) was calculated from two technical replicates per sample . Expression of the transgene was standardized relative to the ribosomal protein gene RpL32 ( TaqMan probe ID Dm02151827 ) . Relative expression values were determined by the ΔΔCt method according the formula 2− ( ΔCtx − ΔCtmin ) , where ΔCtx = Cttransgene − CtRpL32 for a given transformed line x , and ΔCtmin represents the corresponding value of the line displaying the lowest level of transgene relative to RpL32 expression . Statistical analyses were performed as described above for β-galactosidase activity .
The FlyBase ( http://www . flybase . org/ ) accession numbers for the genes discussed in this article are janus B ( CG7931; FBgn0001281 ) , ocnus ( CG7929; FBgn0041102 ) , RpL32 ( CG7939; FBgn0002626 ) , white ( CG2759; FBgn0003996 ) , and yellow ( CG3757; FBgn0004034 ) .
|
During spermatogenesis , the X chromosome is inactivated in the male germline ( sperm cells ) , thereby silencing , or inactivating , genes residing on the X chromosome . X chromosome silencing is thought to be common among species with XY sex determination and has important implications for genome evolution . For example , genes with increased expression in the male tend to be under-represented on the X chromosome , and many testes-specific genes have been “retrotransposed , ” or moved , from the sex to autosomal chromosomes . However , compelling evidence for X chromosome inactivation in the fruit fly Drosophila has been lacking . Here , we used a transgenic technique to test for male germline X inactivation in this important model organism . We randomly inserted a “reporter gene” whose expression requires a regulatory element for an autosomal testis-specific gene into multiple autosomal and X-chromosomal locations . We found that autosomal insertions of the reporter gene have significantly higher expression in the male germline than X-linked insertions . This pattern holds for two different transgenes with nearly 50 independent insertions , providing strong evidence for X chromosome inactivation during spermatogenesis . The silencing of X-linked gene expression in the male germline may contribute to the observed paucity of male-expressed genes on the X chromosome and the excess of retrotransposed genes that have moved from the X chromosome to the autosomes .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] |
[
"genetics",
"and",
"genomics",
"evolutionary",
"biology"
] |
2007
|
X Chromosome Inactivation during Drosophila Spermatogenesis
|
The Gram-positive , spore-forming pathogen Clostridium difficile is the leading definable cause of healthcare-associated diarrhea worldwide . C . difficile infections are difficult to treat because of their frequent recurrence , which can cause life-threatening complications such as pseudomembranous colitis . The spores of C . difficile are responsible for these high rates of recurrence , since they are the major transmissive form of the organism and resistant to antibiotics and many disinfectants . Despite the importance of spores to the pathogenesis of C . difficile , little is known about their composition or formation . Based on studies in Bacillus subtilis and other Clostridium spp . , the sigma factors σF , σE , σG , and σK are predicted to control the transcription of genes required for sporulation , although their specific functions vary depending on the organism . In order to determine the roles of σF , σE , σG , and σK in regulating C . difficile sporulation , we generated loss-of-function mutations in genes encoding these sporulation sigma factors and performed RNA-Sequencing to identify specific sigma factor-dependent genes . This analysis identified 224 genes whose expression was collectively activated by sporulation sigma factors: 183 were σF-dependent , 169 were σE-dependent , 34 were σG-dependent , and 31 were σK-dependent . In contrast with B . subtilis , C . difficile σE was dispensable for σG activation , σG was dispensable for σK activation , and σF was required for post-translationally activating σG . Collectively , these results provide the first genome-wide transcriptional analysis of genes induced by specific sporulation sigma factors in the Clostridia and highlight that diverse mechanisms regulate sporulation sigma factor activity in the Firmicutes .
Clostridium difficile is a Gram-positive , spore-forming , obligate anaerobe that causes gastrointestinal diseases including diarrhea , pseudomembranous colitis , and toxic megacolon [1]–[3] . C . difficile infections and C . difficile-related deaths have risen dramatically in the past decade , increasing the financial burden on health care systems [4]–[7] . While C . difficile is best known for causing hospital-acquired antibiotic-associated infections , recent epidemiologic studies indicate that community-acquired C . difficile infections are increasingly more common and associated with significant morbidity [6] , [7] . A key element to the success of C . difficile as a pathogen is its ability to produce spores . Spores are resistant to most disinfectants and antibiotics , making them difficult to eliminate both from infected humans and the environment [1] , [2] , [8] . As a result , C . difficile spores disseminate readily from person to person and cause high rates of recurrent infections , which can lead to serious illness or even death [1]–[3] , [9] . Although spores are critical to the pathogenesis of C . difficile , their composition and formation remain poorly characterized . Less than 25% of the spore coat proteins identified in the well-characterized spore-former Bacillus subtilis have homologs in C . difficile [10] . In contrast , the regulatory proteins that control spore coat gene expression and other sporulation events in B . subtilis are conserved in C . difficile and all other spore-forming Firmicutes [10]–[13] . These include the master sporulation transcriptional regulator , Spo0A , and the sporulation sigma factors σF , σE , σG , and σK . In B . subtilis the sporulation sigma factors function at discrete stages during spore development to couple changes in gene expression with specific morphological changes in the cell [14]–[16] . The morphological changes begin with the formation of a polar septum , which creates two compartments , the mother cell and the forespore . The mother cell engulfs the forespore and guides the assembly of the spore until it lyses once spore maturation is complete . By coupling these developmental changes to the sequential activation of compartment-specific sporulation sigma factors , the mother cell and forespore produce divergent transcriptional profiles that coordinately lead to the formation of a dormant spore [16] . Sporulation gene transcription in B . subtilis begins with the activation of the transcription factor Spo0A , which in turn activates early sporulation gene transcription , such as the genes encoding the early sigma factors σF and σE . σF is initially held inactive by an anti-σ factor and only undergoes activation after septum formation is complete; this mode of regulation couples σF activation in the forespore to a morphological event [17] , [18] . Active σF induces the transcription of genes whose products mediate cleavage of an inhibitory pro-peptide from σE in the mother cell via trans-septum signaling [19] . Active σE induces the transcription of genes whose products lead to the activation of the late sporulation sigma factor σG in the forespore , which occurs during or after engulfment [20] , [21] . Activated σG in the forespore subsequently induces the expression of genes whose products proteolytically activate σK in the mother cell via trans-septum signaling [22] . Notably , the activity of each sigma factor relies on the activation of the preceding sigma factor [11] , [14]–[16] , [23] . As a result , the sigma factors operate in a sequential , “criss-cross” manner and collectively control the expression of hundreds of genes during sporulation [24]–[26] . The regulatory pathway controlling sporulation sigma factor activation in B . subtilis is thought to be conserved across endospore-forming bacteria , since all four sigma factors are conserved [11] , [12] . However , a growing body of work in the Clostridia suggests that diverse pathways regulate sporulation sigma factor activity in the Firmicutes . In C . perfringens , a sigG− mutant still produces cleaved σK , suggesting that σG does not control the proteolytic activation of σK as it does in B . subtilis [27] . Furthermore , a C . perfringens sigK− mutant exhibits a phenotype more severe than a B . subtilis sigE− mutant in that it fails to initiate asymmetric division or produce σE [28] , suggesting that in C . perfringens σK functions upstream of σE . Indeed , C . perfringens σE and σK have been suggested to be dependent on each other for full activity , in contrast with B . subtilis [28] . A similar early sporulation defect has been observed in a sigK− mutant of C . botulinum , which also exhibits reduced expression of early sporulation genes spo0A and sigF [29] . In contrast with B . subtilis and C . perfringens , however , a C . acetobutylicum sigF− mutant does not initiate asymmetric division [30] , and a sigE− mutant fails to complete asymmetric division [31] . In addition , a C . acetobutylicum sigE− mutant produces wildtype levels of σG [31] in contrast with B . subtilis , and a sigG− mutant exhibits elongated forespores and pleiotropic defects in coat and cortex formation [31] . To determine how these sporulation sigma factors regulate sporulation in C . difficile , we constructed mutations in the genes encoding the sporulation transcription factor Spo0A and the sigma factors σF , σE , σG , and σK and determined the transcriptional profiles of these mutants using RNA-Sequencing ( RNA-Seq ) . The transcriptional analyses , combined with cytological characterization of the sigma factor mutants , suggest that divergent mechanisms regulate the activity of σG and σK in C . difficile relative to B . subtilis and other Clostridium spp . In addition , these analyses have identified a set of 314 genes that are upregulated during sporulation in a Spo0A- , σF- , σE- , σG- , and/or σK-dependent manner . These sporulation-induced genes provide a framework for identifying and characterizing C . difficile spore proteins that may have diagnostic or therapeutic utility .
In order to identify genes that are regulated by the sporulation-specific sigma factors , we used a modified TargeTron gene knockout system to disrupt the genes encoding σF , σE , σG , and σK in C . difficile [32] . This system uses a group II intron to insert an erythromycin resistance cassette into the target gene ( Figure S1A ) . JIR8094 [33] , an erythromycin-sensitive derivative of the sequenced C . difficile strain 630 [34] , was used as the parental strain . As a control , we also constructed a targeted disruption in spo0A , which encodes the master regulator of sporulation Spo0A [35] , [36] . Colony PCR of the intron-disrupted mutants confirmed the expected size change resulting from the intron insertion into the spo0A , sigF , sigE , sigG , and sigK genes ( Figure S1B ) . To determine the effect of blocking sigma factor production on sporulation , the mutants were induced to sporulate on solid sporulation media and visualized by phase contrast microscopy [37] . It should be noted that sporulation is asynchronous in this assay , and the extent and timing of sporulation exhibits variability even between biological replicates ( Figure S2 ) . Nevertheless , after 18 hrs of growth , sufficient numbers of cells have initiated sporulation to detect the production of immature phase-dark forespores and mature phase-bright spores in the wildtype strain ( Figure 1 and S2 ) . In contrast , spo0A− , sigF− , sigE− , sigG− , and sigK− cultures failed to produce phase-bright spores ( Figure 1 ) . No phase-dark or phase-bright forespores were observed in the spo0A− , sigF− , or sigE− strains , suggesting a block early in sporulation . Analysis of live , sporulating cultures with the lipophilic dye FM4-64 ( to stain mother cell and forespore membranes ) and Hoechst 33342 ( to stain cell nucleoids ) revealed polar septum formation in wild type and the sigma factor mutants but not in the spo0A− mutant ( Figure 1 ) . This result was consistent with the observation that Spo0A is necessary to induce the sporulation pathway in C . difficile [35] , [36] . Overall , the proportion of sporulating cells detected by membrane and DNA staining in the culture was 25% , 41% , 24% , 26% , and 18% for wildtype , sigF− , sigE− , sigG− , and sigK− , respectively , as indicated by the presence of a polar septum , immature forespore compartment , or mature forespore ( Table S1 ) . Wildtype cultures contained a heterogenous population of sporulating cells at discrete stages of sporulation: 28% of sporulating cells exhibited intense DNA staining of an FM4-64-labeled forespore compartment ( yellow arrows , Figure 1 , Table S1 ) ; 28% showed phase-dark forespores that stained with both FM4-64 and Hoechst ( Table S1 ) , 28% exhibited phase-dark forespores that stained intensely with FM4-64 but not Hoechst ( green arrows , Figure 1 , Table S1 ) , and 16% contained a phase-bright forespore that failed to be stained with either FM4-64 or Hoechst ( pink arrows , Figure 1 , Table S1 ) . In contrast , sigF− and sigE− sporulating cells were arrested at the asymmetric division stage , with 95% and 92% of sporulating cells , respectively , exhibiting intense DNA staining of an FM4-64-labeled forespore compartment ( yellow arrows , Figure 1 , Table S1 ) . The sigG− mutant strain was arrested at the phase-dark forespore stage , with 69% of sporulating cells exhibiting intense forespore membrane and nucleoid staining ( yellow arrows , Figure 1 , Table S1 ) . While only 4% of the sigG− cells were observed to produce forespores that stained only with FM4-64 , 44% of sporulating sigK− cells were captured at this stage of sporulation , a phenotype that was also observed in wildtype ( green arrows , Figure 1 , Table S1 ) . Taken together , these results indicate that all four sporulation sigma factors are required to complete spore formation and suggest that σG is necessary to complete the stage of sporulation development required to exclude the Hoechst dye from staining the forespore chromosome . The results are also consistent with studies investigating B . subtilis forespore development , which indicate that nucleic acid stains are excluded earlier than membrane stains during spore development [38]–[40] . To confirm that the gene disruptions prevented sigma factor production in each of the respective sigma factor mutants , we performed Western blot analyses using antibodies raised against C . difficile sigma factors . Similar to B . subtilis , Spo0A was required for the production of all the factors , and σF was observed in the sigE− , sigG− , and sigK− strains at wildtype levels ( Figure 2 , [41] ) . σE was detected in both its pro- and cleaved form in wildtype , sigG− and sigK− strains , whereas the majority of σE was unprocessed in the sigF− strain ( Figure 2 ) . This result slightly deviates from the B . subtilis model , where pro-σE processing is completely abrogated in a B . subtilis sigF− strain [42] . In contrast , a C . perfringens sigF− mutant fails to produce pro-σE altogether [27] , and σE processing has not been demonstrated in C . acetobutylicum [31] . σK was present in wildtype and sigG− mutant strains but absent in the sigF− and sigE− strains ( Figure 2 ) , analogous to observations in B . subtilis where σE is required for sigK expression . A C . perfringens sigE− strain in contrast produces low amounts of σK [28] . Consistent with the observation that C . difficile σK lacks an N-terminal pro-peptide [43] , no processing of σK was observed in wildtype C . difficile ( Figure 2 ) , even though σK undergoes proteolytic activation in B . subtilis and C . perfringens [27] . σG was detected in the C . difficile sigF− , sigE− and sigK− mutants ( Figure 2 ) in contrast with studies of other endospore-forming bacteria , where σG activity and auto-activation of sigG transcription is partially dependent on σE in B . subtilis [44]–[46] , and σG production depends on σF in C . perfringens and C . acetobutylicum [27] , [30] . We next performed transmission electron microscopy ( TEM ) to identify the precise developmental stage at which each sigma factor mutant was stalled . Cortex and coat layers were present on forespores in wildtype sporulating cells , while the spo0A− mutant exhibited no signs of spore formation ( Figure 3 ) . The sigF− mutant failed to progress beyond asymmetric division ( Figure 3 ) , similar to a B . subtilis sigF− mutant [47] but in contrast with a C . acetobutylicum sigF− mutant which does not initiate asymmetric division [30] . Nevertheless , unlike B . subtilis , a more electron-translucent region in the mother cell cytosol surrounded by electron dense layers was observed in some sigF− mutant cells; this region resembled mislocalized spore coat ( [37] , Figure 3 ) . The C . difficile sigE− strain was arrested at the asymmetric division stage similar to the sigF− mutant , although electron-translucent regions surrounded by coat-like layers were not observed in any sigE− cell analyzed . The C . difficile sigE− mutant phenotype resembled the phenotype of sigE− mutants of B . subtilis [47] and C . perfringens [28] , with frequent observations of disporic cells or cells with multiple septa at one pole ( Figures 1 and 3 ) . This observation was in contrast with a C . acetobutylicum sigE− mutant , which does not complete asymmetric division [31] . The C . difficile sigG− mutant produced forespores lacking an apparent cortex layer , similar to B . subtilis [21] , [44]; however , unlike B . subtilis , the forespores were surrounded by thin layers that resembled the spore coat layers visible in wildtype cells ( Figure 3 ) . In addition , the C . difficile sigG− mutant exhibited pleiotropic defects including forespore ruffling , incomplete membrane fission during engulfment , and a septated forespore compartment ( Figures 3 and S3 ) . Quantitation of the prevalence of each phenotype revealed that forespore ruffling , incomplete engulfment , and a septated forespore compartment were observed in 98 , 87 and 21% of sigG− cells , respectively ( Figure S3 ) . Lastly , the C . difficile sigK− mutant produced forespores surrounded by a layer that resembled the cortex layer of wildtype , but no coat layers were apparent ( Figure 3 ) . This phenotype was more similar to a B . subtilis sigK− mutant , which lacks both cortex and coat [22] , than C . perfringens , which fails to initiate polar septum formation [28] . To validate that the observed mutant phenotypes were due to the targeted insertions , we complemented the mutant strains by expressing a wildtype copy of the gene encoding the corresponding sigma factor from a plasmid . We used either the pMTL83151 or pMTL84151 multicopy plasmids [48] to express the complementing genes or operons from their native promoters . The complementation constructs all restored production of phase-bright spores when expressed in their respective mutant backgrounds ( Figure S4A ) , although phase-bright spore formation by the sigK complementation strain was delayed relative to wildtype . Western blot analysis further confirmed that the complementation constructs restored production of the respective sigma factor to wildtype levels ( Figure S4B ) . TEM analysis revealed that all four complementation constructs restored coat and cortex formation to their respective mutant strains ( Figure S5 ) . Heat resistance assays to measure complementation strain sporulation efficiency revealed that the sigF− and sigE−constructs fully complemented heat resistance relative to wildtype and that the sigG− and sigK− constructs partially complemented heat resistance ( 70 and 23% , respectively , Figure S4C ) . While these analyses showed that σF , σE , σG , and σK were all required for mature spore formation , they did not reveal which genes were being misregulated in the sporulation sigma factor mutants to produce their respective sporulation defects . To identify these genes and gain insight into the regulatory network controlling sporulation sigma factor activity , we used RNA-Sequencing ( RNA-Seq ) to transcriptionally profile our sporulation mutants and wild type during sporulation . Three biological replicates of wildtype , spo0A− , and sporulation sigma factor mutant strains were grown on sporulation media ( Figure S2 ) , and RNA was isolated . Following DNase-treatment , ribosomal RNA depletion and reverse transcription , Illumina-based RNA-Seq was used to determine the complete transcriptome of wildtype C . difficile and the sporulation mutants . Genome coverage and sequencing counts for each strain and replicate can be found in Table S2 . The DeSeq variance analysis package [49] was used to identify genes that were downregulated by ≥4-fold with an adjusted p-value of ≤0 . 05 in the spo0A− strain relative to wild type . This pair-wise analysis identified 276 genes as being Spo0A-dependent ( Table S3 ) . Consistent with the role of Spo0A as the master regulator of sporulation , 65 of these genes were predicted to be involved in sporulation ( Table S4 ) [11] , [50]–[52] . Six of these Spo0A-dependent genes were recently identified as encoding components of the C . difficile spore coat [50] , [53] , and 36 sporulation-related genes ( Table S4 ) were shown to depend on σH , the stationary phase sigma factor that induces spo0A transcription in C . difficile [51] and B . subtilis [54] . σF- , σE- , σG- , and σK-dependent genes were identified by comparing the transcriptional profiles of the sigF− , sigE− , sigG− , and sigK− strains to wild type , respectively , using the same parameters as above . This analysis identified 183 genes as being dependent on σF for their expression ( Table S5 ) . One hundred eighteen of these σF-dependent genes were also σE-dependent ( Table S6 ) , indicating that σE has some activity in a sigF− mutant consistent with the reduced levels of cleaved σE being detected by Western blot ( Figure 2 ) ; 29 of the σF-dependent genes formed a separate subset of genes that were also σG-dependent but σE-independent . Indeed , the majority of the 34 σG-dependent genes identified in this analysis were not dependent on σE for their expression ( Table S7 ) , since only four of the σG-regulated genes were also σE-regulated . Notably , none of the genes identified as being σG-dependent required σK for their expression ( Table S8 ) , suggesting that the σG produced in the sigE− and sigK− mutants is active ( Figure 2 ) . This result differs from the B . subtilis model where σE is needed to fully activate σG function [20] , [21] , [46] , [55] , [56] . Of the 169 genes that depended on σE for their expression ( Table S6 ) , 85% and 78% of these genes were dependent on Spo0A and σF , respectively ( Figure 4 ) . The expression of 29 of these genes was also σK-dependent ( Table S5 ) . Indeed the majority of the 31 σK-dependent genes were σE-dependent ( Table S8; Figure 4 ) , consistent with σE being required for σK production ( Figure 2 ) . In contrast , as described earlier , no overlap was observed between σG- and σK-dependent genes ( Figure 4 ) . Taken together , the RNA-Seq analyses suggested that ( 1 ) a small subset of σF-dependent genes are neither σE , σG , nor σK-dependent; ( 2 ) σE activity depends on Spo0A and σF but not σG or σK; ( 3 ) σK activity depends on Spo0A , σF , and σE but not σG , and ( 4 ) σG activity depends on Spo0A and σF but not σE or σK . The latter two findings differ from the B . subtilis model , where the σK-dependent genes are also σG-dependent because σK activity depends on σG [11] , [15] , [22] , and σG-dependent genes are σE-dependent because full activation of σG requires σE [20] , [21] , [46] , [55] , [56] . To visually represent the differences in gene expression profiles between the sigma factor mutants and wild type , we generated a heat map for genes downregulated by ≥4-fold with an adjusted p-value of ≤10−5 in the spo0A− strain relative to wild type . The expression levels of wild type and the sigma factor mutants relative to spo0A− strain were centered , scaled , and mapped to a red-green color scale . The heat map revealed a cluster of genes that was poorly expressed in the sigE− mutant relative to the wildtype , sigG− , and sigK− strains; these genes were also expressed at reduced levels in the sigF− mutant ( Figure 5 ) and were primarily σE-dependent ( Table S4 ) . A separate cluster of genes was downregulated in both the sigK− and sigE− mutants relative to the wildtype and sigG− strains ( Figure 5 ) ; these genes were all identified as σK-dependent genes ( Table S5 ) . Another discrete cluster of genes was downregulated in the sigG− and sigF− strains relative to the wildtype , sigE− , and sigK− strains ( Figure 5 ) ; again , most of these genes were identified as σG-dependent genes , although two genes were σF-dependent but not σG-dependent ( Tables S5 and S7 ) . Thus , identification of variably expressed genes between the strains confirmed the findings of our earlier pair-wise analyses: σF-dependent genes were largely Spo0A-dependent , σE-dependent genes were largely σF-dependent , σK-dependent genes were σE-dependent , and σG-dependent genes were σF-dependent but not σE- or σK-dependent . These results support a model where ( 1 ) σF controls the activation of both σE and σG , ( 2 ) σE induces the production and activation of σK , and ( 3 ) σE and σK are dispensable for σG activation . Alternative statistical models were also employed to validate these findings ( see Text S1 and Figures S6 and S7 ) . To validate the RNA-Seq data , we isolated RNA from three separately prepared biological replicates of wildtype , spo0A− , sigF− , sigE− , sigG− , and sigK− strains grown on sporulation media for 18 hrs . RNA was reverse transcribed and quantitative RT-PCR ( qRT-PCR ) was performed using primers specific for three genes within each of the sigma factor-dependent transcriptomes . Gene expression levels in the wildtype and the sigma factor mutant strains relative to spo0A− were determined by comparative CT analysis normalized to the housekeeping gene rpoB . These analyses confirmed that the transcript levels of the σF-dependent gene gpr was reduced by >50-fold ( p<0 . 0001 ) in the sigF− mutant relative to wild type , and reduced in the sigG− mutant by ∼4 fold ( p<0 . 01 ) ; gpr expression was not affected in the sigE− and sigK− mutants . cd0125 ( spoIIQ , [13] ) transcription was reduced by >10-fold in the sigF− mutant relative to wild type ( p<0 . 01 ) , but no reduction in transcript levels was observed in sigE− , sigG− , and sigK− mutants ( Figure 6A ) . Transcription of cd2376 was reduced by 3-fold in the sigF− relative to wild type ( Figure 6A ) . Although this correlation was not statistically significant , it approached statistical significance ( p = 0 . 065 ) ( Figure 6A ) ; this result is likely due to the low number of overall cd2376 transcripts present in the samples . Transcript levels of the σG-dependent genes spoVT , sspB , and dacF showed significant reductions in the sigG− ( p<0 . 0004 , <0 . 0002 and <0 . 0001 , respectively ) and sigF− mutants ( p<0 . 0001 ) compared to wild type but no significant reduction in the sigE− and sigK− mutants relative to wild type ( Figure 6B ) . This observation was consistent with the RNA-Seq data indicating that σG activity depends on σF , although it is likely that σF directly induces the transcription of some σG-dependent genes given the predicted overlap in their promoter specificities [11] . Nevertheless , given that σG is present at wildtype levels in a sigF− strain , these observations suggest that σF regulates σG activity through a post-translational mechanism . σE-dependent genes cd3522 and spoIVA were reduced by >100-fold , and cd1511 by >50-fold , in sigE− relative to wild type , ( p<0 . 0001 , <0 . 0001 , and <0 . 006 , respectively ) , but not in sigG− and sigK− mutants ( Figure 6C ) . Transcript levels of these σE-dependent genes were reduced by ∼5 to 6-fold ( p<0 . 01 ) in the sigF− mutant relative to wildtype , indicating that , in the absence of σF , σE activity is reduced but detectable . Transcript levels of the σK-dependent genes cd1433 , cd1067 and sleC were significantly reduced by >100-fold in the sigE− ( p<0 . 0001 for each gene ) and the sigK− ( p<0 . 0001 for each gene ) strains compared to wild type ( Figure 6D ) . σK-dependent gene expression was reduced in the sigF− mutant by 8 to 10-fold ( p<0 . 01 ) , suggesting that σK has reduced but detectable activity in the sigF− strain . Importantly , no statistically significant change for any of these σK-dependent genes was observed in the sigG− mutant relative to wild type , consistent with the RNA-Seq results indicating that σK activity does not depend on σG ( Figures 4 and 5 ) . Altogether , the qRT-PCR data validated the RNA-Seq data identifying σF , σE , σG , and σK-dependent genes and confirmed that ( 1 ) σE , σG , and σK activity depend on σF , ( 2 ) full σG activity requires σF but not σE , and ( 3 ) σK activity requires σE but not σG . It should be noted however that , although σF is required for full σE and σK activity , some degree of σE- and σK-dependent gene expression is observed even in the absence of σF . Having validated the RNA-Seq data at the transcript level , we next investigated whether changes in transcript levels correlated with changes in protein levels for σF- , σE- , σG- , and σK-regulated genes . To this end , we raised antibodies against proteins encoded by genes identified by RNA-Seq as being σF- , σE- , σG- , and σK-dependent . Western blot analyses of the germination protease Gpr confirmed that only σF is required for gpr expression , while production of the regulatory protein SpoVT and the small acid-soluble protein SspA depended on both σF and σG . These results indicate that σG can directly activate the expression of spoVT and sspA ( Figure 7 ) . Western blot analyses for CD3522 , SpoIVA , and CD1511 demonstrated that their production depends on σE but not σG or σK; these proteins were detected , albeit at greatly reduced levels , in the sigF− mutant ( Figure 7 ) . These results were consistent with the observation that active , processed σE is present in both sigG− and sigK− strains , while only trace amounts of processed σE could be detected in the sigF− strain ( Figure 2 ) . Analysis of σK-dependent protein production using antibodies specific for CD1433 , CD1067 and SleC confirmed that these proteins were absent in the sigE− and sigK− mutants and present in wild type and the sigG− mutant ( Figure 7 ) . Only SleC was reliably detected in the sigF− mutant , even though cd1433 and cd1067 transcripts could be detected in the sigF− strain ( Figure 6D ) . Nevertheless , taken together these observations confirm that ( 1 ) σF does not require σE , σG , or σK for activation , ( 2 ) full σE activation requires σF , ( 3 ) full σG activation requires σF but not σE or σK , and ( 4 ) σK activation requires σF and σE but not σG .
While mutation of all four sporulation sigma factors in C . difficile abrogated functional spore formation as expected [11] , the regulation and function of these sigma factors in C . difficile differed from the regulatory pathways determined for B . subtilis and other Clostridium spp . The differences between C . difficile , C . perfringens , C . acetobutylicum , and B . subtilis sporulation pathways are summarized in Figure 8 , as are the similarities . Similar to B . subtilis , our transcriptional and cytological analyses reveal that C . difficile σK functions downstream of σE to regulate late-stage sporulation events , and σG functions downstream of σF to regulate forespore maturation ( Figures 2 and 6 ) . In contrast with B . subtilis , C . difficile σG is fully active in the absence of σE , and σK is fully active in the absence of σG ( Figures 6 and 7 ) . The latter observation could have been anticipated given that C . difficile σK lacks an N-terminal pro-peptide , in contrast to all other spore formers [43] . However , the former observation was unexpected because σE-regulated gene products function to activate σG in the forespore of B . subtilis , initiating a positive feedback pathway that increases σG levels through auto-activation of the sigG promoter [44] , [46] , [57] . In particular , B . subtilis σG activation requires the formation of a σE-dependent “feeding tube” [20] , [21] , [55] , [58] , [59] , which maintains forespore integrity by transporting small molecules from the mother cell into the forespore [20] , [21] , [55] . This mode of regulation couples the activation of the forespore-specific σG to σE-controlled events in the mother cell . In contrast , our results indicate that C . difficile σG is active in the absence of σE-dependent feeding tube gene expression ( Figures 5 and 6 , Tables S6 and S7 ) . Nevertheless , even though σG was active at wildtype levels in the sigE− mutant ( Figures 6 and 7 ) , it remains possible that σG activity may be mislocalized in the mother cell cytosol , similar to the premature activation of σG in Lon− and anti-σG sigma factor CsfB− cells [57] , [60] , [61] . Even though C . difficile σG can be fully activated in the absence of σE , our results further show that σG is post-translationally activated in a σF-dependent manner ( Figures 2 and 6 ) . These results raise the intriguing question as to how σF activates σG independent of σE in C . difficile . In B . subtilis , multiple post-translational mechanisms control σG activity; however , aside from the feeding tube , these mechanisms are inhibitory rather than activating . In B . subtilis the Lon protease reduces σG activity in the mother cell [60] , while the anti-σ factors SpoIIAB [57] , [62] and CsfB ( also known as Gin ) [61] , [63] , [64] prevent σG activity in the forespore until engulfment is complete . Whether these factors inhibit σG activity in C . difficile is unknown , although C . difficile does not appear to encode a CsfB homolog . In future studies , it will be interesting to determine whether σF functions to activate σG directly or alleviate its inhibition , and whether C . difficile sporulation sigma factors exhibit compartment-specific activity similar to B . subtilis . Interestingly , the morphology of the C . difficile sigG− mutant differed considerably from a B . subtilis sigG− mutant . While B . subtilis sigG− mutant forespores are normal in appearance despite lacking both a coat and cortex [44] , C . difficile sigG− mutant forespores produced layers resembling spore coat around the forespore and exhibited defects in engulfment and structural integrity ( Figures 3 and S3 ) . The forespore membrane ruffling phenotype of C . difficile sigG− mutants was reminiscent of B . subtilis feeding tube mutant phenotypes [21] , suggesting that σG may encode proteins required to “nurture” the C . difficile forespore . Alternatively , σG could regulate a cytoskeletal or cortex component that confers structural integrity to the forespore . Such proteins could be represented in the σG-regulated genes identified in this study ( Table S7 ) . The phenotype of the C . difficile sigF− mutant also differed from its cognate mutant in B . subtilis , since the sigF− mutant produced low levels of σE− and σK− induced gene products ( Figure 7 ) and regions that resembled mislocalized coat in the mother cell cytosol ( Figure 3 ) [47] . In B . subtilis , σF is required to activate the expression of spoIIR , which encodes an intercellular signaling protein that activates SpoIIGA , the protease responsible for activating pro-σE [65] , [66] . Whether the trace amounts of σE processing observed in the C . difficile sigF− mutant results from low-level expression of spoIIR or spoIIGA , or whether an unknown protease activates σE , remains to examined . Comparison of the sporulation pathway of C . perfringens relative to C . difficile indicates that both organisms proteolytically activate σE in a σF-dependent manner ( Figure 2 ) , although it should be noted that a C . perfringens sigF− mutant does not make σE , σG , or σK [27] in contrast with C . difficile ( Figure 2 ) . Since the phenotypes of C . perfringens sigF− and sigG− mutants have not been examined by electron microscopy , the precise stage at which they are arrested remains unclear . Nevertheless , unlike C . perfringens ( and C . botulinum ) where σK is essential for both early and late stage sporulation events ( Figure 8 ) [28] , [29] , C . difficile σK is needed only at late stages of sporulation . Furthermore , C . perfringens σK is produced at low levels in an unprocessed form in a sigE− mutant; σE is made at low levels in a C . perfringens sigK− mutant; and sigE and sigK expression appear to be auto-activated [28] . In contrast , no sigK expression was observed in the absence of σE in C . difficile . The sporulation pathway of C . difficile appears to be most similar to the C . acetobutylicum pathway . Both C . difficile and C . acetobutylicum sigma factors σF , σE , and σG appear to function at similar stages of sporulation , although C . acetobutylicum sigF− and sigE− mutants exhibit more severe phenotypes than in C . difficile in that they fail to initiate and complete asymmetric division , respectively [30] , [31] , and σF is required to activate sigG transcription in C . acetobutylicum [30] unlike C . difficile . Similar to C . difficile , however , C . acetobutylicum σG does not require σE for auto-activation of sigG expression , although it is unclear whether C . acetobutylicum σG is active in the absence of σE [31] . Lastly , loss of sigG expression in C . acetobutylicum results in pleiotropic defects in coat and cortex formation and forespore integrity similar to C . difficile ( Figure 3 , [31] ) . Since a sigK− mutant in C . acetobutylicum has not been described , it will be interesting to determine whether C . acetobutylicum σK function is more similar to C . difficile than to C . perfringens and whether these differences correlate to the presence of the skin element , an ∼15 kb prophage-like element that interrupts the sigK gene in C . difficile but not other Clostridium spp . [43] . Nevertheless , our collective transcriptional and cytological analyses of sporulation sigma factor function in C . difficile suggest that novel mechanisms regulate σG and σK activation relative to other spore-forming organisms ( Figure 8 ) . Further studies are needed to determine the regulatory interplay between C . difficile sporulation sigma factors and their downstream auxiliary regulators SpoVT and SpoIIID , which modulate the expression of σG- and σK-regulated genes , respectively , in B . subtilis [24] , [26] , [67] , [68] and are conserved in Clostridium spp . By performing whole genome transcriptional profiling on the different sporulation sigma factor mutants , we have identified distinct subsets of genes that are σF- , σE- , σG- , and σK-dependent . The number of genes determined to be σG-dependent in C . difficile was relatively small ( 34 genes ) relative to B . subtilis , where the σG regulon comprises ∼100 genes [11] , [25] , [26] . Similarly , the σE and σK-dependent genes ( 169 and 31 genes , respectively ) identified by our study were smaller than their cognate regulons in B . subtilis ( 270 and 150 genes , respectively , [24] ) . While the parameters we used to define genes as being σF- , σE- , σG- , and σK-dependent were relatively stringent , relaxing these parameters did not result in large increases in gene numbers . One explanation for the smaller size is that C . difficile activates fewer genes during sporulation than B . subtilis . A more likely explanation is that the asynchronous population of sporulating cells ( Figures 1 and S2 , [37] ) limits the detection of genes that are transiently expressed during discrete stages of sporulation or genes that are expressed at low levels during sporulation [24]–[26] , [69] . Since the RNA samples used in the RNA-Seq analysis were harvested from a sporulation timepoint in which phase-bright forespores were produced by wildtype cells ( Figure S2 ) , fewer cells in the population are likely to be at early stages sporulation . As a result , early sporulation genes may be under-represented in our data set; for example , σF-dependent early sporulation gene transcripts from spoIIR and spoIIP were almost undetectable by RNA-Seq ( Table S4 ) . In addition , genes that are regulated by more complex mechanisms beyond upregulation by a specific sigma factor are likely to be under-represented in our data set . Sporulation genes that are subject to incoherent feed forward loop regulation , in which their expression is induced by a given sigma factor and repressed by a downstream regulator such as SpoVT-mediated inhibition of sigG transcription in B . subtilis [26] , may not be detected in our data set . Unraveling the complexities of sporulation gene regulation in C . difficile will require further characterization of the kinetics of sporulation and the analysis of mutants defective in auxiliary sporulation regulators . Of the 51 genes proposed to comprise the core set of sporulation genes in spore-forming Firmicutes by bioinformatics analyses that are conserved in C . difficile [13] , 34 were identified in our RNA-Seq analyses , leaving 17 that were not identified in our transcriptional analyses ( Table S4 ) . Seven of these genes do not have detectable homologs in the 630 C . difficile genome , and 6 were expressed at low levels with a base mean of expression less than 15 ( Table S4 ) . Although some sporulation-induced genes were likely not detected in our analyses due to low levels of expression , the transcriptional profiling data presented here identify a promising set of genes that are likely to encode proteins with important roles in spore formation . Of the six spore coat proteins recently identified in a proteomic analysis of C . difficile [50] , [53] , all were identified in our RNA-Seq experiments . Three of these spore coat genes were determined to be σK-dependent , consistent with their predicted role as components of the outer coat ( Table S4 ) . Notably , σK-regulated genes were among the most abundantly expressed genes induced during sporulation , comprising 6 of the 10 most highly expressed sporulation genes ( Table S9 ) . The σK-regulated CD1067 gene was the most highly expressed gene induced during sporulation in C . difficile . Cysteine-rich CD1067 was also one of the most abundant spore proteins identified in proteomic analyses of purified spores and is encoded in a 7 . 5 kB region enriched in genes encoding spore proteins [70] . Western blot analyses of cysteine-rich CD1067 revealed that it forms higher order multimers that are highly resistant to denaturing conditions ( data not shown ) , consistent with the proposal that CD1067 may form a rigid , disulfide-bonded structure around the spore coat upon exposure to atmospheric oxygen , for example during excretion from the host [70] . Intriguingly , CD1067 is encoded in a region enriched in highly expressed , σK-regulated genes encoding hypothetical proteins unique to C . difficile , with 8 of the 9 genes in this region being induced during sporulation and 6 of the 9 being σK-regulated . These genes may encode coat proteins that confer structural integrity and/or resistance to the C . difficile spore coat and thus may play important roles in disease transmission and/or represent good candidates for developing diagnostic reagents . Although the number of σG-dependent genes identified by our study was small , a number of these genes encode proteins with important functions in the forespore of B . subtilis , specifically sspA , sspB , dacF , spoVT , and spoVAD [26] , [67] , [71]–[73] . Since B . subtilis σG induces the expression of genes encoding the germinant receptors ( of which there are no homologs in C . difficile [12] , [34] ) , it seems likely that some of the σG-dependent genes identified in our study encode proteins that transduce the germinant signal into the spore core . It will be interesting to determine whether any of the σG-regulated genes identified in our study play important roles in regulating germination and thus disease transmission . Genes encoding hypothetical proteins were the most abundant class of genes identified in our study ( 82 in total , Table S10 ) . Twenty of these hypothetical proteins were detected in proteomic analyses of C . difficile spores [70] . Indeed , two of the hypothetical proteins were previously shown to be part of the spore coat [50] , and we have validated three additional proteins as localizing to the spore coat ( data not shown ) . BLAST searches with the hypothetical proteins identified by RNA-Seq indicate that 16 have no known homologs . These C . difficile-specific proteins could comprise part of the spore coat , since coat proteins are often poorly conserved , species-specific , and categorized as hypothetical proteins [10] , [11] . Taken together , by examining the regulatory interplay between sporulation sigma factors in C . difficile , our study highlights that diverse pathways regulate sporulation in the Firmicutes and that considerable work is needed to map these pathways in the Clostridia . By using whole genome transcriptional profiling to define a large set of genes that are activated by Spo0A , σF , σE , σG , and/or σK , our study also provides a framework for identifying new proteins that are necessary for sporulation and determining the role of these proteins in forming a functional , infectious spore . Studies of this nature may lead to the identification of biomarkers for C . difficile spores and candidates for vaccine development .
All C . difficile strains are listed in Table 1 and derive from the parent strain JIR8094 [33] , an erythromycin-sensitive derivative of the sequenced clinical isolate 630 [34] . C . difficile strains were grown on solid brain heart infusion media supplemented with yeast extract ( BHIS: 37 g brain heart infusion , 5 g yeast extract , 0 . 1% ( w/v ) L-cysteine , 15 g agar per liter ) [74] . Taurocholate ( TA; 0 . 1% w/v ) , thiamphenicol ( 5–10 µg/mL ) , kanamycin ( 50 µg/mL ) , cefoxitin ( 16 µg/mL ) , FeSO4 ( 50 µM ) , and/or erythromycin ( 10 µg/mL ) were used to supplement the BHIS media as indicated . Cultures were grown at 37°C , under anaerobic conditions using a gas mixture containing 85% N2 , 5% CO2 , and 10% H2 . Sporulation was induced on media containing BHIS and SMC ( 90 g BactoPeptone , 5 g protease peptone , 1 g NH4SO4 , 1 . 5 g Tris base , 15 g agar per liter ) [50] , at 70% SMC and 30% BHIS ( 70∶30 media , 63 g BactoPeptone , 3 . 5 g Protease Peptone , 11 . 1 g BHI , 1 . 5 g yeast extract , 1 . 06 g Tris base , 0 . 7 g NH4SO4 , 15 g agar per liter ) [37] . 70∶30 agar ( supplemented as appropriate with thiamphenicol at 5–10 µg/mL ) was inoculated from a starter culture grown on solid media . HB101/pK424 strains were used for conjugations and BL21 ( DE3 ) strains were used for protein expression . E . coli strains were routinely grown at 37°C , shaking at 225 rpm in Luria-Bertani broth ( LB ) . Media was supplemented with chloramphenicol ( 20 µg/mL ) , ampicillin ( 50–100 µg/mL ) , or kanamycin ( 30 µg/mL ) as indicated . All strains are listed in Table S11; all plasmids are listed in Table S12; and all primers used are listed in Table S13 . For disruption of spo0A , sigE , sigG , sigK , and sigF , a modified plasmid containing the retargeting group II intron , pCE245 ( a gift from C . Ellermeier , University of Iowa ) , was used as the template . Primers used to amplify the targeting sequence from the template carried flanking regions specific for each gene target and are listed as follows: spo0A ( #539 , 540 , 541 and 532 , the EBS Universal primer as specified by the manufacturer ( Sigma Aldrich ) , sigE ( #653 , 654 , 655 and 532 ) , sigG ( #728 , 729 , 730 , and 532 ) , sigK ( #681 , 682 , 683 , and 532 ) and sigF ( #775 , 776 , 777 , and 532 ) . The spo0A disruption mutant was constructed using the same primers as Underwood et al . [36] . The resulting retargeting sequences were digested with BsrGI and HindIII and cloned into pJS107 ( a gift from J . Sorg , University of Texas A&M ) , a derivative of pJIR750ai ( Sigma Aldrich ) [32] . The ligations were transformed into DH5α and confirmed by sequencing . The resulting plasmids were used to transform HB101/pK424 . To construct the sigE complementation construct , primers #725 and 726 were used to amplify a fragment containing 252 bp upstream and 156 bp downstream of the two gene spoIIGA-sigE operon using 630 genomic DNA as the template . To construct the sigG complementation construct , primers #835 and 836 were used to amplify 288 bp upstream and 16 bp downstream of sigG using 630 genomic DNA as the template . The sigK complementation construct was made using PCR splicing by overlap extension ( SOE ) [75] . Primer pair #734 and 736 was used to amplify the 5′ SOE product , while primer pair #735 and 737 was used to amplify the 3′ SOE product . The resulting fragments were mixed together , and the flanking primers #734 and #737 were used to amplify an 898 bp fragment corresponding to the sigK gene including 256 bp region of upstream sequence . This strategy was used to clone an intact sigK gene with the skin element excised [43] . To construct the sigF complementation construct , primers #954 and #956 were used to amplify 88 bp upstream and 19 bp downstream of spoIIAA-spoIIAB-sigF operon , using 630 genomic DNA as the template . All complementation constructs were digested with NotI and XhoI and ligated into pMTL83151 [48] digested with the same enzymes , with the exception of the sigF complementation construct , which was cloned into pMTL84151 digested with the same enzymes [48] . To construct strains producing recombinant CD3522 , σE , σG , σF , Gpr , SpoVT , and SspA for antibody production , primer pairs #498 and 499; #596 and 597; #727 and 688; #723 and 724; #790 and 791; #883 and 884; #975 and 976; and #885 and 886 were used to amplify the cd3522 , sigE , sigG , sigF , gpr , spoVT , and sspA genes lacking stop codons , respectively , using 630 genomic DNA as the template . The sigE expression construct deletes the sequence encoding the first 23 amino acids of σE , which removes its membrane-tethering domain and improves the solubility of the protein in E . coli . The resulting PCR products were digested with NdeI and XhoI , ( or NheI and XhoI for gpr ) ligated to pET22b ( or pET21a for gpr and sspA ) , and used to transform DH5α . To construct a strain producing recombinant σK , PCR SOE was used to amplify the sigK gene lacking the skin element . Primer pair #689 and 736 was used to amplify the 5′ SOE product , while primer pair #735 and 737 was used to amplify the 3′ SOE product . The resulting fragments were mixed together , and the flanking #689 and #737 primers were used to amplify the sigK gene including the TAA stop codon . The resulting PCR product was digested with NcoI and XhoI , ligated to pET30a digested with the same enzymes , and used to transform DH5α . The resulting pET22b-cd3522 , pET22b-sigE , pET22b-sigG , pET30a-sigK , pET22b-sigF , pET21a-gpr , pET22b-spoVT , and pET21a-sspA plasmids were used to transform BL21 ( DE3 ) for protein expression . C . difficile strains were constructed using TargeTron-based gene disruption as described previously ( Figure S1 , [32] , [37] , [76] ) . TargeTron constructs in pJS107 were conjugated into C . difficile using an E . coli HB101/pK424 donor strain . HB101/pK424 strains containing the appropriate pJS107 construct were grown aerobically to exponential phase in 2 mL of LB supplemented with ampicillin ( 50 µg/mL ) and chloramphenicol ( 10 µg/mL ) . Cultures were pelleted , transferred into the anaerobic chamber , and resuspended in 1 . 5 mL of late-exponential phase C . difficile JIR8094 cultures ( grown anaerobically in BHIS broth ) . The resulting cell mixture was plated as seven 100 µL spots onto pre-dried , pre-reduced BHIS agar plates . After overnight incubation , all growth was harvested from the BHIS plates , resuspended in 2 . 5 mL pre-reduced BHIS , and twenty-one 100 µL spots per strain were plated onto BHIS agar supplemented with thiamphenicol ( 10 µg/mL ) , kanamycin ( 50 µg/mL ) , and cefoxitin ( 16 µg/mL ) to select for C . difficile containing the pJS107 plasmid . After 24–48 hrs of anaerobic growth , single colonies were patched onto BHIS agar supplemented with thiamphenicol ( 10 µg/mL ) , kanamycin ( 50 µg/mL ) , and FeSO5 ( 50 µM ) to induce the ferredoxin promoter of the group II intron system . After overnight growth , patches were transferred to BHIS agar plates supplemented with erythromycin ( 10 µg/mL ) for 24–72 hrs to select for cells with activated group II intron systems . Erythromycin-resistant patches were struck out for isolation onto the same media and individual colonies were screened by colony PCR for a 2 kb increase in the size of spo0A ( primer pair #556 and 557 ) , sigE ( primer pair #687 and 688 ) , sigG ( primer pair #723 and 724 ) , sigK ( primer pair #689 and 690 ) , and sigF ( primer pair #790 and 791 ) ( Figure S1 ) . A minimum of two independent clones from each mutant strain was phenotypically characterized . HB101/pK424 donor strains carrying the appropriate complementation construct were grown in LB containing ampicillin ( 50 µg/mL ) and chloramphenicol ( 20 µg/mL ) at 37°C , 225 rpm , under aerobic conditions , for 6 hrs . C . difficile recipient strains spo0A− , sigE− , sigG− , sigK− , and sigF− , containing group II intron disruptions , were grown anaerobically in BHIS broth at 37°C with gentle shaking for 6 hrs . HB101/pK424 cultures were pelleted at 2500 rpm for 5 min and the supernatant was removed . Pellets were transferred to the anaerobic chamber and gently resuspended in 1 . 5 mL of the appropriate C . difficile culture . The resulting mixture was inoculated onto pre-dried , pre-reduced BHIS agar plates , as seven 100 µL spots for 12 hrs . All spots were collected anaerobically and resuspended in 1 mL PBS . The resulting suspension was spread onto pre-dried , pre-reduced BHIS agar plates supplemented with thiamphenicol ( 10 µg/mL ) , kanamycin ( 50 µg/mL ) , and cefoxitin ( 10 µg/mL ) at 100 µL per plate , five plates per conjugation . Plates were monitored for colony growth for 24–72 hrs . Individual colonies were struck out for isolation and analyzed for complementation by phase contrast microscopy , Western blot analysis and transmission electron microscopy . A minimum of two independent clones from each complementation strain was phenotypically characterized . For the sigF complementation , a pMTL84151 plasmid backbone was used . The complementation protocol was followed as described except that after spots were collected from overnight growth on BHIS plates , the resulting PBS suspension was spotted onto three BHIS agar plates supplemented with thiamphenicol ( 10 µg/mL ) , kanamycin ( 50 µg/mL ) , and cefoxitin ( 16 µg/mL ) with 7–100 µL spots per plate . C . difficile strains were grown from glycerol stocks on BHIS plates supplemented with TA ( 0 . 1% w/v ) , or with both TA and thiamphenicol ( 5–10 µg/mL ) for strains with pMTL83151-derived or pMTL84151-derived plasmids . Cultures grown on BHIS agar plates were then used to inoculate 70∶30 agar plates ( with thiamphenicol at 5–10 µg/mL as appropriate ) for 18–48 hrs as previously described [37] . Sporulation induced lawns were harvested in PBS , washed once , resuspended in 0 . 2 mL of PBS , visualized by phase contrast microscopy , and/or further processed for analysis by transmission electron microscopy or Western blotting . C . difficile strains grown from glycerol stocks on BHIS plates supplemented with taurocholate and thiamphenicol ( described above ) were inoculated on to 70∶30 media containing thiamphenicol ( 5–10 µg/mL ) . After 30 hrs of growth , cells were harvested in 1 . 0 mL PBS , and split into two tubes . One tube was heat shocked at 60–65°C for 25 minutes . Both heat-shocked and non-heat shocked cells were serially diluted , and cells were plated on pre-reduced BHIS-TA plates . After 20 hrs on BHIS-TA , colonies were counted , and cell counts were determined . The percent of heat-resistant spores was determined based on the ratio of heat-resistant cells to total cells , and sporulation efficiencies were determined based on the ratio of heat-resistant cells for a strain compared to wild type . Results are based on a minimum of three biological replicates . spo0A− containing empty vector was included as a control for all assays [77] . For fluorescence microscopy studies , C . difficile strains were harvested in PBS after 18 hours of growth on 70∶30 media , pelleted , and resuspended in 1 . 0 mL PBS containing 1 µg/mL FM4-64 ( Molecular Probes ) and 15 µg/mL Hoechst 33342 ( Molecular Probes ) . The bacterial suspension ( 4 µL ) was added to a freshly prepared 1% agarose pad on a microscope slide , covered with a 22×22 mm #1 coverslip and sealed with VALAB ( 1∶1∶1 of vaseline , lanolin , and beeswax ) as previously described [78] . Phase and fluorescence microscopy were performed using a Nikon PlanApo 100× Ph3 oil immersion objective ( 1 . 4 NA ) on a Nikon Eclipse TE300 epifluorescence microscope . Five fields for each sample were acquired with an iXon3 885 EMCCD camera ( Andor ) cooled to −70°C with frame averaging set to 4 and an EM gain setting of 3 , and driven by NIS-Elements software ( Nikon ) . Images were subsequently imported into Adobe Photoshop CS6 for minimal adjustments in brightness/contrast levels and pseudocoloring . Phase-contrast microscopy for imaging the samples used for RNA-Seq was performed as previously described [37] . Quantification of total cells undergoing sporulation was determined by analyzing multiple fields for each strain at random . Greater than 200 cells were enumerated for each strain . For cultures analyzed by fluorescence microscopy , sporulating cells were identified as either having a polar septum with or without DNA staining in the forespore , a phase-dark forespore with or without DNA staining in the forespore compartment , a phase-bright forespore without DNA staining , or a free spore ( no mother cell compartment ) . One hundred microliters of bacterial cell suspension samples from sporulation assays were prepared as previously described [37] . Sporulation assay C . difficile cells ( 50 µL of PBS suspension ) were freeze-thawed three times , diluted in 100 µL EBB buffer ( 8 M urea , 2 M thiourea , 4% ( w/v ) SDS , 2% ( v/v ) β-mercaptoethanol ) , and incubated at 95°C for 20 min with vortexing every 5 min . Samples were centrifuged for 5 min at 15 , 000 rpm and 7 µL of 4× sample buffer ( 40% ( v/v ) glycerol , 1 M Tris pH 6 . 8 , 20% ( v/v ) β-mercaptoethanol , 8% ( w/v ) SDS , and 0 . 04% ( w/v ) bromophenol blue ) , was added . Protein samples were incubated again at 95°C for 15 minutes with vortexing followed by centrifugation for 5 min at 15 , 000 rpm . SDS-PAGE gels ( 12%–15% ) were loaded with 5 µL of protein prep . Gels were transferred to Bio-Rad PVDF membrane and blocked in 50% PBS:50% Odyssey Blocking Buffer with 0 . 1% ( v/v ) Tween for 30 min at RT . Polyclonal rabbit anti-σE , anti-σG , anti-σF , anti-SpoIVA [37] , and anti-CD1433 [76] , anti-CD1067 , anti-Gpr , anti-SpoVT , and anti-SspA antibodies were used at a 1∶1 , 000 dilution and anti-σK , anti-CD1511 , anti-SleC [76] , and anti-CD3522 at a 1∶5 , 000 dilution . Monoclonal mouse anti-Spo0A [37] was used at a 1∶10 , 000 dilution . IRDye 680CW and 800CW infrared dye-conjugated secondary antibodies were used at a 1∶20 , 000 dilutions . The Odyssey LiCor CLx was used to detect secondary antibody fluorescent emissions for Western blots . The anti-Δ230aa-σE , anti-σG , anti-σK , anti-σF , anti-CD3522 , anti-Gpr , anti-SpoVT , and anti-SspA antibodies used in this study were raised in rabbits by Cocalico Biologicals ( Reamstown , PA ) . The antigens Δ230aa-σE-His6 , σG-His6 , His6−σK , σF-His6 , CD3522-His6 , Gpr-His6 , SpoVT-His6 , and SspA-His6 , were purified on Ni2+-affinity resin from E . coli strains #755 , 743 , 756 , 921 , 577 , 853 , 881 , and #SspA respectively , as described above . Cultures were grown and protein expression was induced with 250 µM IPTG overnight at 19°C . E . coli cells were harvested , pelleted , and resuspended in 25 mL of low imidazole buffer ( LIB; 500 mM NaCl , 50 mM Tris-HCl , pH 7 . 5 , 15 mM imidazole , 10% ( v/v ) glycerol ) . Cells were flash frozen in liquid nitrogen , thawed , and lysed by sonication ( 45 sec burst followed by 5 min on ice for 3 cycles ) . For protein affinity purification , the lysate was centrifuged at 16 , 000× g for 30 min , supernatant was collected and added to pre-washed Ni2+-affinity resin for 4 hrs at 4°C . Bound beads were centrifuged at 2 , 000× g for 2 min at 4°C and washed once in LIB . Beads were reconstituted in 375 µL of high imidazole buffer ( HIB; 500 mM NaCl , 50 mM Tris-HCl , pH 7 . 5 , 200 mM imidazole , 10% ( v/v ) glycerol ) , incubated on a nutator for 15 min at RT , centrifuged , and eluate was collected . Beads were reconstituted with HIB for a total of five sequential elutions . Polyclonal antibodies against CD1067 were raised in rabbits against a peptide derived from CD1067 ( INSEDMRGFKKSHHC , Genscript ) ; the polyclonal antibodies were affinity-purified using the indicated peptide ( Genscript ) . RNA for RNA-Seq was extracted from WT , spo0A− , sigE− , sigF− , sigG− , and sigK− C . difficile cell suspensions , from an 18 hr sporulation assay ( described earlier ) , using a FastRNA Pro Blue Kit ( MP Biomedical ) and a FastPrep-24 automated homogenizer ( MP Biomedical , setting 6 . 0 , 45 seconds for 3 cycles ) . Contaminating genomic DNA was depleted using a column-bound DNase treatment with an RNeasy Kit ( Qiagen ) followed by two suspension DNase treatments ( New England Biolabs ) , according to manufacturer's recommendations . Samples were tested for genomic DNA contamination using quantitative PCR for 16S rRNA and the sleC gene . DNAse-treated RNA ( 5 µg ) was mRNA enriched using a Ribo-Zero Magnetic Kit ( Epicentre ) . RNA isolated for qRT-PCR was processed identically except that mRNA enrichment was done using an Ambion MICROBExpress Bacterial mRNA Enrichment Kit ( Invitrogen ) . Reverse transcription of enriched RNA was done using the Super Script First Strand cDNA Synthesis Kit ( Invitrogen ) with random hexamer primers . Enriched mRNA ( 100 ng ) was submitted to the Advanced Technology Genome Center Core Lab at the University of Vermont for massively parallel sequencing on an Illumina HiSeq 1000 . cDNA synthesis was carried out using the Ovation Prokaryotic RNA-Seq System ( Nugen ) , according to manufacturer's instructions . Libraries were prepared using the Ovation ultralow multiplex kit ( Nugen , 0304/0305-32 ) according to manufacturer's instructions . Briefly , samples were end-repaired , mono-adenylated , ligated to index/adaptors , and then amplified for 15 cycles ( after a PCR titration was performed ) . Completed libraries were quantitated using a SYBR Fast Universal qPCR Kit ( KAPA Biosystems ) . Paired end sequencing of samples was performed using a total of 10 pM of library in each flow cell lane . The samples were indexed and pooled in equal amounts to generate equal read coverage . Sequence calls and quality scores were produced in BCL format from images using Illumina RTA v1 . 13 with default parameters . Read pairs were mapped to libraries ( demultiplexed ) and converted to Fastq format using Illumina CASAVA 1 . 8 . 2 with default parameters . Adapters were clipped and reads were trimmed to remove the first 12 and last 11 cycles using Trimmomatic [79] , dropping read pairs for which at least one read was less than 50 bp . The C . difficile 630 genome ( NC_009089 ) sequence was modified by removing the sigK intervening ( skin ) element . The C . difficile 630 genome annotation was modified by the addition of sigK . Read pairs were aligned to the modified C . difficile 630 genome ( NC_009089 ) using BWA 0 . 6 . 1 with default parameters with one exception ( −q 20 ) . Read pairs were mapped to NC_009089 gene annotation using the countOverlaps procedure of the R/Bioconductor IRanges package [80]–[82] . Counts associated with rRNA were removed . Counts associated with the same library were pooled . Reads were of high quality ( median Phred score of 39 and a first quartile of 35 ) as were alignments ( median mapping quality score , MAPQ of 60 ) . Median fragment lengths were between 180 and 250 . The vast majority of unmapped sequences failed to align to sequences in the NCBI non-redundant database using a blastn and blastx search . There was no indication of highly represented reads among unmapped sequences . Since the majority of reads failed to map to known natural sequences , and since sequences can arise during library preparation particularly when the input sample is small , sequences that failed to map to the C . difficile genome likely represent spurious sequences produced during library construction . Differential expression statistics reflecting both effect size ( fold-change ) and statistical significance ( p-value adjusted based on the method of Benjamini and Hochberg [83] ) were calculated using DESeq [49] . Duplicate reads were excluded from these analyses . Differentially expressed genes were identified based on a minimum fold-change ( higher in the reference sample than the query ) and maximum p-value . Tables showing genes whose expression was downregulated by ≥4-fold with an adjusted p-value of ≤0 . 05 during sporulation are provided in the Supplementary Information ( Tables S3 , S4 , S5 , S6 , S7 , S8 , S9 , S10 ) . A table showing genes whose expression was upregulated by ≥4-fold with an adjusted p-value of ≤0 . 05 in a Spo0A-dependent manner are shown in Table S14 . The log2-transformed expression of genes that were downregulated by ≥4-fold with an adjusted p-value of ≤10−5 in the spo0A− strain relative to wild type expression were represented in a heat map using the heatmap . 2 procedure of the R/Bioconductor gplots package with default options [84] . Expression levels in spo0A− were not shown because the differential expression between spo0A− and wild type was biased by the method used to select genes . Expression levels in the other four strains relative to spo0A− were centered , scaled , and mapped to a red-green color scale . For the RNA-Seq validation , expression levels of gpr , cd0125 ( spoIIQ ) , cd2376 , cd1511 , cd3522 , spoIVA , cd1433 , cd1067 , sleC , sspB , spoVT , dacF , and rpoB ( housekeeping gene ) were run on WT , spo0A− , sigF− sigE− , sigG− , and sigK− cDNA templates in three replicate reactions using gene-specific primer pairs #1187 and 1188; #1213 and 1214; #1191 and 1192; #796 and 797; #989 and 990; #798 and 799; #792 and 793; #1030 and 1031; #575 and 576; #810 and 811; #995 and 996; #993 and 994; #1002 and 1003 , respectively . Quantitative real-time PCR was performed using SYBR Green JumpStart Taq Ready Mix ( Sigma ) , 50 nM of gene specific primers ( Table S12 ) , and an ABI PRISM 7900HT Sequence Detection System ( Applied Biosystems ) . Mean CT values were normalized to the spo0A− ( negative control ) sample and further normalized to rpoB . Relative expression values reported are representative of three biological replicates .
|
C . difficile is the leading cause of healthcare-associated infectious diarrhea in the United States in large part because of its ability to form spores . Since spores are resistant to most disinfectants and antibiotics , C . difficile infections frequently recur and are easily spread . Despite the importance of spores to C . difficile transmission , little is known about how spores are made . We set out to address this question by generating C . difficile mutants lacking regulatory factors required for sporulation and identifying genes that are regulated by these factors during spore formation using whole-genome RNA-Sequencing . We determined that the regulatory pathway controlling sporulation in C . difficile differs from related Clostridium species and the non-pathogenic model spore-former Bacillus subtilis and identified 314 genes that are induced during C . difficile spore development . Collectively , our study provides a framework for identifying C . difficile gene products that are essential for spore formation . Further characterization of these gene products may lead to the identification of diagnostic biomarkers and the development of new therapeutics .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"developmental",
"biology",
"infectious",
"diseases",
"genetics",
"biology",
"microbiology"
] |
2013
|
Global Analysis of the Sporulation Pathway of Clostridium difficile
|
The world's oceans contain a complex mixture of micro-organisms that are for the most part , uncharacterized both genetically and biochemically . We report here a metagenomic study of the marine planktonic microbiota in which surface ( mostly marine ) water samples were analyzed as part of the Sorcerer II Global Ocean Sampling expedition . These samples , collected across a several-thousand km transect from the North Atlantic through the Panama Canal and ending in the South Pacific yielded an extensive dataset consisting of 7 . 7 million sequencing reads ( 6 . 3 billion bp ) . Though a few major microbial clades dominate the planktonic marine niche , the dataset contains great diversity with 85% of the assembled sequence and 57% of the unassembled data being unique at a 98% sequence identity cutoff . Using the metadata associated with each sample and sequencing library , we developed new comparative genomic and assembly methods . One comparative genomic method , termed “fragment recruitment , ” addressed questions of genome structure , evolution , and taxonomic or phylogenetic diversity , as well as the biochemical diversity of genes and gene families . A second method , termed “extreme assembly , ” made possible the assembly and reconstruction of large segments of abundant but clearly nonclonal organisms . Within all abundant populations analyzed , we found extensive intra-ribotype diversity in several forms: ( 1 ) extensive sequence variation within orthologous regions throughout a given genome; despite coverage of individual ribotypes approaching 500-fold , most individual sequencing reads are unique; ( 2 ) numerous changes in gene content some with direct adaptive implications; and ( 3 ) hypervariable genomic islands that are too variable to assemble . The intra-ribotype diversity is organized into genetically isolated populations that have overlapping but independent distributions , implying distinct environmental preference . We present novel methods for measuring the genomic similarity between metagenomic samples and show how they may be grouped into several community types . Specific functional adaptations can be identified both within individual ribotypes and across the entire community , including proteorhodopsin spectral tuning and the presence or absence of the phosphate-binding gene PstS .
The concept of microbial diversity is not well defined . It can either refer to the genetic ( taxonomic or phylogenetic ) diversity as commonly measured by molecular genetics methods , or to the biochemical ( physiological ) diversity measured in the laboratory with pure or mixed cultures . However , we know surprisingly little about either the genetic or biochemical diversity of the microbial world [1] , in part because so few microbes have been grown under laboratory conditions [2 , 3] , and also because it is likely that there are immense numbers of low abundance ribotypes that have not been detected using molecular methods [4] . Our understanding of microbial physiological and biochemical diversity has come from studying the less than 1% of organisms that can be maintained in enrichments or cultivated , while our understanding of phylogenetic diversity has come from the application of molecular techniques that are limited in terms of identifying low-abundance members of the communities . Historically , there was little distinction between genetic and biochemical diversity because our understanding of genetic diversity was based on the study of cultivated microbes . Biochemical diversity , along with a few morphological features , was used to establish genetic diversity via an approach called numerical taxonomy [5 , 6] . In recent years the situation has dramatically changed . The determination of genetic diversity has relied almost entirely on the use of gene amplification via PCR to conduct taxonomic environmental gene surveys . This approach requires the presence of slowly evolving , highly conserved genes that are found in otherwise very diverse organisms . For example , the gene encoding the small ribosomal subunit RNA , known as 16S , based on sedimentation coefficient , is most often used for distinguishing bacterial and archaeal species [7–10] . The 16S rRNA sequences are highly conserved and can be used as a phylogenetic marker to classify organisms and place them in evolutionary context . Organisms whose 16S sequences are at least 97% identical are commonly considered to be the same ribotype [11] , otherwise referred to as species , operational taxonomic units , or phylotypes . Although rRNA-based analysis has revolutionized our view of genetic diversity , and has allowed the analysis of a large part of the uncultivated majority , it has been less useful in predicting biochemical diversity . Furthermore , the relationship between genetic and biochemical diversity , even for cultivated microbes , is not always predictable or clear . For instance , organisms that have very similar ribotypes ( 97% or greater homology ) may have vast differences in physiology , biochemistry , and genome content . For example , the gene complement of Escherichia coli O157:H7 was found to be substantially different from the K12 strain of the same species [12] . In this paper , we report the results of the first phase of the Sorcerer II Global Ocean Sampling ( GOS ) expedition , a metagenomic study designed to address questions related to genetic and biochemical microbial diversity . This survey was inspired by the British Challenger expedition that took place from 1872–1876 , in which the diversity of macroscopic marine life was documented from dredged bottom samples approximately every 200 miles on a circumnavigation [13–15] . Through the substantial dataset described here , we identified 60 highly abundant ribotypes associated with the open ocean and aquatic samples . Despite this relative lack of diversity in ribotype content , we confirm and expand upon previous observations that there is tremendous within-ribotype diversity in marine microbial populations [4 , 7 , 8 , 16 , 17] . New techniques and tools were developed to make use of the sampling and sequencing metadata . These tools include: ( 1 ) the fragment recruitment tool for performing and visualizing comparative genomic analyses when a reference sequence is available; ( 2 ) new assembly techniques that use metadata to produce assemblies for uncultivated abundant microbial taxa; and ( 3 ) a whole metagenome comparison tool to compare entire samples at arbitrary degrees of genetic divergence . Although there is tremendous diversity within cultivated and uncultivated microbes alike , this diversity is organized into phylogenetically distinct groups we refer to as subtypes . Subtypes can occupy similar environments yet remain genetically isolated from each other , suggesting that they are adapted for different environmental conditions or roles within the community . The variation between and within subtypes consists primarily of nucleotide polymorphisms but includes numerous small insertions , deletions , and hypervariable segments . Examination of the GOS data in these terms sheds light on patterns of evolution and also suggests approaches towards improving the assembly of complex metagenomic datasets . At least some of this variation can be associated with functional characters that are a direct response to the environment . More than 6 . 1 million proteins , including thousands of new protein families , have been annotated from this dataset ( described in the accompanying paper [18] ) . In combination , these papers bring us closer to reconciling the genetic and biochemical disconnect and to understanding the marine microbial community . We describe a metagenomic dataset generated from the Sorcerer II expedition . The GOS dataset , which includes and extends our previously published Sargasso Sea dataset [19] , now encompasses a total of 41 aquatic , largely marine locations , constituting the largest metagenomic dataset yet produced with a total of ~7 . 7 million sequencing reads . In the pilot Sargasso Sea study , 200 l surface seawater was filtered to isolate microorganisms for metagenomic analysis . DNA was isolated from the collected organisms , and genome shotgun sequencing methods were used to identify more than 1 . 2 million new genes , providing evidence for substantial microbial taxonomic diversity [19] . Several hundred new and diverse examples of the proteorhodopsin family of light-harvesting genes were identified , documenting their extensive abundance and pointing to a possible important role in energy metabolism under low-nutrient conditions . However , substantial sequence diversity resulted in only limited genome assembly . These results generated many additional questions: would the same organisms exist everywhere in the ocean , leading to improved assembly as sequence coverage increased; what was the global extent of gene and gene family diversity , and can we begin to exhaust it with a large but achievable amount of sequencing; how do regions of the ocean differ from one another; and how are different environmental pressures reflected in organisms and communities ? In this paper we attempt to address these issues .
Microbial samples were collected as part of the Sorcerer II expedition between August 8 , 2003 , and May 22 , 2004 , by the S/V Sorcerer II , a 32-m sailing sloop modified for marine research . Most specimens were collected from surface water marine environments at approximately 320-km ( 200-mile ) intervals . In all , 44 samples were obtained from 41 sites ( Figure 1 ) , covering a wide range of distinct surface marine environments as well as a few nonmarine aquatic samples for contrast ( Table 1 ) . Several size fractions were isolated for every site ( see Materials and Methods ) . Total DNA was extracted from one or more fractions , mostly from the 0 . 1–0 . 8-μm size range . This fraction is dominated by bacteria , whose compact genomes are particularly suitable for shotgun sequencing . Random-insert clone libraries were constructed . Depending on the uniqueness of each sampling site and initial estimates of the genetic diversity , between 44 , 000 and 420 , 000 clones per sample were end-sequenced to generate mated sequencing reads . In all , the combined dataset includes 6 . 25 Gbp of sequence data from 41 different locations . Many of the clone libraries were constructed with a small insert size ( <2 kbp ) to maximize cloning efficiency . As this often resulted in mated sequencing reads that overlapped one another , overlapping mated reads were combined , yielding a total of ~6 . 4 M contiguous sequences , totaling ~5 . 9 Gbp of nonredundant sequence . Taken together , this is the largest collection of metagenomic sequences to date , providing more than a 5-fold increase over the dataset produced from the Sargasso Sea pilot study [19] and more than a 90-fold increase over the other large marine metagenomic dataset [20] . Assembling genomic data into larger contigs and scaffolds , especially metagenomic data , can be extremely valuable , as it places individual sequencing reads into a greater genomic context . A largely contiguous sequence links genes into operons , but also permits the investigation of larger biochemical and/or physiological pathways , and also connects otherwise-anonymous sequences with highly studied “taxonomic markers” such as 16S or recA , thus clearly identifying the taxonomic group with which they are associated . The primary assembly of the combined GOS dataset was performed using the Celera Assembler [21] with modifications as previously described [19] and as given in Materials and Methods . The assembly was performed with quite stringent criteria , beginning with an overlap cutoff of 98% identity to reduce the potential for artifacts ( e . g . , chimeric assemblies or consensus sequences diverging substantially from the genome of any given cell ) . This assembly was the substrate for annotation ( see the accompanying paper by Yooseph et al . [18] ) . The degree of assembly of a metagenomic sample provides an indication of the diversity of the sample . A few substantial assemblies notwithstanding , the primary assembly was strikingly fragmented ( Table 2 ) . Only 9% of sequencing reads went into scaffolds longer than 10 kbp . A majority ( 53% ) of the sequencing reads remained unassembled singletons . Scaffolds containing more than 50 kb of consensus sequence totaled 20 . 7 Mbp; of these , >75% were produced from a single Sargasso Sea sample and correspond to the Burkholderia or Shewanella assemblies described previously [19] . These results highlight the unusual abundance of these two organisms in a single sample , which significantly affected our expectations regarding the current dataset . Given the large size of the combined dataset and the substantial amount of sequencing performed on individual filters , the overall lack of assembly provides evidence of a high degree of diversity in surface planktonic communities . To put this in context , suppose there were a clonal organism that made up 1% of our data , or ~60 Mbp . Even a genome of 10 Mbp—enormous by bacterial standards—would be covered ~6-fold . Such data might theoretically assemble with an average contig approaching 50 kb [22] . While real assemblies generally fall short of theory for various reasons , Shewanella data make up <1% of the total GOS dataset , and yet most of the relevant reads assemble into scaffolds >50 kb . Thus , with few scaffolds of significant length , we could conclude that there are very few clonal organisms present at even 1% in the GOS dataset . To investigate the nature of the implied diversity and to see whether greater assembly could be achieved , we explored several alternative approaches . Breaks in the primary assembly resulted from two factors: incomplete sequence coverage and conflicts in the data . Conflicts can break assemblies when there is no consistent way to chain together all overlapping sequencing reads . As it was possible that there would be fewer conflicts within a single sample ( i . e . , that diversity within a single sample would be lower ) , assemblies were attempted with individual samples . However , the results did not show any systematic improvements even in those samples with greater coverage ( unpublished data ) . Upon manual inspection , most assembly-breaking conflicts were found to be local in nature . These observations suggested that reducing the degree of sequence identity required for assembly could ameliorate both factors limiting assembly: effective coverage would increase and many minor conflicts would be resolved . Accordingly , we produced a series of assemblies based on 98% , 94% , 90% , 85% , and 80% identity overlaps for two subsets of the GOS dataset , again using the Celera Assembler . Assembly lengths increased as the overlap cutoff decreased from 98% to 94% to 90% , and then leveled off or even dropped as stringency was reduced below 90% ( Table 3 ) . Although larger assemblies could be generated using lower identity overlaps , significant numbers of overlaps satisfying the chosen percent identity cutoff still went unused in each assembly . This is consistent with a high rate of conflicting overlaps and in turn diagnostic of significant polymorphism . In mammalian sequencing projects the use of larger insert libraries is critical to producing larger assemblies because of their ability to span repeats or local polymorphic regions [23] . The shotgun sequencing libraries from the GOS filters were typically constructed from inserts shorter than 2 kb . Longer plasmid libraries were attempted but were much less stable . We obtained paired-end sequences from 21 , 419 fosmid clones ( average insert size , 36 kb; [24 , 25] ) from the 0 . 1-micron fraction of GS-33 . The effect of these long mate pairs on the GS-33 assembly was quite dramatic , particularly at high stringency ( e . g . , improving the largest scaffold from 70 kb to 1 , 247 kb and the largest contig from 70 kb to 427 kb ) . At least for GS-33 this suggests that many of the polymorphisms affect small , localized regions of the genome that can be spanned using larger inserts . This degree of improvement may be greater than what could be expected in general , as the diversity of GS-33 is by far the lowest of any of the currently sequenced GOS samples , yet it clearly indicates the utility of including larger insert libraries for assembly . In the absence of substantial assembly , direct comparison of the GOS sequencing data to the genomes of sequenced microbes is an alternative way of providing context , and also allows for exploration of genetic variation and diversity . A large and growing set of microbial genomes are available from the National Center for Biotechnology Information ( NCBI; http://www . ncbi . nlm . nih . gov ) . At the time of this study , we used 334 finished and 250 draft microbial genomes as references for comparison with the GOS sequencing reads . Comparisons were carried out in nucleotide-space using the sequence alignment tool BLAST [26] . BLAST parameters were designed to be extremely lenient so as to detect even distant similarities ( as low as 55% identity ) . A large proportion of the GOS reads , 70% in all , aligned to one or more genomes under these conditions . However , many of the alignments were of low identity and used only a portion of the entire read . Such low-quality hits may reflect distant evolutionary relationships , and therefore less information is gained based on the context of the alignment . More stringent criteria could be imposed requiring that the reads be aligned over nearly their entire length without any large gaps . Using this stringent criterion only about 30% of the reads aligned to any of the 584 reference genomes . We refer to these fully aligned reads as “recruited reads . ” Recruited reads are far more likely to be from microbes closely related to the reference sequence ( same species ) than are partial alignments . Despite the large number of microbial genomes currently available , including a large number of marine microbes , these results indicate that a substantial majority of GOS reads cannot be specifically related to available microbial genomes . The amount and distribution of reads recruited to any given genome provides an indication of the abundance of closely related organisms . Only genomes from the five bacterial genera Prochlorococcus , Synechococcus , Pelagibacter , Shewanella , and Burkholderia yielded substantial and uniform recruitment of GOS fragments over most of a reference genome ( Table 4 ) . These genera include multiple reference genomes , and we observed significant differences in recruitment patterns even between organisms belonging to the same species ( Figure 2A–2I ) . Three genera , Pelagibacter ( Figure 2A ) , Prochlorococcus ( Figure 2B–2F ) , and Synechococcus ( Figure 2G–2I ) , were found abundantly in a wide range of samples and together accounted for roughly 50% of all the recruited reads ( though only ~15% of all GOS sequencing reads ) . By contrast , although every genome tested recruited some GOS reads , most recruited only a small number , and these reads clustered at lower identity to locations corresponding to large highly conserved genes ( for typical examples see Figure 2E–2F ) . We refer to this pattern as nonspecific recruitment as it reflects taxonomically nonspecific signals , with the reads in question often recruiting to distantly related sets of genomes . Most microbial genomes , including many of the marine microbes ( e . g . , the ubiquitous genus Vibrio ) , demonstrated this nonspecific pattern of recruitment . The relationship between the similarity of an individual sequencing read to a given genome and the sample from which the read was isolated can provide insight into the structure , evolution , and geographic distribution of microbial populations . These relationships were assessed by constructing a “percent identity plot” [27] in which the alignment of a read to a reference sequence is shown as a bar whose horizontal position indicates location on the reference and whose vertical position indicates the percent identity of the alignment . We colored the plotted reads according to the samples to which they belonged , thus indirectly representing various forms of metadata ( geographic , environmental , and laboratory variables ) . We refer to these plots that incorporate metadata as fragment recruitment plots . Fragment recruitment plots of GOS sequences recruited to the entire genomes of Pelagibacter ubique HTCC1062 , Prochlorococcus marinus MIT9312 , and Synechococcus WH8102 are presented in Poster S1 . Characteristic patterns of recruitment emerged from each of these abundant marine microbes consisting of horizontal bands made up of large numbers of GOS reads . These bands seem constrained to a relatively narrow range of identities that tile continuously ( or at least uniformly , in the case when abundance/coverage is lower ) along ~90% of the reference sequence . The uninterrupted tiling indicates that environmental genomes are largely syntenic with the reference genomes . Multiple bands , distinguished by degree of similarity to the reference and by sample makeup , may arise on a single reference ( Poster S1D and S1F ) . Each of these bands appears to represent a distinct , closely related population we refer to as a subtype . In some cases , an abundant subtype is highly similar to the reference genome , as is the case for P . marinus MIT9312 ( Poster S1 ) and Synechococcus RS9917 ( unpublished data ) . P . ubique HTCC1062 and other Synechococcus strains like WH8102 show more complicated banding patterns ( Poster S1D and S1F ) because of the presence of multiple subtypes that produce complex often overlapping bands in the plots . Though the recruitment patterns can be quite complex they are also remarkably consistent over much of the reference genome . In these more complicated recruitment plots , such as the one for P . ubique HTCC1062 , individual bands can show sudden shifts in identity or disappear altogether , producing a gap in recruitment that appears to be specific to that band ( see P . ubique recruitment plots on Poster S1B and S1E , and specifically between 130–140 kb ) . Finally , phylogenetic analysis indicates that separate bands are indeed evolutionarily distinct at randomly selected locations along the genome . The amount of sequence variation within a given band cannot be reliably determined from the fragment recruitment plots themselves . To examine this variation , we produced multiple sequence alignments and phylogenies of reads that recruited to several randomly chosen intervals along given reference genomes to show that there can be considerable within-subtype variation ( Figure 3A–3B ) . For example , within the primary band found in recruitment plots to P . marinus MIT9312 , individual pairs of overlapping reads typically differ on average between 3%–5% at the nucleotide level ( depending on exact location in the genome ) . Very few reads that recruited to MIT9312 have perfect ( mismatch-free ) overlaps with any other read or to MIT9312 , despite ~100-fold coverage . While many of these differences are silent ( i . e . , do not change amino acid sequences ) , there is still considerable variation at the protein level ( unpublished data ) . The amount of variation within subtypes is so great that it is likely that no two sequenced cells contained identical genomes . Variation in genome structure in the form of rearrangements , duplications , insertions , or deletions of stretches of DNA can also be explored via fragment recruitment . The use of mated sequencing reads ( pairs of reads from opposite ends of a clone insert ) provides a powerful tool for assessing structural differences between the reference and the environmental sequences . The cloning and sequencing process determines the orientation and approximate distance between two mated sequencing reads . Genomic structural variation can be inferred when these are at odds with the way in which the reads are recruited to a reference sequence . Relative location and orientation of mated sequences provide a form of metadata that can be used to color-code a fragment recruitment plot ( Figure 4 ) . This makes it possible to visually identify and classify structural differences and similarities between the reference and the environmental sequences ( Figure 5 ) . For the abundant marine microbes , a high proportion of mated reads in the “good” category ( i . e . , in the proper orientation and at the correct distance ) show that synteny is conserved for a large portion of the microbial population . The strongest signals of structural differences typically reflect a variant specific to the reference genome and not found in the environmental data . In conjunction with the requirement that reads be recruited over their entire length without interruption , recruitment plots result in pronounced recruitment gaps at locations where there is a break in synteny . Other rearrangements can be partially present or penetrant in the environmental data and thus may not generate obvious recruitment gaps . However , given sufficient coverage , breaks in synteny should be clearly identifiable using the recruitment metadata based on the presence of “missing” mates ( i . e . , the mated sequencing read that was recruited but whose mate failed to recruit; Figure 4 ) . The ratio of missing mates to “good” mates determines how penetrant the rearrangement is in the environmental population . In theory , all genome structure variations that are large enough to prevent recruitment can be detected , and all such rearrangements will be associated with missing mates . Depending on the type of rearrangement present other recruitment metadata categories will be present near the rearrangements' endpoints . This makes it possible to distinguish among insertions , deletions , translocations , inversions , and inverted translocations directly from the recruitment plots . Examples of the patterns associated with different rearrangements are presented in Figure 5 . This provides a rapid and easy visual method for exploring structural variation between natural populations and sequenced representatives ( Poster S1A and S1B ) . Variation in genome structure potentially results in functional differences . Of particular interest are those differences between sequenced ( reference ) microbes and environmental populations . These differences can indicate how representative a cultivated microbe might be and shed light on the evolutionary forces driving change in microbial populations . Fragment recruitment in conjunction with the mate metadata helped us to identify both the consistent and the rare structural differences between the genomes of microbial populations in the GOS data and their closest sequenced relatives . Our analysis has thus far been confined to the three microbial genera that were widespread in the GOS dataset as represented by the finished genomes of P . marinus MIT9312 , P . ubique HTCC1062 , and to a lesser extent Synechococcus WH8102 . Each of these genomes is characterized by large and small segments where little or no fragment recruitment took place . We refer to these segments as “gaps . ” These gaps represent reference-specific differences that are not found in the environmental populations rather than a cloning bias that identifies genes or gene segments that are toxic or unclonable in E . coli . The presence of missing mates flanking these gaps indicates that the associated clones do exist , and therefore that cloning issues are not a viable explanation for the absence of recruited reads . Although the reference-specific differences are quite apparent due to the recruitment gaps they generate , there are also sporadic rearrangements associated with single clones , mostly resulting from small insertions or deletions . Careful examination of the unrecruited mates of the reads flanking the gaps allowed us to identify , characterize , and quantify specific differences between the reference genome and their environmental relatives . The results of this analysis for P . ubique and P . marinus have been summarized in Table 5 . With few exceptions , small gaps resulted from the insertion or deletion of only a few genes . Many of the genes associated with these small insertions and deletions have no annotated function . In some cases the insertions display a degree of variability such that different sets of genes are found at these locations within a portion of the population . In contrast , many of the larger gaps are extremely variable to the extent that every clone contains a completely unrelated or highly divergent sequence when compared to the reference or to other clones associated with that gap . These segments are hypervariable and change much more rapidly than would be expected given the variation in the rest of the genome . Sites containing a hypervariable segment nearly always contained some insert . We identified two exceptions both associated with P . ubique . The first is approximately located at the 166-kb position in the P . ubique HTCC1062 genome . Though no large gap is present , the mated reads indicate that under many circumstances a highly variable insert is often present . The second is a gap on HTCC1062 that appears between 50 and 90 kb . This gap appears to be less variable than other hypervariable segments and is occasionally absent based on the large numbers of flanking long mated reads ( Poster S1A ) . Interestingly , the long mated reads around this gap seem to be disproportionately from the Sargasso Sea samples , suggesting that this segment may be linked to geographic and/or environmental factors . Thus , hypervariable segments are highly variable even within the same sample , can on occasion be unoccupied , and the variation , or lack thereof , can be sample dependent . Hypervariable segments have been seen previously in a wide range of microbes , including P . marinus [28] , but their precise source and functional role , especially in an environmental context , remains a matter of ongoing research . For clues to these issues we examined the genes associated with the missing mates flanking these segments and the nucleotide composition of the gapped sequences in the reference genomes . In some rare cases the genes identified on reads that should have recruited within a hypervariable gap were highly similar to known viral genes . For example , a viral integrase was associated with the P . ubique HTCC1062 hypervariable gap between 516 and 561 kb . However , in the majority of cases the genes associated with these gaps were uncharacterized , either bearing no similarity to known genes or resembling genes of unknown function . If these genes were indeed acquired through horizontal transfer then we might expect that they would have obvious compositional biases . Oligonucleotide frequencies along the P . ubique HTCC1062 and Synechococcus WH8102 genomes are quite different in the large recruitment gaps in comparison to the well-represented portions of the genome ( Poster S1 ) . Surprisingly , this was less true for P . marinus MIT9312 , where the gaps have been linked to phage activity [28] . These results suggest that these hypervariable segments of the genome are widespread among marine microbial populations , and that they are the product of horizontal transfer events perhaps mediated by phage or transposable elements . These results are consistent with and expand upon the hypothesis put forward by Coleman et al . [28] suggesting that these segments are phage mediated , and conflicts with initial claims that the HTCC1062 genome was devoid of genes acquired by horizontal transfer [29] . Though insertions and deletions accounted for many of the obvious regions of structural variation , we also looked for rearrangements . The high levels of local synteny associated with P . ubique and P . marinus suggested that large-scale rearrangements were rare in these populations . To investigate this hypothesis we used the recruitment data to examine how frequently rearrangements besides insertions and deletions could be identified . We looked for rearrangements consisting of large ( greater than 50 kb ) inversions and translocations associated with P . marinus; however , we did not identify any such rearrangements that consistently distinguished environmental populations from sequenced cultivars . Rare inversions and translocations were identified in the dominant subtype associated with MIT9312 ( Table 6 ) . Based on the amount of sequence that contributed to the analysis , we estimate that one inversion or translocation will be observed for every 2 . 6 Mbp of sequence examined ( less than once per P . marinus genome ) . A further observation concerns the uniformity along a genome of the evolutionary history among and within subtypes . For instance , the similarity between GOS reads and P . marinus MIT9312 is typically 85%–95% , while the similarity between MIT9312 and P . marinus MED4 is generally ~10% lower . However , there are several instances where the divergence of MIT9312 and MED4 abruptly decreases to no more than that between the GOS sequences and MIT9312 ( Poster S1G ) . These results are consistent either with horizontal transfer ( recombination ) or with inhomogeneous selectional pressures . Similar patterns are present in the two high-identity subtypes seen on the P . ubique HTCC1062 genome ( Poster S1D ) . Other regions show local increases in similarity between MIT9312 and the dominant subtype that are not reflected in the MIT9312/MED4 divergence ( e . g . , near positions 50 kb , 288 kb , 730 kb , 850 kb , and 954 kb on MIT9312; also see Poster S1G ) . These latter regions might reflect either regions of homogenizing recombination or regions of higher levels of purifying selection . However , the lengths of the intervals ( several are 10 kb or more ) are longer than any single gene and correspond to genes that are not extremely conserved over greater taxonomic distances ( in contrast to the ribosomal RNA operon ) . Equally , if widespread horizontal transfer of an advantageous segment explains these intervals , the transfers occurred long enough ago for appreciable variation to accumulate ( unpublished data ) . The analyses described above have been confined to those organisms with representatives in culture and for which genomes were readily available . Producing assemblies for other abundant but uncultivated microbial genera would provide valuable physiological and biochemical information that could eventually lead to the cultivation of these organisms , help elucidate their role in the marine community , and allow similar analyses of their evolution and variation such as those performed on sequenced organisms . Previous assembly efforts and the fragment recruitments plots showed that there is considerable and in many cases conflicting variation among related organisms . Such variation is known to disrupt whole-genome assemblers . This led us to try an assembly approach that aggressively resolves conflicts . We call this approach “extreme assembly” ( see Materials and Methods ) . This approach currently does not make use of mate-pairing data and , therefore produces only contigs , not scaffolded sequences . Using this approach , contigs as large as 900 kb could be aligned almost in their entirety to the P . marinus MIT9312 and P . ubique HTCC1062 genomes ( Figure 2J–2L ) . Consistent patterns of fragment recruitment ( see below ) generally provided evidence of the correctness of contigs belonging to otherwise-unsequenced organisms . Accordingly , large contigs from these alternate assemblies were used to investigate genetic and geographic population structure , as described below . However , the more aggressive assemblies demonstrably suffered from higher rates of assembly artifacts , including chimerism and false consensus sequences ( Figure 6 ) . Thus , the more stringent primary assembly was employed for most assembly-based analyses , as manual curation was not practical . As just noted , many of the large contigs produced by the more aggressive assembly methods described above did not align to any great degree with known genomes . Some could be tentatively classified based on contained 16S sequences , but the potential for computationally generated chimerism within the rRNA operon is sufficiently high that inspection of the assembly or other means of confirming such classifications is essential . An alternative to an unguided assembly that facilitates the association of assemblies with known organisms is to start from seed fragments that can be identified as belonging to a particular taxonomic group . We employed fragments outside the ribosomal RNA operon that were mated to a 16S-containing read , limiting extension to the direction away from the 16S operon . This produced contigs of 100 kb or more for several of the ribotypes that were abundant in the GOS dataset . When evaluated via fragment recruitment ( Figure 2M–2O ) , these assemblies revealed patterns analogous to those seen for the sequenced genomes described above: multiple subtypes could be distinguished along the assembly , differing in similarity to the reference sequence and sample distribution , with occasional gaps . Hypervariable segments by definition were not represented in these assemblies , but they may help explain the termination of the extreme assemblies for P . marinus and SAR11 and provide a plausible explanation for termination of assemblies of the other deeply sampled populations as well . This directed approach to assembly can also be used to investigate variation within a group of related organisms ( e . g . , a 16S ribotype ) . We explored the potential to assemble distinct subtypes of SAR11 by repeatedly seeding extreme assembly with fragments mated to a SAR11-like 16S sequence . Figure 7 compares the first 20 kb from each of 24 independent assemblies . Eighteen of these segments could be aligned full-length to a portion of the HTCC1062 genome just upstream of 16S , while six appeared to reflect rearrangements relative to HTCC1062 . The rearranged segments were associated with more divergent 16S sequences ( 8%–14% diverged from the 16S of HTCC1062 ) , while those without rearrangements corresponded to less divergent 16S ( averaging less than 3% different from HTCC1062 ) . In each segment , many reads were recruited above 90% identity , but different samples dominated different assemblies . Phylogenetic trees support the inference of evolutionarily distinct subtypes with distinctive sample distributions ( Figure 8 ) . Environmental surveys provide a cultivation-independent means to examine the diversity and complexity of an environmental sample and serve as a basis to compare the populations between different samples . Typically , these surveys use PCR to amplify ubiquitous but slowly evolving genes such as the 16S rRNA or recA genes . These in turn can be used to distinguish microbial populations . Since PCR can introduce various biases , we identified 16S genes directly from the primary GOS assembly . In total , 4 , 125 distinct full-length or partial 16S were identified . Clustering of these sequences at 97% identity gave a total of 811 distinct ribotypes . Nearly half ( 48% ) of the GOS ribotypes and 88% of the GOS 16S sequences were assigned to ribotypes previously deposited in public databases . That is , more than half the ribotypes in the GOS dataset were found to be novel at what is typically considered the species level [30] . The overall taxonomic distribution of the GOS ribotypes sampled by shotgun sequencing is consistent with previously published PCR based studies of marine environments ( Table 7 ) [31] . A smaller amount ( 16% ) of GOS ribotypes and 3 . 4% of the GOS 16S sequences diverged by more than 10% from any publicly available 16S sequence , thus being novel to at least the family level . A census of microbial ribotypes allows us to identify the abundant microbial lineages and estimate their contribution to the GOS dataset . Of the 811 ribotypes , 60 contain more than 8-fold coverage of the 16S gene ( Table 8 ) ; jointly , these 60 ribotypes accounted for 73% of all the 16S sequence data . All but one of the 60 have been detected previously , yet only a few are represented by close relatives with complete or nearly complete genome sequencing projects ( see Fragment Recruitment for further details ) . Several other abundant 16S sequences belong to well-known environmental ribotypes that do not have cultivated representatives ( e . g . , SAR86 , Roseobacter NAC-1–2 , and branches of SAR11 other than those containing P . ubique ) . Interestingly , archaea are nearly absent from the list of dominant organisms in these near-surface samples . The distribution of these ribotypes reveals distinct microbial communities ( Figure 9 and Table 8 ) . Only a handful of the ribotypes appear to be ubiquitously abundant; these are dominated by relatives of SAR11 and SAR86 . Many of the ribotypes that are dominant in one or more samples appear to reside in one of three separable marine surface habitats . For example , several SAR11 , SAR86 , and alpha Proteobacteria , as well as an Acidimicrobidae group , are widespread in the surface waters , while a second niche delineated by tropical samples contains several different SAR86 , Synechococcus and Prochlorococcus ( both cyanobacterial groups ) , and a Rhodospirillaceae group . Other ribotypes related to Roseobacter RCA , SAR11 , and gamma Proteobacteria are abundant in the temperate samples but were not observed in the tropical or Sargasso samples . Not surprisingly , samples taken from nonmarine environments ( GS33 , GS20 , GS32 ) , estuaries ( GS11 , GS12 ) , and larger-sized fraction filters ( GS01a , GS01b , GS25 ) have distinguishing ribotypes . Furthermore , as the complete genomes of these dominant members are obtained , the capabilities responsible for their abundances may well lend insight into the community metabolism in various oceanic niches . The most common approach for comparing the microbial community composition across samples has been to examine the ribotypes present as indicated by 16S rRNA genes or by analyzing the less-conserved ITS located between the 16S and 23S gene sequences [7 , 8 , 16 , 17] . However , a clear observation emerging from the fragment recruitment views was that the reference ribotypes recruit multiple subtypes , and that these subtypes were distributed unequally among samples ( Figures 2 , 7 , 8; Poster S1D , S1F , and S1I ) . We developed a method to assess the genetic similarity between two samples that potentially makes use of all portions of a genome , not just the 16S rRNA region . This similarity measure is assembly independent; under certain circumstances , it is equivalent to an estimate of the fraction of sequence from one sample that could be considered to be in the other sample . Whole-metagenomic similarities were computed for all pairs of samples . Results are presented for comparisons at ≥98% and 90% identity . No universal cutoff consistently divides sequences into natural subsets , but the 98% identity cutoff provides a relatively high degree of resolution , while the 90% cutoff appears to be a reasonable heuristic for defining subtypes . For instance , a 90% cutoff treats most of the reads specifically recruited to P . marinus MIT9312 as similar ( those more similar to MED4 notably excepted ) , while reasonably separating clades of SAR11 ( Figures 7 and 8 ) . Reads with no qualifying overlap alignment to any other read in a pair of samples are uninformative for this analysis , as they correspond to lineages that were so lightly sequenced that their presence in one sample and absence in another may be a matter of chance . For the 90% cutoff , 38% of the sequence reads contributed to the analysis . The resulting similarities reveal clear and consistent groupings of samples , as well as the outlier status of certain samples ( Figures 10 and 11 ) . The broadest contrast was between samples that could be loosely labeled “tropical” ( including samples from the Sargasso Sea [GS00b , GS00c , GS00d] and samples that are temperate by the formal definition but under the influence of the Gulf Stream [GS14 , GS15] ) and “temperate . ” Further subgroups can be identified within each of these categories , as indicated in Figures 10 and 11 . In some cases , these groupings were composed of samples taken from different ocean basins during different legs of the expedition . A few pairs of samples with strikingly high similarity were observed , including GS17 and GS18 , GS23 and GS26 , GS27 and GS28 , and GS00b and GS00d . In each case , these pairs of samples were collected from consecutive or nearly consecutive samples . However , the same could be said of many other pairs of samples that do not show this same degree of similarity . Indeed , geographically and temporally separated samples taken in the Atlantic ( GS17 , GS18 ) and Pacific ( GS23 , GS26 ) during separate legs of the expedition are more similar to one another than were most pairs of consecutive samples . The samples with least similarity to any other sample were from unique habitats . Thus , similarity cannot be attributed to geographic separation alone . The groupings described above can be reconstructed from taxonomically distinct subsets of the data . Specifically , the major groups of samples visible in Figure 10 were reproduced when sample similarities were determined based only on fragments recruiting to P . ubique HTCC1062 ( unpublished data ) . Likewise , the same groupings were observed when the fragments recruiting to either HTCC1062 or P . marinus MIT9312 , or both , were excluded from the calculations ( unpublished data ) . Thus , the factors influencing sample similarities do not appear to rely solely on the most abundant organisms but rather are reflected in multiple microbial lineages . It is tempting to view the groups of similar samples as constituting community types . Sample similarities based on genomic sequences correlated significantly with differences in the environmental parameters ( Table 1 ) , particularly water temperature and salinity ( unpublished data ) . Samples that are very similar to each other had relatively small differences in temperature and salinity . However , not all samples that had similar temperature and salinity had high community similarities . Water depth , primary productivity , fresh water input , proximity to land , and filter size appeared consistent with the observed groupings . Other factors such as nutrients and light for phototrophs and fixed carbon/energy for chemotrophs may ultimately prove better predictors , but these results demonstrate the potential of using metagenomic data to tease out such relationships . Examining the groupings in Figure 11 in light of habitat and physical characteristics , the following may be observed . The first two samples , a hypersaline pond in the Galapagos Islands ( GS33 ) and the freshwater Lake Gatun in the Panama Canal ( GS20 ) are quite distinct from the rest . Salinity—both higher and lower than the remaining coastal and ocean samples—is the simplest explanation . Twelve samples form a strong temperate cluster as seen in the similarity matrix of Figure 11 as a darker square bounded by GS06 and GS12 . Embedded within the temperate cluster are three subclusters . The first subcluster includes five samples from Nova Scotia through the Gulf of Maine . This is followed by a subcluster of four samples between Rhode Island and North Carolina . The northern subcluster was sampled in August , the southern subcluster in November and December . Though all samples were collected in the top few meters , the southern samples were in shallower waters , 10 to 30 m deep , whereas most of the northern samples were in waters greater than 100 m deep . Monthly average estimates of chlorophyll a concentrations were typically higher in the southern samples as well ( Table 1 ) . All of these factors—temperature , system primary production , and depth of the sampled water body—likely contribute to the differences in microbial community composition that result in the two well-defined clusters . The final temperate subgroup includes two estuaries , Chesapeake Bay ( GS12 ) and Delaware Bay ( GS11 ) , distinguished by their lower salinity and higher productivity . However , GS11 is markedly similar not only to GS12 but also to coastal samples , whereas the latter appears much more unique . Interestingly , the Bay of Fundy estuary sample ( GS06 ) clearly did not group with the two other estuaries , but rather with the northern subgroup , perhaps reflecting differences in the rate or degree of mixing at the sampling site . Continuing to the right and downward in Figure 11 , one can see a large cluster of 25 samples from the tropics and Sargasso Sea , bounded by GS47 and GS00b . This can be further subdivided into several subclusters . The first subcluster ( a square bounded by GS47 and GS14 ) includes 14 samples , about half of which were from the Galapagos . The second distinct subcluster ( a square bounded by GS16 and GS26 ) includes seven samples from Key West , Florida , in the Atlantic Ocean to a sample close to the Galapagos Islands in the Pacific Ocean . Loosely associated with this subcluster is a sample from a larger filter size taken en route to the Galapagos ( GS25 ) . The remaining samples group weakly with the tropical cluster . GS32 was taken in a coastal mangrove in the Galapagos . The thick organic sediment at a depth of less than a meter is the likely cause for it being unlike the other samples . Sample 00a was from the Sargasso Sea and contained a large fraction of sequence reads from apparently clonal Burkholderia and Shewanella species that are atypical . When this sample is reanalyzed to exclude reads identified as belonging to these two groups , sample GS00a groups loosely with GS00b , GS00c , and GS00d ( unpublished data ) . Finally , three subsamples from a single Sargasso sample ( GS01a , GS01b , GS01c ) group together , despite representing three distinct size fractions ( 3 . 0–20 , 0 . 8–3 . 0 , and 0 . 1–0 . 8 μm , respectively; Table 1 ) . The complete set of sample similarities is more complex than described above , and indeed is more complex than can be captured by a hierarchical clustering . For instance , the southern temperate samples are appreciably more similar to the tropical cluster than are the northern temperate samples . GS22 appears to constitute a mix of tropical types , showing strong similarity not only to the GS47–GS14 subcluster to which it was assigned , but also to the other tropical samples . These results may be compared to the more traditional view of community structure afforded by 16S sequences ( Figure 9 ) . Some of the same groupings of samples are visible using both analyses . Several ribotypes recapitulated the temperate/tropical clustering described above . Others were restricted to the single instances of nonmarine habitats . Several of the most abundant organisms from the coastal mangrove , hypersaline lagoon , and freshwater lake were found exclusively in these respective samples . However , while several ribotypes recapitulated the temperate/tropical distinction revealed by the genomic sequence , others crosscut it . A few dominant 16S ribotypes , related to SAR11 , SAR86 , and SAR116 , were found in every marine sample . The brackish waters from two mid-Atlantic estuaries ( GS11 and GS12 ) contained a mixture of otherwise exclusively marine and freshwater ribotypes; similarity of these sites to the freshwater sample ( GS20 ) was minimal at the metagenomic level , while the greater similarity of GS11 to coastal samples visible at the metagenomic level was not readily visible here . A fuller comparison of metagenome-based measurements of diversity based on a large dataset of PCR-derived 16S sequences will be presented in another paper ( in preparation ) . Differences in gene content between samples can identify functions that reflect the lifestyles of the community in the context of its local environment [20 , 32] . We examined the relative abundance of genes belonging to specific functional categories in the distinct GOS samples . Genes were binned into functional categories using TIGRFAM hidden Markov models [18] , which are well annotated and manually curated [33] . The results can be filtered in various ways to highlight genes associated with specific environments . One catalog of possible interest is genes that were predominantly found in a single sample . We identified 95 TIGRFAMs that annotated large sets of genes ( 100 or more ) that were significantly more frequent ( greater than 2-fold ) in one sample than in any other sample ( Table 9 ) . Not surprisingly , this approach disproportionately singles out genes from the samples collected on larger filters ( GS01a , GS01b , and GS25 ) and from the nonmarine environments , particularly the hypersaline pond ( sample GS33 ) . Another contrast might be between the temperate and tropical clusters ( Figures 10 and 11 ) . We identified 32 proteins that were more than 2-fold more frequent in one or the other group ( Table 10 ) . The presence of various Prochlorococcus-associated genes in this list highlights some of the potential challenges with this sort of approach . Overrepresentation may reflect: a direct response to particular environmental pressures ( as the excess of salt transporters plausibly do in the hypersaline pond ) ; a lineage-restricted difference in functional repertoire ( as exemplified by the excess of photosynthesis genes in samples containing Prochlorococcus ) ; or a more incidental “hitchhiking” of a protein found in a single organism that happens to be present . We explored whether clearer and more informative differences could be discovered between communities by focusing on groups of samples that are highly similar in overall taxonomic/genetic content . Two pairs of samples provide a particularly nice illustration of this approach . Samples GS17 and GS18 from the western Caribbean Sea and samples GS23 and GS26 from the eastern Pacific Ocean were all very similar based on the presence of abundant ribotypes and overall similarity in genetic content ( Figures 9–11 ) . Despite these similarities , several genes are found to be up to seven times more common in the pair of Caribbean samples than the Pacific pair ( Table 11 ) . No genes are more than 2-fold higher in the Pacific than the Caribbean pair of samples . Several of the most differentially abundant genes are related to phosphate transport and utilization . It is very plausible that this is a reflection of a functional adaptation: these differences correlate well with measured differences in phosphate abundance between the Atlantic and eastern Pacific samples [34 , 35] , and phosphate abundance plays a critical role in microbial growth [36 , 37] . Indeed , the ability to acquire phosphate , especially under conditions where it is limited , is thought to determine the relative fitness of Prochlorococcus strains [38] . The single greatest difference between GS17 and GS18 on the one hand and GS23 and GS26 on the other was attributed to a set of genes annotated by the hidden Markov model TIGR02136 as a phosphate-binding protein ( PstS ) . This TIGRFAM identified a single gene in both P . marinus MIT9312 and P . ubique HTCC1062 . In P . marinus MIT9312 , this gene is located at 672 kb lying roughly in the middle of a 15-kb segment of the genome that recruits almost no GOS sequences from the Pacific sampling sites ( Poster S1H ) . In P . ubique HTCC1062 , the PstS gene is found at 1 , 133 kb in a 5-kb segment that also recruited far fewer GOS sequences from all the Pacific samples except for GS51 ( Poster S1E ) . These genomic segments differ structurally among isolates but they are no more variable than the flanking regions , and thus are not hypervariable in the sense used previously ( unpublished data ) . Nor are they particularly conserved when present , indicating that they are not the result of a recent lateral transfer . Phylogenetic analyses outside these segments did not produce any evidence of a Pacific versus Caribbean clade of either Prochlorococcus or SAR11 ( Figure 3A–3B ) . The presence or absence of phosphate transporters is not limited to these two types of organisms . The number of phosphate transporters that were found in the Caribbean far exceeds the number that can be attributed to HTCC1062- and MIT9312-like organisms . However , these results indicate that within individual strains or subtypes the ability to acquire phosphate ( in one or more of its forms ) can vary without detectable differences in the surrounding genomic sequences . Variation in gene content is only one aspect of the tremendous diversity in the GOS data . The functional significance of all the polymorphic differences between homologous proteins remains largely unknown . To look for functional differences , we analyzed members of proteorhodopsin gene family . Proteorhodopsins are fast , light-driven proton pumps for which considerable functional information is available though their biological role remains unknown . Proteorhodopsins were highly abundant in the Sargasso Sea samples [19] and continue to be highly abundant and evenly distributed ( relative to recA abundance ) in all the GOS samples . A total of 2 , 674 putative proteorhodopsin genes were identified in the GOS dataset . Although many of the sequences are fragmentary , 1 , 874 of these genes contain the residue that is primarily responsible for tuning the light-absorbing properties of the protein [39–41] , and these properties have been shown to be selected for under different environmental conditions [42] . Variation at this residue is strongly correlated with sample of origin ( Figure 12 ) . The leucine ( L ) or green-tuned variant was highly abundant in the North Atlantic samples and in the nonmarine environments like the fresh water sample from Lake Gatun ( GS20 ) . The glutamine ( Q ) or blue-tuned variant dominated in the remaining mostly open ocean samples . Given our limited understanding of the biological role for proteorhopsin , the reason for this differential distribution is not immediately clear . In coastal waters where nutrients are more abundant , phytoplankton is dominant . Phytoplankton absorbs primarily in the blue and red spectra; consequently , the water appears green [43] . Conversely , in the open ocean nutrients are rare and phytoplanktonic biomass is low , so waters appear blue because in the absence of impurities the red wavelengths are absorbed preferentially [44] . It may be that proteorhodopsin-carrying microbes have simply adapted to take advantage of the most abundant wavelengths of light in these systems . Proteorhodopsins encoded on reads that were recruited to P . ubique HTCC1062 account for a fraction ( ~25% ) of all the proteorhodopsin-associated reads , suggesting that the remainder must be associated with a variety of marine microbial taxa ( see also [45–47] ) . Phylogenetic analysis of the SAR11-associated proteins revealed that each variant has arisen independently at least two times in the SAR11 lineage ( Figure 3C ) . Consistent with other findings that proteorhodopsins are widely distributed throughout the microbial world [48] , we conclude that multiple microbial lineages are responsible for proteorhodopsin spectral variation and that the abundance of a given variant reflects selective pressures rather than taxonomic effects . Similar mechanisms seem to be involved in the evolution and diversification of opsins that mediate color vision in vertebrates [49] .
Our data demonstrate to an unprecedented degree the nature and evolution of genetic variation below the species level . Variation can be analyzed in several ways , including observed differences in sequence , genomic structure , and gene complement . The observed patterns of variation shed light on the mechanisms by which marine prokaryotes evolve . Gene synteny seems to be more highly conserved than the nucleotide and protein sequences . This variation is seen over essentially the entire genome in every abundant group of organisms sufficiently related for us to recognize a population by fragment recruitment . ( These include , but are not limited to , the organisms shown in Figure 2 and Poster S1 . ) Notably , we found no evidence of widespread low-diversity organisms such as B . anthracis [50] . Phylogenetic trees and fragment recruitment plots ( Figures 7 and 8 ) indicate that the variation within a species is not an unstructured swarm or cloud of variants all equally diverged from one another . Instead , there are clearly distinct subtypes , in terms of sequence similarity , gene content , and sample distribution . Similar findings have been shown for specific organisms , based on evaluation of one or a few loci [2 , 51–53] . These results rule out certain trivial models of population history and evolution for what is commonly considered a bacterial “species . ” For instance , it argues against a recent explosive population growth from a single successful individual ( selective sweep ) [54] . Equally , it argues against a perfectly mixed population , suggesting instead some barriers to competition and exchange of genetic material . In principle , this variation could reflect some combination of physical barriers ( true biogeography ) , short-term stochastic effects , and/or functional differentiation . Given the confounding variables of geography , time , and environmental conditions in the current collection of samples , it is difficult to definitively separate these effects , but various observations argue for functional differentiation between subtypes ( i . e . , they constitute distinct ecotypes ) . First , individual subtypes may be found in a wide range of locations; P . ubique HTCC1062 was isolated in the Pacific Ocean off the coast of Oregon [55] , but closely related sequences are relatively abundant in our samples taken in the Atlantic Ocean . Second , geography per se cannot fully explain differences in subtype distributions , as multiple subtypes are found simultaneously in a single sample . Third , the collection of samples in which a given subtype was found generally exhibits similar environmental conditions . A strong independent illustration of this comes from the correlation of temperature with the distribution of Prochlorococcus subtypes [56] . Fourth , the extensive variation within each subtype ( i . e . , the fact that subtypes are not clonal populations ) indicates that it cannot be chance alone that makes genetically similar organisms have similar observed distributions . Taken together , these results argue that subtype classification is more informative for categorizing microbial populations than classification using 16S-based ribotypes , or fingerprinting techniques based on length polymorphism , such as T-RFLPs [57] or ARISA [58] . For example , the grouping of such disparate microbial populations under the umbrella P . marinus dilutes the significance of the term “species . ” Indeed , numerous papers have been devoted to comparing and contrasting the differences and variability in P . marinus isolates to better understand how this particularly abundant group of organisms has evolved and adapted within the dynamic marine environment [28 , 52 , 56 , 59–66] . Prior to the widespread use of marker-based phylogenetic approaches , microbial systematics relied on a wide range of variables to distinguish microbial populations [67] . Subtypes bring us back to these more comprehensive approaches since they reflect the influences of a wide range of factors in the context of an entire genome . Although subtypes are a salient feature of our data , variation within a ribotype does not stop at the level of subtypes . Variation within subtypes is so extensive that few GOS reads can be aligned at 100% identity to any other GOS read , despite the deep coverage of several taxonomic groups . Related findings have been shown for the ITS region in various organisms [2 , 51 , 52] , and in a limited number of organisms for individual protein coding and intergenic regions [2 , 53 , 68] . High levels of diversity within the ribotype can be convincingly demonstrated in the 16S gene itself [69] . The applicability of these results over the entire genome were recently shown for P . marinus [28] using data from the Sargasso Sea samples taken as a pilot project for the expedition reported here [19] . We have definitively demonstrated the generality of these findings , greatly increased our understanding of the minimum number of variants of a given organism , and shown that these observations apply to the entire genome for a wide range of abundant taxonomic groups and across a wide range of geographic locations . Average pairwise differences of several percent between overlapping P . marinus or SAR11 reads imply that this variation did not arise recently . If one uses substitution rates estimated for E . coli [70] , one could conclude that on average any two P . marinus cells must have diverged millions of years ago . Mutational rates are notoriously variable and hard to estimate , and assumptions of molecular clocks are equally chancy , but clearly within-subtype variants have persisted side by side for quite some time . This raises a question related to the classic “Paradox of the Plankton”: how can so many similar organisms have coexisted for so long [71 , 72] ? One explanation , which we favor , is that not only subtypes but also individual variants are sufficiently different phenotypically to prevent any one strain from completely replacing all others ( discussed further below; see [71] for a recent theoretical treatment ) . An alternative is that recombination might prevent selective sweeps within ecotypes , as proposed by Cohan ( reviewed in [73] ) . Given the apparent generality of subtypes and intra-subtype variation , it is important to understand if and how these subpopulations are functionally distinct . At the level of DNA sequence , a substantial fraction of substitutions are silent in terms of amino acid sequence , and others may be nonsynonymous but functionally neutral . However , two organisms that differ by 5% in their genetic sequence ( e . g . , 100 , 000 substitutions in 2 Mbp of shared sequence ) will inevitably have at least minor functional differences such as in the optimal temperature or pH for the activity of some enzyme . At the level of gene content , the observation of hypervariable segments ( [28] and here ) implies that there is an additional dimension to functional variability . Hypervariable genomic islands with preferential insertion sites could potentially be associated with a wide range of functions , though to date they have been most closely examined for their role in pathogenicity ( for a review , see [74] ) . However , given their apparent variability within even a single sampling site , it seems unlikely that these elements reflect a specific adaptive advantage to the local population . Identifying the source ( s ) , diversity , and range of functionality associated with these islands by fully sequencing a large number of these segments and understanding how their individual abundances fluctuate should be quite informative . Some might still argue that these differences must be moot for the purpose of understanding the role these organisms play in an ecosystem . Yet even small differences in optimal conditions may have profound effects . They may prevent any single genotype from being universally fittest , allowing and/or necessitating the coexistence of multiple variants [2 , 51 , 69] . Moreover , variation within subtype might afford a form of functional “buffering , ” such that the population as a whole may be more stable in its ecosystem role than any one clone could be ( see also [51] ) . That is , while any one strain of Prochlorococcus might thrive and provide energy input to the rest of the community at a limited range of temperatures , light conditions , etc . , the ensemble might provide such inputs over a wider range of environmental conditions . In this way , microdiversity might provide system stability or robustness through functional redundancy and the “insurance effect” ( reviewed in [75] ) . Thus , while the extent of microdiversity suggests that knowing the behavior of any one isolate in exquisite detail might not be as useful to reductionist modeling as one might hope , this buffering could afford a more stable ensemble behavior , facilitating the development and maintenance of an ecosystem and allowing for system-level modeling . A direct equation of subtypes with ecotypes is tempting , but not entirely clear-cut . The correlation of PstS distribution with phosphate abundance suggests a functional adaptation , but within Prochlorococcus and SAR11 the presence or absence of PstS subdivides subtypes without apparent respect for phylogenetic structure . This contrasts markedly with the distribution of proteorhodopsin-tuning variants within SAR11 , which , despite a few convergent substitutions , are strongly congruent with phylogeny . It is interesting to ask what distinguishes pressures or adaptations that respect ( or that lead to ) lineage splits from those that show little or no phylogenetic structuring . These two specific examples plausibly reflect two different mechanisms ( i . e . , convergent but independent mutation in proteorhodopsin genes and the acquisition by horizontal transfer of genes involved in phosphate uptake ) . Yet , we must wonder: given the evidence that proteorhodopsin has been transferred laterally [48] , and that only a small number of mutations , in some circumstances even a single base-pair change , are required to switch between the blue-absorbing and green-absorbing forms [39 , 40] , why should proteorhodopsin variants show any lineage restriction ? Perhaps this relates to the modularity of the system in question: proteorhodopsin tuning may be part of a larger collection of synergistic adaptations that are collectively not easily evolved , acquired , or lost , while the PstS and surrounding genes may represent a functional unit that can be readily added and removed over relatively short evolutionary time scales . If so , perhaps subtypes are indeed ecotypes , but rapidly evolving characters can lead to phenotypes that crosscut or subdivide ecotypes . Phage provide one possible mechanism for rapid evolution of microbial populations or strains , and have been found in abundance with this and other marine metagenomic datasets [18 , 20] . It has been proposed that hypervariable islands are phage mediated [28] . However , there are reasons to be cautious about invoking phage as an explanation for rapidly evolving characteristics . While we see variability of PstS and neighboring genes in both SAR11 and P . marinus populations , this variation does not seem to be linked to recent phage activity . Initially , the distribution of PstS seems similar to the variation associated with the hypervariable islands , which may be phage mediated [28] . Indeed , phosphate-regulating genes including PstS have been identified in phage genomes [64] , presumably because enhanced phosphate acquisition is required during the replication portion of their life cycle . However , the regions containing the PstS genes in both SAR11 and Prochlorococcus do not behave in the same fashion as clearly hypervariable regions , being effectively bimorphic ( modulo the level of sequence variation observed elsewhere in the genome ) , whereas clearly hypervariable regions are so diverse that nearly every sampled clone falling in such a region appears completely unrelated to every other . Nor do the other genes in PstS-containing regions appear to be phage associated . These observations suggest that differences in PstS presence or absence arose in the distant past , or that different mechanisms are at work . It seems likely that phage may mediate lateral transfer of PstS and other phosphate acquisition genes , but it is unclear whether these genes then can become fixed within the population . Phage require enhanced phosphate acquisition as part of their life cycle [64] , so regulatory or functional differences in these genes may limit their suitability for being acquired by the host cell for its own purposes . The rate of phage-mediated horizontal transfer of genes may reflect a combination of the gene's value to the host and to the agent mediating the transfer ( e . g . , phage ) , suggesting that PstS may have much greater immediate value than do proteorhodopsin genes and their variants . In practical terms , these results highlight the limitations associated with marker-based analysis and the use of these approaches to infer the physiology of a particular microbial population . At the resolution used here , marker-based approaches are not always informative regarding differences in gene content ( e . g . , the PstS gene as well as neighboring genes ) , especially those associated with hypervariable segments . Though phosphate acquisition is known to vary within different strains of P . marinus [64 , 76] , our results clearly show that this variability can happen within a single subtype ( as represented by MIT9312 ) , effectively identifying distinct ecotypes . Given the correct samples from the appropriate environments , other core genes might also show similar variation and allow us to more fully assess the reliability of reference genomes as indicators of physiological potential . Analysis of the GOS dataset has benefited from the development of new tools and techniques . Many of these approaches rely on fairly well-known techniques but have been modified to take greater advantage of the metadata . The technique of fragment recruit and the corresponding fragment recruitment plots have proven highly useful for examining the biogeography and genomic variation of abundant marine microbes when a close reference genome exists . Ultimately , this approach derives from the percent identity plots of PipMaker [27] . Similar approaches have been used to examine variation with respect to metagenomic datasets . For example , hypervariable segments and sequence variation have been visualized in P . marinus MIT9312 using the Sargasso Sea data [28] and in human gut microbes Bifidobacterium longum and Methanobrevibacter smithii [77] . Our primary advance associated with fragment recruitment plots is the incorporation of metadata associated with the isolation or production of the sequencing data . While simple in nature , the resulting plots can be extremely informative due to the volume of data being presented . Being able to present the sequence similarity and metadata visually allows a researcher to quickly identify interesting portions of the data for further examination . This is one of the first tools to make extensive use of the metadata collected during a metagenomic sequencing project . The use of sample and recruitment metadata is just the beginning . It is not difficult to imagine displaying other variations such as water temperature , salinity , phosphate abundance , and time of year with this approach . Even sample independent metadata such as phylogenetic information may produce informative views of the data . The usefulness of this and related approaches will only grow as the robust collection of metadata becomes routine and the variables that are most relevant to microbial communities are further elucidated . The greatest limitation of fragment recruitment is the lack of appropriate reference sequences , particularly finished genomes . Using a series of modifications to the Celera Assembler referred to as “extreme assembly , ” we have produced large assemblies for cultivated and uncultivated marine microbes . On its own , the extreme assembly approach would be excessively prone to producing chimeric sequences . However , when extreme assemblies are used as references for fragment recruitment , the metadata provides additional criteria to validate the sampling consistency along the length of the scaffold . Chimeric joins can be rapidly detected and avoided . This argues that future metagenomic assemblers could be specifically designed to make use of the metadata to produce more accurate assemblies , and that metagenomic assemblies will be improved by using data from multiple sources . Finding ways to represent the full diversity in these assemblies remains a pressing issue . Extreme assembly can produce much larger assemblies but it is still limited by overall coverage . While many ribotypes are presumably present in sufficient quantities that reasonable assemblies of these genomes might be expected , this did not occur even for the most abundant organisms , including SAR11 and P . marinus . Many of the problems can be attributed to the diversity associated with the hypervariable segments where the effective coverage drops precipitously . If these are indeed commonplace in the microbial world , it is unlikely that complete genomes will be produced using the small insert libraries presented here . However , the ability to bin the larger sequences based on their coverage profiles across multiple samples , oligonucleotide frequency profiles , and phylogenetic markers suggests that large portions of a microbial genome can be reconstructed from the environmental data . This in turn should provide critical insights into the physiology and biochemistry of these microbial lineages that will inform culture techniques to allow cultivation of these recalcitrant organisms under laboratory conditions . Not every technique described herein relies on metadata . The marker-less , overlap-based metagenomic comparison provides a quantitative approach to comparing the overall genetic similarity of two samples ( Figures 10 and 11 ) . In essence , genomic similarity acts as a proxy for community similarity . Marker-based approaches such as ARISA including the use of 16S sequences described herein can also be used to infer community similarity , though these approaches more aptly generate a census of the community members [51 , 69 , 78 , 79] . This census is biased to the extent that 16S genes can vary in copy number and relies on linkage of the marker gene to infer genome composition . While our metagenome comparison does not directly provide a census , the sensitivity can be tuned by restricting the identity of matches . This means that even subtype-level differences can be detected across samples . It would also identify the substantial gene content differences between the K12 and O157:H7 E . coli strains [12] . Such large-scale gene content differences have yet to be seen between closely related marine microbes , but may be a factor in other environments . Although the requisite amount of data will vary with the complexity of the environment or the degree of resolution required , we have found that 10 , 000 sequencing reads is sufficient to reliably measure the similarity of two surface water samples ( unpublished data ) . This analysis may become a general tool for allocating sequencing resources by allowing a shallow survey of many samples followed by deep sequencing of a select number of “interesting” ones . The application of this technique for comparing samples along with detailed analysis of fragments recruiting to a given reference sequence can also help explicate differences among communities in gene content or sequence variation . For example , recent metagenomic studies have reported differences in abundance of various gene families or differing functional roles between samples . Some of these differences correspond to plausible differences in physiology and biochemistry , such as the relative overabundance of photosynthetic or light-responsive genes in surface water samples [20 , 32] . Other differences however are less obvious , such as the abundance of ribosomal proteins at 130 m or the abundance of tranposase at 4 , 000 m [20]; some of these may reflect “taxonomic hitchhiking , ” such that a sample rich in Archaea or Firmicutes or Cyanobacteria , etc . , has an overrepresentation of genes more reflective of their recent evolutionary history than of a response to environmental conditions . Being able to control or account for these taxonomic effects is crucial to understanding how microbial populations have adapted to environmental conditions and how they may behave under changing conditions . The metagenomic comparison method described here provides a new tool to more accurately measure the impact of taxonomic effects . In conclusion , this study reveals the wealth of biological information that is contained within large multi-sample environmental datasets . We have begun to quantify the amount and structure of the variation in natural microbial populations , while providing some information about how these factors are structured along phylogenetic and environmental factors . At the same time , many questions remain unanswered . For example , although microbial populations are structured and therefore genetically isolated , we do not understand the mechanisms that lead to this isolation . Their isolation seems contradictory given overwhelming evidence that horizontal gene transfer associated with hypervariable islands is a common phenomenon in marine microbial populations . Whatever the mechanism , the role and rate at which gene exchange occurs between populations will be crucial to understanding population structure within microbial communities and whether these communities are chance associations or necessary collections . The hypervariable islands could be a source for tremendous genetic innovation and novelty as evidenced by the rate of discovery of novel protein families in the GOS dataset [18] . However , it is not clear whether these entities are the main source of this novelty or whether this novelty resides in the vast numbers of rare microbes [4] that cannot be practically accessed using current metagenomic approaches . Altogether , this research reaffirms our growing wealth and complexity of data and paucity of understanding regarding the biological systems of the oceans .
A more detailed description of the sampling sites provides additional context in which to understand the individual samples . The northernmost site ( GS05 ) was at Compass Buoy in the highly eutrophic Bedford Basin , a marine embayment encircled by Halifax , Nova Scotia , that has a 15-y weekly record of biological , physical , and chemical monitoring ( http://www . mar . dfo-mpo . gc . ca/science/ocean/BedfordBasin/index . htm ) . Other temperate sites included a coastal station sample near Nova Scotia ( GS4 ) , a station in the Bay of Fundy estuary at outgoing tide ( GS06 ) , and three Gulf of Maine stations ( GS02 , GS03 , and GS07 ) . These were followed by sampling coastal stations from the New England shelf region of the Middle Atlantic Bight ( Newport Harbor through Delaware Bay; GS08–GS11 ) . The Delaware Bay ( GS11 ) was one of several estuary samples along the Global Expedition path . Estuaries are complex hydrodynamic environments that exhibit strong gradients in oxygen , nutrients , organic matter , and salinity and are heavily impacted by anthropogenic nutrients . The Chesapeake Bay ( GS12 ) is the largest estuary in the United States and has microbial assemblages that are diverse mixtures of freshwater and marine-specific organisms [80] . GS13 was collected near Cape Hatteras , North Carolina , inside and north of the Gulf Stream , and GS14 was taken along the western boundary frontal waters of the Gulf Stream off the coast of Charleston , South Carolina . The vessel stopped at five additional stations as it transited through the Caribbean Sea ( GS15–GS19 ) to the Panama Canal . In Panama , we sampled the freshwater Lake Gatun , which drains into the Panama Canal ( GS20 ) . The first of the eastern Pacific coastal stations GS21 , GS22 , and GS23 were sampled on the way to Cocos Island ( ~500 km southwest of Costa Rica ) , followed by a coastal Cocos Island sample ( GS25 ) . Near the island , ocean currents diverge and nutrient rich upwellings mix with warm surface waters to support a highly productive ecosystem . Cocos Island is distinctive in the eastern Pacific because it belongs to one of the first shallow undersea ridges in the region encountered by the easterly flowing North Equatorial Counter/Cross Current in the Far Eastern Pacific [81 , 82] . After departing Cocos Island , the vessel continued southwest to the Galapagos Islands , stopping for an open ocean station ( GS26 ) . An intensive sampling program was then conducted in the Galapagos . The Galapagos Archipelago straddles the equator 960 km west of mainland Ecuador in the eastern Pacific . These islands are in a hydrographically complex region due to their proximity to the Equatorial Front and other major oceanic currents and regional front systems [83] . The coastal and marine parts of the Galapagos Islands ecosystem harbor an array of distinctive habitats , processes , and endemic species . Several distinct zones were targeted including a shallow-water , warm seep ( GS30 ) , below the thermocline in an upwelling zone ( GS31 ) , a coastal mangrove ( GS32 ) , and a hypersaline lagoon ( GS33 ) . The last stations were collected from open ocean sites ( GS37 and GS47 ) and a coral reef atoll lagoon ( GS51 ) in the immense South Pacific Gyre . The open ocean samples come from a region of lower nutrient concentrations where picoplankton are thought to represent the single most abundant and important factor for biogeochemical structuring and nutrient cycling [84–87] . In the atoll systems , ambient nutrients are higher , and bacteria are thought to constitute a large biomass that is one to three times as large as that of the phytoplankton [88–90] . A YSI ( model 6600 ) multiparameter instrument ( http://www . ysi . com ) was deployed to determine physical characteristics of the water column , including salinity , temperature , pH , dissolved oxygen , and depth . Using sterilized equipment [91] , 40–200 l of seawater , depending on the turbidity of the water , was pumped through a 20-μm nytex prefilter into a 250-l carboy . From this sample , two 20-ml subsamples were collected in acid-washed polyethylene bottles and frozen ( −20 °C ) for nutrient and particle analysis . At each station the biological material was size fractionated into individual “samples” by serial filtration through 20-μm , 3-μm , 0 . 8-μm , and 0 . 1-μm filters that were then sealed and stored at −20 °C until transport back to the laboratory . Between 44 , 160 and 418 , 176 clones per station were picked and end sequenced from short-insert ( 1 . 0–2 . 2 kb ) sequencing libraries made from DNA extracted from filters [19] . Data from these six Sorcerer II expedition legs ( 37 stations ) were combined with the results from samples in the Sargasso Sea pilot study ( four stations; GS00a–GS00d and GS01a–GS01c; [19] . The majority of the sequence data presented came from the 0 . 8- to 0 . 1-μm size fraction sample that concentrated mostly bacterial and archaeal microbial populations . Two samples ( GS01a , GS01b ) from the Sargasso Sea pilot study dataset and one GOS sample ( GS25 ) came from other filter size fractions ( Table 1 ) . Microbes were size fractionated by serial filtration through 3 . 0-μm , 0 . 8-μm , and 0 . 1-μm membrane filters ( Supor membrane disc filter; Pall Life Sciences , http://www . pall . com ) , and finally through a Pellicon tangential flow filtration ( Millipore , http://www . millipore . com ) fitted with a Biomax-50 ( polyethersulfone ) cassette filter ( 50 kDa pore size ) to concentrate a viral fraction to 100 ml . Filters were vacuum sealed with 5 ml sucrose lysis buffer ( 20 mM EDTA , 400 mM NaCl , 0 . 75 M sucrose , 50 mM Tris-HCl [pH 8 . 0] ) and frozen to −20 °C on the vessel until shipment back to the Venter Institute , where they were transferred to a −80 °C freezer until DNA extraction . Glycerol was added ( 10% final concentration ) as a cryoprotectant for the viral/phage sample . In the laboratory , the impact filters were aseptically cut into quarters for DNA extraction . Unused quarters of the filter were refrozen at −80 °C for storage . Quarters used for extraction were aseptically cut into small pieces and placed in individual 50-ml conical tubes . TE buffer ( pH 8 ) containing 50 mM EGTA and 50 mM EDTA was added until filter pieces were barely covered . Lysozyme was added to a final concentration of 2 . 5 mg/ml−1 , and the tubes were incubated at 37 °C for 1 h in a shaking water bath . Proteinase K was added to a final concentration of 200 μg/ml−1 , and the samples were frozen in dry ice/ethanol followed by thawing at 55 °C . This freeze–thaw cycle was repeated once . SDS ( final concentration of 1% ) and an additional 200 μg/ml−1 of proteinase K were added to the sample , and samples were incubated at 55 °C for 2 h with gentle agitation followed by three aqueous phenol extractions and one phenol/chloroform extraction . The supernatant was then precipitated with two volumes of 100% ethanol , and the DNA pellet was washed with 70% ethanol . Finally , the DNA was treated with CTAB to remove enzyme inhibitors . Size fraction samples not utilized in this study were archived for future analysis . DNA was randomly sheared via nebulization , end-polished with consecutive BAL31 nuclease and T4 DNA polymerase treatments , and size-selected using gel electrophoresis on 1% low-melting-point agarose . After ligation to BstXI adapters , DNA was purified by three rounds of gel electrophoresis to remove excess adapters , and the fragments were inserted into BstXI-linearized medium-copy pBR322 plasmid vectors . The resulting library was electroporated into E . coli . To ensure construction of high-quality random plasmid libraries with few to no clones with no inserts , and no clones with chimeric inserts , we used a series of vectors ( pHOS ) containing BstXI cloning sites that include several features: ( 1 ) the sequencing primer sites immediately flank the BstXI cloning site to avoid excessive resequencing of vector DNA; ( 2 ) elimination of strong promoters oriented toward the cloning site; and ( 3 ) the use of BstXI sites for cloning facilitates the preparation of libraries with a low incidence of no-insert clones and a high frequency of single inserts . Clones were sequenced from both ends to produce pairs of linked sequences representing ~820 bp at the end of each insert . Libraries were transformed , and cells were plated onto large format ( 16 × 16cm ) diffusion plates prepared by layering 150 ml of fresh molten , antibiotic-free agar onto a previously set 50-ml layer of agar containing antibiotic . Colonies were picked for template preparation using the Qbot or QPix colony-picking robots ( Genetix , http://www . genetix . com ) , inoculated into 384-well blocks containing liquid media , and incubated overnight with shaking . High-purity plasmid DNA was prepared using the DNA purification robotic workstation custom-built by Thermo CRS ( http://www . thermo . com ) and based on the alkaline lysis miniprep [92] . Bacterial cells were lysed , cell debris was removed by centrifugation , and plasmid DNA was recovered from the cleared lysate by isopropanol precipitation . DNA precipitate was washed with 70% ethanol , dried , and resuspended in 10 mM Tris HCl buffer containing a trace of blue dextran . The typical yield of plasmid DNA from this method is approximately 600–800 ng per clone , providing sufficient DNA for at least four sequencing reactions per template . Sequencing protocols were based on the di-deoxy sequencing method [93] . Two 384-well cycle-sequencing reaction plates were prepared from each plate of plasmid template DNA for opposite-end , paired-sequence reads . Sequencing reactions were completed using the Big Dye Terminator chemistry and standard M13 forward and reverse primers . Reaction mixtures , thermal cycling profiles , and electrophoresis conditions were optimized to reduce the volume of the Big Dye Terminator mix ( Applied Biosystems , http://www . appliedbiosystems . com ) and to extend read lengths on the AB3730xl sequencers ( Applied Biosystems ) . Sequencing reactions were set up by the Biomek FX ( Beckman Coulter , http://www . beckmancoulter . com ) pipetting workstations . Robots were used to aliquot and combine templates with reaction mixes consisting of deoxy- and fluorescently labeled dideoxynucleotides , DNA polymerase , sequencing primers , and reaction buffer in a 5 μl volume . Bar-coding and tracking promoted error-free template and reaction mix transfer . After 30–40 consecutive cycles of amplification , reaction products were precipitated by isopropanol , dried at room temperature , and resuspended in water and transferred to one of the AB3730xl DNA analyzers . Set-up times were less than 1 h , and 12 runs per day were completed with average trimmed sequence read length of 822 bp . Fosmid libraries [24] were constructed using approximately 1 μg DNA that was sheared using bead beating to generate cuts in the DNA . The staggered ends or nicks were repaired by filling with dNTPs . A size selection process followed on a pulse field electrophoresis system with lambda ladder to select for 39–40 Kb fragments . The DNA was then recovered from a gel , ligated to the blunt-ended pCC1FOS vector , packaged into lambda packaging extracts , incubated with the host cells , and plated to select for the clones containing an insert . Sequencing was performed as described for plasmid ends . Assembly was conducted with the Celera Assembler [21] , with modifications as follows . The “genome length” was artificially set at the length of the dataset divided by 50 to allow unitigs of abundant organisms to be treated as unique , as previously described [19] . Several distinct assemblies were computed . In the primary assembly , all pairs of mated reads were tested to see whether the paired reads overlapped one another; if so , they were merged into a single pseudo-read that replaced the two original reads; further , only overlaps of 98% identity or higher were used to construct unitigs . A second assembly was conducted in the same fashion with the exception of using a 94% identity cutoff to construct unitigs . Finally , series of assemblies at various stringencies were computed for subsets of the GOS data; in these assemblies , overlapping mates were not preassembled and the Celera Assembler code was modified slightly to allow for overlapping and multiple sequence alignment at lower stringency . An all-against-all comparison of unassembled ( but merged and duplicate-stripped ) sequences from the combined dataset was performed using a modified version of the overlapper component of the Celera Assembler [21] . The code was modified to find overlap alignments ( global alignments allowing free end gaps ) starting from pairs of reads that share an identical substring of at least 14 bp . An alignment extension was then performed with match/mismatch scores set to yield a positive outcome if an overlap alignment was found with ≥65% identity . Overlaps involving alignments of ≥40 bp were retained for various analyses . For the GOS dataset described here , this process resulted in a dataset of 1 . 2 billion overlaps . Due to the 14-bp requirement and certain heuristics for early termination of apparently hopeless extensions , not all alignments at ≥65% were found . In addition , some of the lowest-identity overlaps are bound to be chance matches; however , this was a relatively uncommon event . Approximately one in 5 × 106 pairs of 800-bp random sequences ( all sites independent , A = C = G = T = 25% ) can be aligned to overlap ≥40 bp at ≥65% identity using the same procedure . At a 70% cutoff , the value is reduced to one in 4 × 107 , and one in 5 × 108 at a 75% cut off . Like many assembly algorithms , the extreme assembler proceeds in three phases: overlap , layout , and consensus . The overlap phase is provided by the all-against-all comparison described above . The consensus phase is performed by a version of the Celera Assembler , modified to accept higher rates of mismatch . The layout phase begins with a single sequencing read ( “seed” ) that is chosen at random or specified by the user and is considered the “current” read . The following steps are performed off one or both ends of the seed . ( 1 ) Starting from the current fragment end , add the fragment with the best overlap off that end and mark the current fragment as “used , ” thus making the added fragment the new current fragment . ( 2 ) Mark as used any alternative overlap that would have resulted in a shorter extension . The simplest notion of “best overlap” is simply the one having the highest identity alignment , but more complicated criteria have certain advantages . A simple but useful refinement is to favor fragments whose other ends have overlaps over those which are dead ends . For an unsupervised extreme assembly , when the sequence extension terminates because there are no more overlaps , a new unused fragment is chosen as the next seed and the process is repeated until all fragments have been marked used . Sequencing reads mated to SAR11-like 16S sequences but themselves outside of the ribosomal operon ( n = 348 ) were used as seeds in independent extreme assemblies . Since the assemblies were independent , the results were highly redundant , with a given chain of overlapping fragments typically being used in multiple assemblies . A subset of 24 assemblies that shared no fragments over their first 20 kb was identified as follows . ( 1 ) Connected components were determined in a graph defined by nodes corresponding to extreme assemblies . If the assemblies shared at least one fragment in the first 20 kb of each assembly , the two nodes were connected by an edge . ( 2 ) A single assembly was chosen at random from each of the connected components . The consensus sequence over the 20-kb segment of each such representative was used as the reference for fragment recruitment . Phylogenies of sequences homologous to a given portion of a reference sequence ( typically 500 bp ) were determined in the following manner . A set of homologous fragments was identified based on fragment recruitment to the reference as described above . Fragments that fully spanned the segment of interest and had almost full-length alignments to the reference sequence of a user-defined percent identity ( typically , 70% ) were used for further analysis . A preliminary master–slave multiple sequence alignment of the recruited reads ( slaves ) to the reference segment ( master ) was performed with a modified version of the consensus module of the Celera Assembler . Based on this alignment , reads were trimmed to the portion aligning to the reference segment of interest . A refined multiple sequence alignment was then computed with MUSCLE [94] . Distance based phylogenies were computed using the programs DNADIST and NEIGHBOR from the PHYLIP package [95] using default settings . Trees were visualized using HYPERTREE [96] . Based on the low-identity overlap database described above , the similarity of a library i to another library j at a given percent identity cutoff was computed as follows . For each sequence s of i , let ns , i = the number of overlaps to other fragments of i satisfying the cutoff; ns , j = the number of overlaps to fragments of j satisfying the cutoff; and fs , i = ns , i/ ( ns , i + ns , j ) = fraction of reads overlapping s from i or j that are from i . A read that can be overlapped to another at sufficiently high-sequence identity was taken to indicate that they were from similar organisms , and , relatedly , that similar genes were present in the samples . Only reads with such overlaps contributed to the calculation . Other reads reflect genes or segments of genomes that were so lightly sampled ( i . e . , at such low abundance ) that they were not informative regarding the similarity of two samples . Consequently , the analysis automatically corrects for differences in the amount of sequencing , and can be computed over sets of samples that vary considerably in diversity . The resulting measure of similarity Si , j takes on a value between 0 and 1 , where 0 implies no overlaps between i and j , and 1 implies that a fragment from i and a fragment from j are as likely to overlap one another as are two fragments from i or two fragments from j . As with the Bray-Curtis coefficient [97] , abundance of categories affects the computation . In an idealized situation where two libraries can each be divided into some number k of “species” at equal abundance , and the libraries have l of the species in common , the similarity statistic will approach l/k for large samplings; in this sense , Si , j = x indicates that the two samples share approximately a fraction x of their genetic material . It is frequently useful to define Di , j = 1 − Si , j , the “dissimilarity” or distance between two samples . An all-against-all comparison of predicted 16S sequences was performed to determine the alignment between pairs of overlapping sequences using a version of an extremely fast bit-vector algorithm [98] . A hierarchical clustering was determined using percent-mismatch in the resulting alignments as the distance between pairs of sequences . Order of clustering and cluster identity scores were based on the average-linkage criterion , with distances between nonoverlapping partial sequences treated as missing data . Ribotypes were the maximal clusters with an identity score above the cutoff ( typically 97% ) . Representative sequences were chosen for each cluster based on both length and highest average identity to other sequences in the cluster . Taxonomic classification of 16S sequences was conducted using phylogenetic techniques based on clade membership of similar sequences with 16S sequences with defined taxonomic membership . Representative sequences from clustered sequences were analyzed as described previously [19 , 99] and by addition into an ARB database of small subunit rDNAs [100 , 101] . Results were spot-checked against the Ribosomal Database Project II Classifier server [102] and the taxonomic labels of the best BLASTN hits against the nonredundant database at NCBI . Global ocean sequences were aligned to genomic sequences of different bacteria and phage using NCBI BLASTN [26] . The following blast parameters were designed to identify alignments as low as 55% identity that could contain large gaps: -F “m L” -U T -p blastn -e 1e-4 -r 8 -q -9 -z 3000000000 -X 150 . Reads were filtered in several steps to identify the reads that were aligned over more or less their entire length . Reads had to be aligned for more than 300 bp at >30% identity with less than 25 bp of unaligned bases on either end , or reads had to be aligned over more than 100 bp at >30% identity with less than 20 bp of overhang off either end . Identity was calculated ignoring gaps . In some instances a read might be placed , but the mate would not be placed under these criteria . In such cases , if 80% or more of the mate were successfully aligned , then the mate would be rescued and considered successfully aligned . Random pieces of DNA from the genome in question with a length between 1 , 800 to 2 , 500 bp were selected . For each piece a read length N1 was selected from the distribution of lengths using the GOS dataset . If that GOS sequence had a mate pair , then a second length N2 was again randomly selected . The length N1 was used to generate a read from the 5′ end of the DNA . The piece of DNA was then reverse complemented and if appropriate , a second length N2 was used to generate a second read . The relationship between these two reads was then recorded and used to produce a fasta file . This approach successfully mimics the types of reads found in the GOS data with similar rates of missing mates . A total of 2 , 644 proteorhodopsin genes were identified from the clustered open reading frames derived from the GOS assembly [18] . These genes could be linked back to 3 , 608 GOS clones . Open reading frames were predicted from these clones as described in [18] . The peptide sequences were aligned with NCBI blastpgp with the following parameters: -j 5 -U T -e 10 -W 2 -v 5 -b 5000 -F “m L” -m 3 . The search was performed with a previously described blue-absorbing proteorhodopsin protein BPR ( gi|32699602 ) as the query . The amino acid associated with light absorption is found within a short conserved motif RYVDWLLTVPL*IVEF , where the asterisk indicates tuning amino acid [39–41] . In total , 1 , 938 clones were found to contain this motif . Clones and the sample metadata were then associated with the tuning amino acid to determine the relative abundance of the different amino acids at these positions . Clones could be associated with SAR11 if both mated sequencing reads ( when available ) were recruited to P . ubique HTCC1062 . Given a set of genes identified on the GOS sequences , we can identify the scaffolds on which these genes were annotated . A vector indicating the number of sequences contributed by every sample is determined for every gene . This vector reflects the number of sequences from every sample that assembled into the scaffold on which the gene was identified after normalizing for the proportion of scaffold covered by the gene . For example , if a 10-kb scaffold contains a 1-kb gene , then each sample will contribute one GOS sequence for every ten GOS sequences it contributed to the entire scaffold . The vectors are then summed and normalized to account for either the total number of GOS sequences obtained from each sample or based on the number of typically single copy recA genes ( identified as in [18] ) . Unless stated otherwise , recA was used to normalize abundance across samples . When comparisons using groups of samples were performed , the average value for the samples was compared . A 1-D profile representing oligonucleotide frequencies was computed as follows . A sequence was converted into a series of overlapping 10 , 000-bp segments , each segment offset by 1 , 000 bp from the previous one , using perl and shell scripting . Dinucleotide frequencies are computed on each segment using a C program written for this purpose . Higher-order oligonucleotides were examined and gave similar results for the genomes of interest . Remaining calculations were performed using the R package [103] . Principle component analysis ( function princomp with default settings ) was applied to the matrix of frequencies per window position . The value of the first component for each position was normalized by the standard deviation of these values , and truncated to the range [−5 , 5] . For visualizations , the resulting values were plotted at the center of each window . The unrecruited mated sequencing reads of reads recruited to P . marinus MIT9312 at or above 80% identity were examined . An unrecruited mate indicated a potential translocation or inversion if it aligned to the MIT9312 genome in two and only two distinct alignments separated by at least 50 kb , if each aligned portion was at least 250 bp long , if there was less than 100 bp of unaligned sequence and no more than 100 bp of overlapping sequence between the two aligned portions in read coordinates , and if each aligned portion was anchored to one end of the sequencing read with less than 25 bp of unaligned sequence from each end . In total , 18 rearrangements were identified , six of which appear to be unique events . The rate of discovery was estimated by determining the number of rearrangements in a given volume of sequence . We estimated the volume of sequence that was potentially examined by identifying recruited mated sequencing reads that fit the “good” category ( i . e . , which were recruited in the correct orientation at the expected distance from each other ) . For a given read , if the mate was recruited at greater than or equal to 80% identity , then the expected amount of sequence examined should be the current ( as opposed to mate ) read length minus 500 bp . This produces an estimate of the search space to be ~47 Mbp . Given 18 rearrangements , this leads to an estimate of one rearrangement per 2 . 6 Mbp . GOS reads assigned to the “missing mate” category that were recruited at greater than 80% identity outside the gap in question were identified . The mates of these reads were then identified and clustering was attempted with Phrap ( http://www . phrap . org ) . Reads that were incorporated end to end into the Phrap assemblies were identified . For most small gaps a single assembly included all the missing mate reads and identified the precise difference between the reference and the environmental sequences . For the hypervariable segments , most of the reads failed to assemble at all , and those that did show greater sequence divergence than typically seen . In the case of SAR11-recruited reads , to increase the number of reads associated with the hypervariable gaps we identified reads that did not recruit to the P . ubique HTCC1062 but aligned in a single HSP ( high-scoring pair ) over at least 500 bp with one end unaligned because it extended into the hypervariable gap . To facilitate continued analysis of this and other metagenomic datasets , the tools presented here along with their source code will be available via the Cyberinfrastructure for Advanced Marine Microbial Ecology Research and Analysis ( CAMERA ) website ( http://camera . calit2 . net ) . The dataset and associated metadata will be accessible via CAMERA ( using the dataset tag CAM_PUB_Rusch07a ) . Given the exceptional abundance of Burkholderia and Shewanella sequences in the first Sargasso Sea sample and the feeling that these may be contaminants , we are also providing a list of the scaffold IDs and sequencing read IDs associated with these organisms to facilitate analyses with or without the sequences . In addition to CAMERA , the GOS scaffolds and annotations will be available via the public sequence repositories such as NCBI ( http://www . ncbi . nlm . nih . gov/entrez/query . fcgi ? db=genomeprj&cmd=Retrieve&dopt=Overview&list_uids=13694 ) , and the reads will be available via the Trace Archive ( http://www . ncbi . nlm . nih . gov/Traces/trace . cgi ? ) .
The GenBank ( http://www . ncbi . nlm . nih . gov/Genbank ) accession number for proteorhodopsin protein BPR is gi|32699602 .
|
Marine microbes remain elusive and mysterious , even though they are the most abundant life form in the ocean , form the base of the marine food web , and drive energy and nutrient cycling . We know so little about the vast majority of microbes because only a small percentage can be cultivated and studied in the lab . Here we report on the Global Ocean Sampling expedition , an environmental metagenomics project that aims to shed light on the role of marine microbes by sequencing their DNA without first needing to isolate individual organisms . A total of 41 different samples were taken from a wide variety of aquatic habitats collected over 8 , 000 km . The resulting 7 . 7 million sequencing reads provide an unprecedented look at the incredible diversity and heterogeneity in naturally occurring microbial populations . We have developed new bioinformatic methods to reconstitute large portions of both cultured and uncultured microbial genomes . Organism diversity is analyzed in relation to sampling locations and environmental pressures . Taken together , these data and analyses serve as a foundation for greatly expanding our understanding of individual microbial lineages and their evolution , the nature of marine microbial communities , and how they are impacted by and impact our world .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] |
[
"viruses",
"archaea",
"ecology",
"virology",
"microbiology",
"computational",
"biology",
"evolutionary",
"biology",
"genetics",
"and",
"genomics",
"eubacteria"
] |
2007
|
The Sorcerer II Global Ocean Sampling Expedition: Northwest
Atlantic through Eastern Tropical Pacific
|
The adaptive cytotoxic T lymphocyte ( CTL ) -mediated immune response is critical for clearance of many viral infections . These CTL recognize naturally processed short viral antigenic peptides bound to human leukocyte antigen ( HLA ) class I molecules on the surface of infected cells . This specific recognition allows the killing of virus-infected cells . The T cell immune T cell response to Chikungunya virus ( CHIKV ) , a mosquito-borne Alphavirus of the Togaviridae family responsible for severe musculoskeletal disorders , has not been fully defined; nonetheless , the importance of HLA class I-restricted immune response in this virus has been hypothesized . By infection of HLA-A*0201-transgenic mice with a recombinant vaccinia virus that encodes the CHIKV structural polyprotein ( rVACV-CHIKV ) , we identified the first human T cell epitopes from CHIKV . These three novel 6K transmembrane protein-derived epitopes are presented by the common HLA class I molecule , HLA-A*0201 . One of these epitopes is processed and presented via a complex pathway that involves proteases from different subcellular locations . Specific chemical inhibitors blocked these events in rVACV-CHIKV-infected cells . Our data have implications not only for the identification of novel Alphavirus and Togaviridae antiviral CTL responses , but also for analyzing presentation of antigen from viruses of different families and orders that use host proteinases to generate their mature envelope proteins .
The mosquito-borne Chikungunya virus ( CHIKV ) , a member of the Alphavirus genus of the Togaviridae family , causes an acute febrile infection in patients that leads to debilitating arthralgia and arthritis . Identified in the former Tanganyika territory in 1952 [1–3] , this arboviral pathogen caused numerous epidemics in Africa and Asia from the 1960s–1980s [4 , 5] . Following several decades of relative inactivity , CHIKV re-emerged in 2005 to cause an explosive epidemic in the Indian Ocean area , mainly on Reunion Island . In this French overseas department , the outbreak affected about half of its 700 , 000 inhabitants , with more than 250 deaths [5] . In 2006 , several million people were infected by this virus in another large outbreak in India [6] . In recent years , this infectious disease has spread quickly from Africa and Asia to the Americas [7] , causing outbreaks in tropical and subtropical countries of more severe forms than previously reported [8 , 9] . Morbidity due to CHIKV infection is a serious threat to global health and this virus is considered a priority emerging pathogen [10] . CHIKV is an enveloped virus with a positive-sense , single-stranded RNA genome that encodes two large polyproteins [11] . The nonstructural P1234 precursor is autocatalytically processed by the C-terminal domain of the nonstructural protein 2 ( nsP2 ) and releases the four multifunctional nsP proteins . In contrast , in maturation of the structural polyprotein , viral and host proteases are both involved in producing capsid , E1 , E2 , and E3 envelope and 6K transmembrane proteins [11] . Although the immune mechanisms involved in CHIKV disease are not fully understood , CHIKV-infected humans show CD8+ T lymphocyte responses in early disease stages [12]; a large percentage of these activated CD8+ T cells can be detected more than 7 weeks postinfection in patient blood samples [13] . The nature and function of CD8+ T cells during acute and chronic CHIKV infection is largely unknown , as is their association with rheumatic disorders . Although the importance of the HLA class I-restricted immune response has been hypothesized [14] , to date , no human T cell epitope has been described in CHIKV infection . In cellular immunity , CD8+ T lymphocytes recognize short viral peptides exposed at the membrane of infected cells [15] . Most of these epitopes are generated by proteolytic degradation of the fraction of newly synthesized viral proteins whose sequence or folding are in some way defective ( defective ribosomal products; DRiP ) and are thus degraded immediately by the combined action of proteasomes and other cytosol degradative peptidases [16] . The antigen processing products are translocated to the endoplasmic reticulum ( ER ) lumen by transporters associated with antigen processing ( TAP ) , where N-terminal trimming by the ER aminopeptidase ( ERAP ) is frequently necessary [17 , 18] . Some of these final peptides might bind the human histocompatibility complex ( human leukocyte antigen; HLA ) class I heavy chain and β2-microglobulin . The stable trimolecular peptide-HLA-β2-microglobulin complexes are then exported to the cell surface for cytotoxic T lymphocyte ( CTL ) recognition [15] . In addition to this classical antigen processing pathway , several alternative routes have been described that contribute to endogenous HLA class I-restricted antigen processing ( reviewed in [19] ) . During maturation of the viral structural polyprotein , the short CHIKV 6K transmembrane protein is efficiently cleaved by the host ER signal peptidase , rendering it a possible source of viral epitopes via alternative pathways . To search for CHIKV 6K protein T cell epitopes , we infected HLA-A*0201-transgenic mice with a recombinant vaccinia virus that encodes the CHIKV structural polyprotein; we identified three epitopes presented by the HLA class I molecule , one of which is processed and presented in a pathway that involves proteases from distinct subcellular locations .
H-2 class I knockout HLA-A*0201-transgenic mice [20] , a versatile animal model for the study of viral and cancer antigen processing and presentation by the human major histocompatibility complex , were bred in the animal facilities at Centro Nacional de Microbiología , Instituto de Salud Carlos III , in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the Spanish Comisión Nacional de Bioseguridad of the Ministerio de Medio Ambiente y Medio Rural y Marino ( accreditation n° 28079-34A ) . The protocol was approved by the Research Ethics and Animal Welfare Committee of the Carlos III Health Institute ( permit n°: PI-283 ) . All surgery was performed under isoflurane anesthesia , and all efforts were made to minimize suffering . The murine cell line RMA-S ( TAP negative ) transfected with HLA-A*0201 α1α2 domains , and the mouse H-2Db α3 transmembrane and cytoplasmic domains have been described [21] . The cell line was cultured in RPMI 1640 medium supplemented with 10% fetal bovine serum ( FBS ) and 5 μM β-mercaptoethanol ( β-ME ) . A vaccinia virus ( VACV ) Western Reserve ( WR ) strain expressing the CHIKV structural genes ( rVACV-CHIKV ) was constructed by inserting the capsid ( CP ) , E3 , E2 , 6K and E1 structural genes of CHIKV clone LR2006-OPY1 into the TK locus of the WR genome [22] . The rVACV-CHIKV virus expresses the same CHIKV structural genes as those in the reported MVA-CHIKV vaccine candidate [22] . The WR strain used as the parental vector to generate rVACV-CHIKV is an optimized attenuated WR with deletions in the vaccinia immunomodulatory genes A48R , B19R and C11R ( manuscript in preparation ) . CHIKV structural gene expression is under the transcriptional control of the viral synthetic early/late promoter . The rVACV-CHIKV virus was generated , grown in primary chicken embryo fibroblast cells and purified through two 36% ( w/v ) sucrose cushions . Correct CHIKV gene insertion was confirmed by PCR and sequencing , and correct CHIKV protein expression was analyzed by western blot . rVACV-CHIKV was free of contamination with mycoplasma , bacteria or fungi . Peptides were purchased from Biomatik ( Cambridge , Ontario , Canada ) . The correct molecular mass and composition of the peptides at >90% purity was established by quadrupole ion trap micro-high performance liquid chromatography ( HPLC ) . Brefeldin A ( BFA ) and all protease inhibitors were purchased from Sigma-Aldrich ( Saint Louis , MO , USA ) , with the exception of lactacystin ( from Dr . E . J . Corey , Harvard University , Cambridge , MA , USA ) , leupeptin ( Amersham , Little Chalfont , Bucks . , UK ) , pepstatin ( Boehringer Mannheim , Mannheim , Germany ) , and Z-VAD-FMK ( Enzyme System Products , Livermore , CA , USA ) . The specificity of inhibitors used is summarized in Table 1 . SYFPEITHI software ( http://www . syfpeithi . de/Scripts/MHCServer . dll/EpitopePrediction . htm ) was used to predict HLA-A*0201-specific ligands of the 61-residue CHIKV 6K protein . Two synthetic peptides were used as positive and negative controls in complex stability assays , VACV A10L688-696 ( ILDRIITNA , HLA-A*0201-restricted ) [23] and CMV pp657-15 ( RCPEMISVL , HLA-C*01-restricted ) [24] , respectively . HLA-A*0201 RMA-S transfectants were incubated in RPMI 1640 medium with 10% heat-inactivated FBS ( 16 h , 26°C ) . Cells were washed and incubated in the same medium ( 2 h , 26°C ) with different peptide concentrations , further incubated ( 2 h , 37°C ) , and collected for flow cytometry . HLA levels were measured using the PA2 . 1 monoclonal antibody ( anti-HLA-A*02; Abnova , Taipei , Taiwan ) , as described [25] . Samples were acquired on a FACSCanto flow cytometer ( BD Biosciences , San Jose , CA , USA ) and analyzed with FlowJo software ( TreeStar Inc , Ashland , OR , USA ) . The fluorescence index ( FI ) was calculated as the ratio of the mean channel fluorescence of the sample to that of control cells incubated without peptides . Peptide binding was also expressed as the EC50 , which is the molar concentration of peptides that produces 50% maximum fluorescence in a concentration range between 0 . 001 and 100 μM . Polyclonal CHIKV 6K peptide-monospecific CD8+ T cell lines were generated by immunizing transgenic mice with 107 plaque-forming units ( PFU ) of rVACV-CHIKV [26] . Splenocytes from immunized mice were restimulated in vitro with mitomycin C-treated spleen cells pulsed with 10−6 M peptide and cultured Minimum Essential Medium ( Alpha modification; α-MEM ) with 10% FBS , 10−7 M peptide and 5 μM β-ME . Recombinant human interleukin-2 used for long-term propagation of peptide-specific CD8+ T cell lines was generously provided by Hoffmann-LaRoche ( Basel , Switzerland ) . Freshly prepared bone marrow cells were cultured in 200 U/ml GM-CSF ( granulocyte-macrophage colony-stimulating factor; PeproTech , London , UK ) , which was renewed on days 3 and 6 . After 7 days , nonadherent cells with a typical dendritic cell ( DC ) morphology and a myeloid DC phenotype ( MHC class II+ , CD11c+ , CD8− ) were collected as described [27] . ICS assays to detect recognition of peptide-pulsed or infected DC from HLA-A*0201-transgenic mice by polyclonal CTL cell lines were performed as reported [28] . Briefly , CD8+ T cell lines were stimulated ( 4 h ) in the presence of 5 μg/ml BFA and of target DC previously infected with VACV-WR strain or rVACV-CHIKV ( 16 h ) . Cells were then incubated with FITC-conjugated anti-CD8 monoclonal antibody ( mAb; ProImmune , Oxford , UK; 30 min , 4°C ) , fixed with Intrastain kit reagent A ( DakoCytomation , Glostrup , Denmark ) , and incubated with phycoerythrin ( PE ) -conjugated anti-interferon ( IFN ) γ mAb ( BD PharMingen , San Diego , CA , USA ) in Intrastain kit permeabilizing reagent B ( 30 min , 4°C ) . Events were acquired and analyzed as for MHC/peptide stability assays . When protease inhibitors were used , all drugs were added 15 min before the virus and maintained at a 2-fold higher concentration during the 1-h adsorption period than during infection . After washing the virus inoculum , inhibitors were maintained at indicated concentrations for individual experiments . The inhibitors were not toxic at these concentrations , as they did not affect antigen presentation by the VACV D12I251-259-specific CD8+ T cell line . To analyze statistical significance , an unpaired Student t test was used . P values <0 . 05 were considered significant .
The epitope prediction tool SYFPEITHI , a reverse immunology algorithm for MHC ligand motifs [29] , was used to identify possible candidate HLA-A*0201-binding peptides from CHIKV 6K protein . The five nonamers and three decamers ranked as potential HLA-A*0201 ligands ( score >20 ) are depicted in Fig 1 . To study the binding ability of the eight predicted peptides to the HLA-A*0201 molecule , we performed MHC-peptide complex stability assays using HLA-A*0201-transfected , TAP-deficient RMA-S cells . Four peptides ( 6K31-39 , 6K37-46 , 6K45-54 , 6K51-59 ) were bound to the HLA-A*02:01 class I molecules ( Fig 2 ) , with EC50 values in the range commonly found among natural high-affinity ligands such as the VACV A10L HLA-A*0201 epitope . In contrast , HLA affinity was substantially lower for 6K22-30 and 6K22-31 peptides , and both were considered medium-affinity ligands ( Fig 2 ) . 6K21-29 peptide binding to HLA-A*02:01 was residual , with a EC50 value >200 μM ( Fig 2 ) . Stable numbers of HLA-peptide surface complexes were not detected with the 6K28-36 peptide ( Fig 2 ) . These data suggest that most of these peptides could be presented by the HLA-A*02:01 molecule in CHIKV-infected cells . In contrast to HLA-B*0702 transgenic mice , in which strong ex vivo VACV-specific T cell responses were detected [28] , peptide-specific IFNγ-secreting cells from VACV-immunized HLA-A*0201 transgenic mice were usually detected only after in vitro stimulation . The cause of these differences is unclear , especially as both transgenic mouse types were generated in the same laboratory [20 , 30] . From rVACV-CHIKV-immunized HLA-A*0201 transgenic mice , we produced polyclonal CTL lines monospecific for each of seven CHIKV 6K peptides with stable numbers of HLA-peptide surface complexes detected in MHC-peptide complex stability assays ( Table 2 ) . The CTL lines stimulated with three of the four HLA-A*0201 high-affinity peptides ( 6K31-39 , 6K45-54 , 6K51-59 ) specifically recognized peptide-pulsed DC ( Fig 3 ) . There was no specific recognition of peptide-pulsed cells by the other four CHIKV 6K peptides ( 6K21-29 , 6K22-30 , 6K22-31 , and 6K37-46; Table 2 ) ; this lack of response was confirmed using several immunization and in vitro stimulation protocols ( not shown ) . These data indicated that CHIKV 6K31-39 , 6K45-54 , and 6K51-59 peptides are HLA-A*0201-restricted CTL epitopes , and were recognized simultaneously as part of the memory response to rVACV-CHIKV . The small ( 61-residue ) CHIKV 6K protein thus contains at least three distinct HLA-A*0201-restricted epitopes , two of which overlap partially . As the three CHIKV 6K viral epitopes derive from the same 6K protein , we studied the CD8+ CTL line specific for the CHIKV 6K51-59 epitope as a representative of antigen processing of this viral protein . The CHIKV 6K51-59 epitope-specific CD8+ CTL line specifically recognized rVACV-CHIKV- but not wild type VACV-infected cells , while another T cell line specific for VACV D12I peptide 251–259 recognized both infected cells ( Fig 4 ) . CHIKV 6K is a structural protein necessary both for virus budding and entry , which is incorporated in small amounts into the virion [11] . As rVACV-CHIKV expresses the five structural proteins of the pathogen , we cannot rule out the presence of CHIKV virus-like particles and possible exogenous antigen presentation . To test whether the CHIKV 6K51-59 HLA-A*0201-restricted epitope requires endogenous processing , we analyzed its presentation in the presence of BFA . Brefeldin A blocks class I export beyond the cis-Golgi compartment [31 , 32] , preventing surface expression of newly assembled HLA class I-peptide complexes of endogenous origin ( Table 1 summarizes the specificity of all inhibitors used ) . BFA addition during infection completely inhibited specific IFNγ secretion by the CHIKV 6K51-59 epitope-specific CD8+ T cell line ( Fig 5 ) , which demonstrated that this epitope was generated from CHIKV 6K protein endogenously processed in rVACV-CHIKV-infected cells . We also observed complete inhibition of specific IFNγ secretion by the VACV D12I251-259 epitope-specific CD8+ T cell line ( Fig 5 ) . To study the antigen processing pathways involved in endogenous generation of the CHIKV 6K51-59 epitope , we performed ICS assays with several specific protease inhibitors on rVACV-CHIKV-infected cells . We tested E64 [33] , leupeptin ( LEU ) [34] , pepstatin ( PEPST ) [34 , 35] , 1 , 10-phenanthroline ( PHE ) , and phenylmethylsulfonyl fluoride ( PMSF ) [36] inhibitors , as they are specific for different protease families and cover a wide range of protease classes ( Table 1 ) . None of these inhibitors affected specific recognition of rVACV-CHIKV-infected target cells by the CHIKV 6K51-59-specific CD8+ T cell line ( Fig 6 ) . The enzymes inhibited by these drugs are thus not involved in generation of this epitope . We also tested specific inhibitors of several cellular proteases , most of which were not relevant for antigen processing of the CHIKV 6K51-59 viral epitope ( Fig 7 ) . In contrast , dec-RVKR , an inhibitor of furin and other proprotein convertases ( Table 1 ) , partially inhibited CHIKV 6K51-59-specific CD8+ T cell recognition of infected cells ( 43 ± 20%; Fig 7 ) . To exclude the possibility that this inhibition was due to toxic effects on target cells or on VACV replication rather than to a specific protease block , we performed parallel experiments using the rVACV-CHIKV-infected target cells with another T cell line . These infected cells were recognized efficiently by the VACV D12I251-259-specific CD8+ T cell line , and no inhibition was detected ( 4 ± 6%; Fig 7 ) . These data indicate that the dec-RVKR-induced inhibition of specific recognition by CHIKV 6K51-59-restricted CD8+ T cells was due to protease blockade and not to nonspecific effects . These data indicate that proprotein convertases are involved in the generation of the CHIKV 6K epitope . The inhibitor puromycin ( PURO ) [37] ( Table 1 ) partially blocked specific recognition of rVACV-CHIKV-infected target cells by CHIKV 6K51-59-specific CD8+ T cells ( 47 ± 21% ) , but had no effect on VACV D12I251-259 epitope presentation ( 4 ± 6% ) ( Fig 8 ) . PURO is a reversible inhibitor of the cytosol alanyl aminopeptidase and of lysosomal DPPII . To identify the specific peptidase involved in CHIKV 6K51-59 peptide processing , we treated rVACV-CHIKV-infected target cells with additional inhibitors . CHIKV 6K51-59-specific CD8+ T cell recognition was unaffected by two distinct inhibitory compounds that block cytosol alanyl aminopeptidase activity , bestatin ( BEST ) and EDTA ( ethylenediaminetetraacetic acid ) ( Table 1 and Fig 8 ) , which excludes this cytosolic enzyme from antigen processing of the CHIKV 6K51-59 epitope . The antimalarial drug chloroquine ( CQ ) , a lysosomotropic agent that affects DPPII and other lysosomal enzymes ( Table 1 ) , nonetheless blocked recognition of infected cells by CHIKV 6K51-59-specific CD8+ T cells ( 69 ± 19%; Fig 8 ) . These data indicate that DPPII is involved in CHIKV 6K epitope generation . The inhibition of antigen recognition by dec-RVKR ( Fig 7 ) or PURO ( Fig 8 ) indicated that furin-like proteases and DPPII peptidase are both involved in antigen presentation of the CHIKV 6K51-59 epitope . The similar partial inhibition of rVACV-CHIKV-infected cell recognition by both drugs ( dec-RVKR , 43 ± 20%; PURO , 47 ± 21% ) is compatible with two explanations . The CHIKV 6K51-59 epitope might be processed sequentially by the two proteases . Alternatively , this epitope could be processed in parallel by proprotein convertases or by DPPII independently; in this case , both antigen processing pathways would have to be inhibited simultaneously to fully abrogate CHIKV 6K51-59 epitope presentation . To discriminate between these possibilities , we analyzed the effect on antigen presentation of the combined inhibitors on rVACV-CHIKV-infected cells . We observed a moderately increased blockage of presentation in target cells treated simultaneously with PURO and dec-RVKR ( 66 ± 6%; Fig 9 ) , comparable and not statistically different to that observed when CQ and dec-RVKR were combined ( 71 ± 18%; Fig 9 ) or with CQ alone ( 69 ± 19%; Fig 8 ) . The inhibitory effect of PURO and dec-RVKR was CHIKV 6K epitope-specific , as recognition of the VACV D12I251-259 epitope was not reduced in their presence ( Fig 9 ) . These results show that furin-like proteases and DPPII are found in the same CHIKV 6K51-59 epitope presentation pathway . To test whether the classical antigen processing pathway is involved in CHIKV 6K51-59 epitope generation , we used the proteasome inhibitor lactacystin ( LC ) [38 , 39] , and leucinthiol ( Leu-SH ) , which has activity against ERAP and other metallo-aminopeptidases [40] ( Table 1 ) . Both LC ( 83 ± 3% ) and Leu-SH ( 91 ± 13% ) blocked specific recognition of rVACV-CHIKV-infected target cells by CHIKV 6K51-59-specific CD8+ T cells ( Fig 10 ) . In contrast , in the same experiment , these drugs had a lesser effect on VACV D12I251-259 epitope presentation ( Fig 10 ) . CHIKV 6K51-59 epitope presentation to a specific T cell line was partially blocked by dec-RVKR ( 43 ± 20% ) , PURO ( 47 ± 21% ) or both ( 66 ± 6% ) ( Figs 7 , 8 and 9 ) , whereas CHIKV 6K51-59 recognition by these CD8+ T cells was strongly inhibited by LC ( 83 ± 3% ) and Leu-SH ( 91 ± 13% ) ( Fig 10 ) . These differences were statistically significant ( Table 3 ) , which suggested that the CHIKV 6K51-59 epitope is generated by two distinct pathways , the classical antigen processing pathway and a second antigen presentation pathway that includes the four proteases ( dec-RVKR , PURO , LC and Leu-SH ) .
In this study , we undertook identification of HLA-A*0201 epitopes from the CHIKV 6K protein and explored their antigen presentation pathways . Our results define several CHIKV 6K protein restricted epitopes , being to our knowledge the first time that epitopes from CHIKV are defined associated to human MHC class I molecules . Extended epitope prediction using the SYFPEITHI tool suggests that ligands of this small viral protein could be presented by a notable proportion of the HLA class I alleles tested ( 12 of 30; 40% , S1 Table ) . According to the Immune Epitope Database ( IEDB ) population coverage tool ( http://tools . iedb . org/population/ ) , these class I molecules are present in 86% of the human population ( S2 Table ) . The short viral CHIKV 6K protein is thus of interest for targeting the cellular immune system . Further studies are needed to analyze cellular immune responses in CHIKV-infected individuals . Here we identified three HLA-A*0201-restricted epitopes in the CHIKV small 6K protein . Using several protease inhibitors ( Table 1 ) , we report that various proteolytic activities ( probably in two distinct antigen processing pathways ) are necessary to generate one of these epitopes , the CHIKV 6K51-59 epitope . These results are consistent with a model for CHIKV maturation and processing and , by extrapolation , that of other Alphavirus structural polyproteins ( Fig 11 ) . Although no furin cleavage motif was found in the 6K protein , 6K51-59 peptide presentation was dependent on dec-RVKR-sensitive proteases , which indicates that proprotein convertase activity is needed to generate this epitope . Like other host and viral proteases [41] , furin are involved in processing structural polyproteins in all Alphaviruses to yield the mature structural proteins that will form the virion . Maturation of the CHIKV structural polyprotein thus affects antigen processing of the 6K51-59 epitope . With regard to the CHIKV replication cycle , only limited information can be extrapolated from comparison between CHIKV structural proteins and those of other Alphaviruses . The translation order of these Alphavirus polyproteins is capsid , PE2 precursor ( that includes envelope glycoproteins E3 and E2 , 6K , and envelope protein E1 ) [41] . Immediately after the ribosome starts translation of the PE2 precursor , the capsid protein ( whose C-terminal domain has protease activity ) is released in the cytosol by autoproteolysis . The new N terminus of the polyprotein thus bears a signal sequence for translocation of the PE2 precursor across the ER membrane . Additional signal sequences in the C terminus of E2 and 6K proteins allow their translocation to the ER . In the ER , signal peptidase cleavage of the C terminus of both the PE2 precursor and the 6K protein releases three viral protein products ( PE2 , 6K , and E1 ) . The PE2 precursor and E1 protein remain attached to the membrane by their C terminus , and 6K remains as a short transmembrane protein . In the Alphavirus Sindbis virus , E1 and PE2 glycoproteins form a heterodimer in the ER , and this interaction is sufficient for transport beyond this organelle [42] . The E1-PE2 heterodimer reaches the trans-Golgi network , but prior to the cell membrane , CHIKV PE2 is cleaved by furin and other proprotein convertases such as PC5A , PC5B , and PACE4 to generate the mature E2 and E3 proteins [43] . CHIKV 6K51-59 epitope-dependent presentation by dec-RVKR-sensitive proteases thus indicates that CHIKV envelope proteins are transported from the ER to the trans-Golgi network as heterotrimers that also include CHIKV 6K protein , as also described for Semliki Forest virus [44] . In Alphaviruses , this cleavage induces conformational changes in E1 and E2 proteins , thus promoting extensive contacts between these two proteins to yield the spike architecture of activated viral envelope complex in 1:1 stoichiometry [41] . The role of the E3 structural protein is unclear; E3 is associated with virions in Semliki Forest virus [45] , but not in other Alphaviruses including CHIKV [46 , 47] . In Semliki Forest and Sindbis viruses , substoichiometric amounts of 6K are incorporated into the virion [44 , 48] . Most of this small protein must thus be discarded in the infected cells , although the fate of the CHIKV 6K protein is nonetheless unclear . As both PURO and CQ impaired antigen recognition of target cells by CHIKV 6K51-59-specific CD8+ T cells , the lysosomal DPPII must have a role in processing this epitope . This data also indicated that at least a fraction of the CHIKV 6K protein must be degraded in the lysosomes . DPPII-processed CHIKV 6K protein or a fragment that includes the viral epitope must be transported to the cytosol for proteasome processing , as indicated by LC inhibition of the CHIKV 6K51-59 antigen presentation . How these fragments reached the cytosol remains unclear , but the proteasome is involved in the generation of some epitopes of the Epstein-Barr virus ( EBV ) latent membrane protein 2 ( LMP2 ) transmembrane nucleoprotein , albeit by uncharacterized mechanisms [49] . Host cell transmembrane protein processing might be involved in both CHIKV and EBV epitopes [50] . The block in CHIKV 6K51-59-specific recognition by Leu-SH , but not by two drugs that do not inhibit ERAP activity ( the general metalloproteinase inhibitor PHE and the aminoprotease inhibitor BEST ) [51 , 52] , indicates that ERAP or a similar metalloproteinase produces the final CHIKV 6K51-59 epitope , probably in the ER , after transport of proteasomal products by TAP . The statistically different percentages of inhibition by LC and Leu-SH inhibitors vs . dec-RVKR , PURO and CQ drugs also suggest a direct contribution of the classical antigen processing pathway , with proteasome degradation of DRiPs from viral polyprotein followed by ERAP trimming . The relative contribution of both pathways to antigen presentation was quantified using the same percentage of inhibition from one-third to half by the classical antigen processing pathway and half to two-thirds by the circular antigen presentation pathway . Our results show a broad diversity of proteases involved in a complex antigen presentation pathway to yield the viral CHIKV 6K epitope . In addition to proteasome and ERAP , several proteases are implicated in processing endogenously synthesized HLA class I antigens ( reviewed in [19] ) . Many proteases included here , such as signal peptidase [53 , 54] , furin [55 , 56] , and uncharacterized lysosomal CQ-sensitive enzymes [57 , 58] , have been linked independently to the processing of several epitopes , although sequential activity of these peptidases to generate a specific HLA class I epitope has not been described . These proteases and the supplementary involvement of DPPII in CHIKV 6K51-59 antigen presentation define the most complex antigen processing and presentation pathway reported to date; this route begins in the ER and includes the trans-Golgi network , lysosomes , retrograde transport to cytosol , and ER re-entry . Lastly , the results reported here also have implications for analysis of the cellular immune response . Only proteasome and ERAP , but not other protease inhibitors , are generally used to analyze antigen presentation of different HLA class I ligands or epitopes . Inhibition is normally sufficient to formally assign presentation of an epitope to the classical antigen processing pathway , excluding additional protease activities ( Fig 10 ) . In addition to the Alphavirus genus and the Togaviridae family , however , many viruses of different families and orders use host proteases from distinct subcellular locations to generate mature envelope and even nuclear proteins . In other viral epitopes , it would thus not be unexpected to find complex antigenic processing and presentation pathways similar to those reported here , if the antiviral cellular immune response was analyzed in depth with a broad spectrum of protease inhibitors as was carry out in the current investigation . In conclusion , the results of the present report highlight the diversity of peptidases involved in HLA class I antigen presentation and expose the complexity of antigen processing pathways , as represented by the CHIKV 6K protein . Definition of the importance of this epitope in natural infection nonetheless awaits studies in CHIKV-infected individuals . This process could have broad implications when applied to other viral proteins .
|
The arboviral pathogen Chikungunya virus ( CHIKV ) is a serious threat to global health , and is considered a priority re-emerging virus . This pathogen causes acute febrile infection in patients , leading to debilitating arthralgia and arthritis . In recent years , CHIKV has spread quickly in tropical and subtropical countries , causing outbreaks of more severe forms of the disease than previously reported . The nature and function of the T cell immune response , critical for clearance of viral infections , is largely unknown during acute and chronic CHIKV disease and their association with rheumatic disorders . In this study , we identified the three first CHIKV epitopes recognized by human T cells . We studied how one of these epitopes is generated in virus-infected cells , a process that involves the sequential proteolytic activity of several proteases at distinct subcellular locations . We postulate that this process could have broad implications when applied to other viral proteins .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"blood",
"cells",
"innate",
"immune",
"system",
"medicine",
"and",
"health",
"sciences",
"antigen",
"presentation",
"lysosomes",
"immune",
"cells",
"pathology",
"and",
"laboratory",
"medicine",
"togaviruses",
"chikungunya",
"infection",
"enzymes",
"pathogens",
"immunology",
"tropical",
"diseases",
"microbiology",
"enzymology",
"alphaviruses",
"viruses",
"chikungunya",
"virus",
"rna",
"viruses",
"neglected",
"tropical",
"diseases",
"cytotoxic",
"t",
"cells",
"cellular",
"structures",
"and",
"organelles",
"infectious",
"diseases",
"white",
"blood",
"cells",
"animal",
"cells",
"proteins",
"medical",
"microbiology",
"microbial",
"pathogens",
"t",
"cells",
"immune",
"system",
"biochemistry",
"cell",
"biology",
"antigen",
"processing",
"and",
"recognition",
"viral",
"pathogens",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"proteases",
"viral",
"diseases",
"organisms"
] |
2017
|
Complex antigen presentation pathway for an HLA-A*0201-restricted epitope from Chikungunya 6K protein
|
Species barriers , expressed as hybrid inviability and sterility , are often due to epistatic interactions between divergent loci from two lineages . Theoretical models indicate that the strength , direction , and complexity of these genetic interactions can strongly affect the expression of interspecific reproductive isolation and the rates at which new species evolve . Nonetheless , empirical analyses have not quantified the frequency with which loci are involved in interactions affecting hybrid fitness , and whether these loci predominantly interact synergistically or antagonistically , or preferentially involve loci that have strong individual effects on hybrid fitness . We systematically examined the prevalence of interactions between pairs of short chromosomal regions from one species ( Solanum habrochaites ) co-introgressed into a heterospecific genetic background ( Solanum lycopersicum ) , using lines containing pairwise combinations of 15 chromosomal segments from S . habrochaites in the background of S . lycopersicum ( i . e . , 95 double introgression lines ) . We compared the strength of hybrid incompatibility ( either pollen sterility or seed sterility ) expressed in each double introgression line to the expected additive effect of its two component single introgressions . We found that epistasis was common among co-introgressed regions . Interactions for hybrid dysfunction were substantially more prevalent in pollen fertility compared to seed fertility phenotypes , and were overwhelmingly antagonistic ( i . e . , double hybrids were less unfit than expected from additive single introgression effects ) . This pervasive antagonism is expected to attenuate the rate at which hybrid infertility accumulates among lineages over time ( i . e . , giving diminishing returns as more reproductive isolation loci accumulate ) , as well as decouple patterns of accumulation of sterility loci and hybrid incompatibility phenotypes . This decoupling effect might explain observed differences between pollen and seed fertility in their fit to theoretical predictions of the accumulation of isolation loci , including the ‘snowball’ effect .
Intrinsic postzygotic isolation ( hybrid inviability and sterility that occurs independently of the environment ) is often due to deleterious genetic interactions between loci that have functionally diverged during the evolution of new species ( i . e . , Dobzhansky-Muller incompatibilities , or DMIs; [1] ) . Several models of this process assume that individual DMIs are due to epistasis among loci in the diverging lineages ( e . g . , two loci , one in each lineage , would form a pairwise genetic interaction ) , and that each DMI contributes independently to total hybrid incompatibility ( i . e . , there is no epistasis between DMIs; but see [2–4] for models that relax this assumption ) . If hybrid dysfunction is due to more complex interactions among loci , however , there can be important consequences for the temporal accumulation of species barriers and the number of loci required to complete speciation [4–6] . In the simplest case , if epistasis between different hybrid incompatibility loci is antagonistic ( i . e . , if the combined effect of two different DMIs is less than expected based on their individual effects on hybrid incompatibility ) , then a greater time to speciation is expected and correspondingly more loci are required . Conversely , if epistasis between different conspecific loci is typically synergistic ( i . e . , the combined effect on hybrid incompatibility is greater than expected based on individual effects ) , fewer DMIs will be required for the expression of complete reproductive isolation , with a correspondingly shorter time to speciation . In the latter case , although the initial observation of incompatibility phenotypes requires more than one substitution per lineage , hybrid incompatibility can be expressed rapidly between two diverging species once substitutions have begun to accumulate . The prevalence of epistasis , and whether this epistasis is synergistic or antagonistic , can therefore be critical in determining rates of evolution of isolation between diverging lineages by governing the accumulation dynamics of alleles that contribute to species barriers ( e . g . , [4 , 5]; and see below ) . Empirically , however , very little is known about the nature of epistasis among different loci contributing to hybrid incompatibility ( ‘complex epistasis’ , c . f . [7] ) and whether these interactions typically act to enhance or retard the expression of hybrid incompatibility between species . Some evidence suggests that complex interactions might be common [8 , 9] . In Drosophila , for example , quantitative trait loci ( QTL ) for male sterility show evidence of complex epistasis . Several individual genomic regions are simultaneously required for the expression of some male sterility phenotypes ( e . g . , [10]; other evidence is reviewed in [1] ) , consistent with synergism ( i . e . , greater-than-additive effects ) between conspecific loci for the expression of hybrid sterility . However , epistasis has been difficult to assess in many genetic analyses of quantitative trait loci because early recombinant generations have low power to detect these interactions [11] . The most promising method of directly estimating epistasis among loci is to subtract background effects by isolating two individual QTL in an otherwise isogenic background ( e . g . , [12] ) . For example , this approach revealed less-than-additive effects for quantitative and ecophysiological traits in tomato , consistent with antagonistic interactions between different conspecific loci [13 , 14] . Similarly , studies in microorganisms have used serial pairwise combinations of target loci to examine the size and direction of pairwise epistasis on the phenotypic effects of , for example , double deletion strains in yeast [15] , and metabolic flux mutants in yeast and E . coli [16] . Nonetheless , the serial combination of many pairs of conspecific loci has yet to be used to assess prevalence and direction of interactions influencing hybrid incompatibility phenotypes [17] . The goal of this study was to assess the prevalence and direction of genetic interactions between different chromosomal regions from one species , when combined pairwise in the background of a second species ( S1 Fig ) , focusing specifically on the expression of hybrid incompatibility . We studied fifteen chromosomal regions ( Table 1 ) , drawing from a set of 71 introgression lines ( ILs ) between two plant species in the genus Solanum Section Lycopersicon ( the tomato clade ) [18] . Each IL contains a unique short chromosomal region from the wild species S . habrochaites ( ‘hab’ ) introgressed into the otherwise isogenic genetic background of the domesticated tomato , S . lycopersicum ( ‘lyc’ ) ( [18]; see [19] for a previous summary ) . The fifteen ILs included in our study , several of which have effects on either pollen or seed sterility ( [19] , and see Results ) , were previously chosen and crossed to generate nearly 100 double introgression lines ( DILs; [20] ) . Here , we draw on this collection of DILs to study the effect of interactions among conspecific introgressed regions on the expression of known hybrid incompatibility phenotypes . Specifically , we compared fertility phenotypes in homozygous DILs to those of their corresponding parental ILs to infer frequency , magnitude and direction of epistasis among conspecific introgressions . We found that complex interactions were common in both pollen and seed fertility phenotypes , although significantly more frequent for introgressions affecting pollen sterility . These observed interactions were frequently antagonistic , whereby the combined effect of pairwise introgressions produced a less severe effect on fitness than predicted from individual effects . For pollen sterility phenotypes , we found that some chromosomal regions were considerably more prone to interaction than others: most interactions occurred among introgressions with significant individual effects . In contrast , we found no evidence of a similar pattern for seed sterility . This pervasive antagonism has critical implications for the predicted rate and pattern of accumulation of hybrid incompatibility between these lineages , and for future empirical studies of the evolution of post-zygotic isolation . We found unexpectedly abundant interactions among genes underlying the phenotypes studied , which implies a violation of key assumptions—regarding the independence among different pairs of incompatibility loci—made by many speciation models . We propose that this high connectivity among loci that contribute to pollen sterility could explain the observed flattened accumulation of ( pairwise ) pollen incompatibilities in this clade [20] . In addition to explaining this apparent lack of “snowball” effect , our results also imply that standard QTL mapping approaches can underestimate the number of incompatibility-causing interactions present in systems with high levels of epistasis .
Each unique genotype was grown and assessed in replicate ( initiated with at least 6 individuals per genotype; total experiment size = 702 individuals; S2 Table ) in a fully randomized common garden ( greenhouse ) experiment at the IU Biology greenhouse facility . Cultivation methods have been previously described [14] . Briefly , all experimental seeds were germinated under artificial lights on moist filter paper , transferred to soil post-germination ( at the cotyledon stage ) , and repotted into 1gallon pots at 4–6 weeks post-germination . A total of 652 individuals reached reproductive maturity . Due to mortality and/or fecundity variation among individuals , we were able to collect pollen and seed fertility data from 569 and 620 individuals , respectively ( S2 Table ) . Because of these inviability or fertility effects , not all DIL genotypes had a minimum of two biological replicates and so were not included in analyses . In total , we analyzed pollen and seed fertility data for 95 and 93 DILs , respectively ( see S3 Table , Results and Discussion ) . Seed fertility ( seeds per fruit ) was determined by quantifying seed production from self-pollinated fruits ( ‘self-seed set’ ) , as with previous QTL analyses [21] . At least two flowers per plant were allowed to produce fruit via selfing; when fruits did not develop automatically , flowers were self-pollinated manually to ensure that floral morphology was not responsible for preventing self-fertilization . Upon maturation , individual fruits were harvested , seeds extracted by hand , and seed fertility determined by counting the number of visible seeds from each fruit . Average seed per fruit for each plant was used to generate self-seed set estimates for each introgression genotype and the lyc control parent . Pollen fertility ( quantified as the proportion of fertile pollen ) was estimated on two unopened flowers on each plant , as previously described [19] . Briefly , all pollen ( the entire anther cone ) from each target flower was collected into lactophenol-aniline blue histochemical stain , homogenized , and a known sub-sample of homogenate used to count inviable and viable pollen grains using microscopy . Pollen inviability was indicated by the absence of a stained cytoplasm , a conservative measure of pollen infertility [22] . Seed set phenotypes may be affected by three components: ovule viability , pollen fertility , and gamete compatibility in the zygote . Therefore , pollen sterility in our lines may influence estimates of seed sterility . In fact , we found a weak but significant correlation between pollen fertility and self-seed set across our experimental lines ( P = 0 . 0014; negative-binomial GLM ) . However , pollen fertility explained only about 5% of variation in seed fertility ( pseudo-R2 = 0 . 046 ) , which is consistent with previous studies [19 , 21] . Nonetheless , we removed this small pollen effect by carrying out our analyses of self-seed set on their residuals from the regression on pollen fertility ( including the genotype as a random effect in a GLM ) . For completeness , we carried out the same analyses on the uncorrected self-seed set values; our results do not differ qualitatively between these analyses ( S2 Fig ) but , as the ‘pollen-corrected’ data is more conservative with respect to inferring sterility effects , we report these results here . Given this , the factors affecting ‘seed sterility’ estimates that we report are expected to be ovule inviability and early zygotic lethality ( such that an initiated seed would not be visible in the mature fruit ) .
We found three prevalent patterns of conspecific epistasis for hybrid incompatibility . First , epistatic interactions between conspecific loci are common , especially among introgressions with larger individual effects on fertility ( Figs 1 and 2 ) . Second , these interactions are predominantly antagonistic ( i . e . , less-than-additive ) in their effects on hybrid incompatibility phenotypes . Third , both these patterns are significantly more pronounced for pollen than seed sterility effects . For seed sterility , we found evidence of fertility-affecting interactions between introgressions in about one fifth of DILs analyzed ( Fig 1 ) . Out of 93 DILs , 16 showed departures from fitness models of independence ( P < 0 . 01 , FDR = 5 . 6% ) . Interactions were qualitatively more common for pollen fertility phenotypes ( Fig 2 ) , where 25 out of 95 analyzed DILs were best fit by epistatic fitness models ( P < 0 . 01 , FDR = 3 . 4% ) . Remarkably , most of these interactions were antagonistic: in seed phenotypes , 10 DIL hybrids were less sterile than expected by their IL parents’ individual effects . This pattern was considerably more dramatic for pollen fertility , where all 25 detected epistatic interactions were antagonistic ( Fig 2 , S3 Table ) . We found little difference between pollen and seed sterility phenotypes in the magnitude of epistasis or the best fit fitness model ( S3 Table ) . Epistatic interactions caused , on average , a 37% excess in pollen fitness when compared to the fitness predicted from the joint effect of parental ILs . In seed sterility , we found similar effect sizes in synergistic and antagonistic interactions ( 33% and 45% , respectively ) . In most DILs with significant epistatic effects , an additive fitness model had a better fit than a multiplicative one ( 10 out of 16 in seed; 21 out of 25 in pollen ) ; this was also true in DILs with no significant epistatic effects for seed ( 39 out of 77 DILs ) but not for pollen phenotypes ( 16 out of 70 ) . Note , however , that most differences in likelihood values between the two fitness models are probably not significant: only 6 cases ( out of 188 analyzed DIL phenotypes ) showed likelihood differences that would be significant in a chi-squared test ( i . e . , greater than 3 log-likelihood units , roughly P = 0 . 01 ) . This apparent bias in favor of an additive fitness model could be because this model predicts slightly larger variance than its multiplicative counterpart , resulting in a better fit even when there are no biologically meaningful differences in predicted means . Epistatic interactions for fitness were common across the genome: most parental ILs were involved in at least one interaction with another conspecific introgression ( all except LA3948 for seed and LA3931 for pollen fertility; Figs 1 and 2 ) . In pollen phenotypes , however , interactions more frequently involved introgressions that had individual sterility effects . This bias becomes more apparent in highly significant interactions ( 14 DILs at FDR = 1%; bordered panels in Fig 2 ) , which always involved DILs with at least one pollen-sterile IL parent genotype; moreover , 10 of these 14 highly significant interactions involved two sterile ILs . Individually , each of the observed interactions could be used to infer whether the incompatibility they participate in is pairwise or higher-order ( Fig 3 ) . A pairwise DMI could be inferred when an antagonistic interaction is observed between a sterile and a non-sterile IL ( found in 4 cases for seed , and 4 for pollen phenotypes ) , since phenotype rescue in the DIL is consistent with complementation of loci carried by parental ILs . On the other hand , higher-order DMIs could be inferred from two types of interactions . First , all synergistic interactions ( 2 cases in seed phenotypes ) imply introgressions that are part of a third-order ( or higher ) DMI in which a specific combination of alleles is necessary for sterility . Second , antagonistic epistasis between sterile ILs ( the most common in our data; 10 cases , all in pollen phenotypes ) implies an incompatibility involving at least three loci ( that is , a fitness-affecting interaction involving at least two loci in one lineage and a third locus in the other lineage ) . We can rule out pairwise DMIs because in that case only one of the parental ILs is expected to be individually sterile ( not both ) as the allele that complements is expected to be compatible with the heterospecific background ( i . e . , there should be asymmetry of DMI allelic effects [32] ) . However , we cannot determine how these three or more loci interact: the parental ILs could carry loci involved in a single higher-order DMI , or the interaction could be the result of complex epistasis between two pairwise incompatibilities , for example . Regardless of these individual inferences , the overall pattern of interactions we detect indicates there is a complex landscape of interdependence among sterility loci . This pattern can be visualized by representing the distribution of interactions as a network ( Fig 4 ) . In pollen , ILs formed a single interconnected cluster in which each line had , on average , 1 . 9 highly significant interactions . Sterile ILs for pollen had more interactions ( i . e . , a higher node degree; 3 . 5 on average ) than non-sterile introgressions ( 0 . 5 on average ) . In contrast , the overall network of interactions in seed phenotypes was sparser than in pollen phenotypes , with three separate clusters of up to 4 ILs . The seed sterility network showed no direct connections between two sterile ILs , and sterile ILs had similar number of interactions than non-sterile ILs ( 0 . 8 on average for both groups ) . The number of interactions in which an IL was involved is not correlated with its mean effect size for seed sterility ( unlike for pollen sterility; S3 Fig ) , suggesting that the lower overall connectivity and reduced enrichment for individually sterile ILs we observed in seed phenotypes is robust to our arbitrary cut-off for significance of individual IL sterility ( FDR of 1% for pollen and 5% for seed ) .
By examining multiple pairwise combinations of conspecific loci , co-introgressed into a heterospecific background , we have shown evidence for pervasive non-independence among introgressions . This non-independence is likely due to a combination of pairwise interactions , higher-order incompatibilities , and complex epistatic effects , and suggests that hybrid phenotypes might frequently be the product of epistatic interactions among several factors . For pollen sterility phenotypes , we found that some chromosomal regions were more prone to interaction than others: most interactions occurred among introgressions with significant individual effects . In contrast , we found no evidence of a similar pattern for seed sterility . Finally , the observed interactions in pollen phenotypes were overwhelmingly antagonistic , whereby the combined effect of pairwise introgressions produced a less severe effect on fitness than predicted from individual effects . This observed pervasive antagonism could have important implications for understanding both the predicted patterns and the mechanism of accumulation of hybrid incompatibility between lineages , as well as for the detection of such loci in QTL mapping efforts . Interactions were common in both pollen and seed fertility phenotypes , although considerably more frequent for introgressions affecting pollen sterility . Interactions in seed suggest at least two higher-order incompatibilities , composed of three and four loci ( if we focus on highly significant edges in Fig 4 ) . In the pollen network , ILs form a single cluster of highly significant interactions , and most of these interacting ILs are epistatic with at least two other introgressions ( 8 out of 10 ILs; Fig 4 ) . Speculating on the genetic details underlying these patterns of interdependence is not straightforward . One extreme interpretation is that this large module represents a single incompatibility involving ten loci . This hypothetical tenth-order DMI could result , for example , if a new substitution triggers incompatibility between the recipient locus and some combination of previously substituted alleles at nine other loci ( similar to how high-order DMIs were modeled in [32] ) . Alternatively , each of these antagonistic interactions represent the gradual accumulation of increasingly complex DMIs ( in which new substitutions interact with previously established DMIs ) , which implies that pollen-sterility loci are frequently involved in more than one DMI ( since many ILs show more than one interaction ) . This latter interpretation is also consistent with our observation that pollen sterile ILs show more interactions than non-sterile ones , since each interaction in which they engage could contribute to their observed fitness ( i . e . , ILs involved in more DMIs would , on average , show larger fitness effects ) . While our approach does not allow us to distinguish between these alternatives , our observations likely represent a mix of higher-order DMIs and loci involved in multiple pairwise incompatibilities . However , the involvement of loci in more than one hybrid incompatibility violates a central assumption in classical models of their evolutionary accumulation ( an issue that we examine below ) . The differences in connectivity between seed and pollen phenotypes could explain previously observed differences in the pattern of accumulation of hybrid incompatibilities ( or ‘snowball’ effect ) in this group . Moyle & Nakazato [20] showed that , while the accumulation of seed sterility QTL among increasingly divergent Solanum lineages follows the predicted snowball ( i . e . , it is faster than linear ) , this does not seem to be the case for QTL involved in pollen fertility . At first , this linear accumulation of pollen fertility QTLs may seem at odds with our observation of high connectivity for the phenotype , because theory predicts a faster increase in the number of incompatibilities with increasing connectivity among loci . In the simplest version of the snowball model , the number of incompatibilities is roughly proportional to the square of K—the number of genetic differences between two species , and p—the probability that any pair of these is incompatible [32] . Therefore , factors that increase K or p should increase the speed of the snowball . For instance , K could be elevated for phenotypes experiencing selective pressures , such as antagonistic coevolution , that result in sustained elevated rates of substitution [33 , 34] . Similarly , the probability that any pair of genetic differences is incompatible ( p ) is elevated when there are more opportunities for incompatibilities to arise , such as when the phenotype is controlled by a highly connected network of genes . Given these predictions , our results would suggest that the number of detected pollen sterility loci should snowball faster than seed sterility loci ( yet , they do not ) . A possible explanation for this discordance lies in the assumption that the number of loci involved in incompatibilities represents an adequate proxy for the total number of incompatibilities . Because of the challenges of identifying underlying molecular loci , empirical studies have used the number of sterility QTL as an indicator of the number of DMIs , with the underlying assumption that each QTL is involved in only one ( pairwise ) incompatibility . Under a pairwise snowball model , each locus is involved in ( K-1 ) p incompatibilities on average [3] . Since p ( the probability of an incompatible pair ) is assumed to be very small compared to K , such that ( K-1 ) p << 1 , each QTL is therefore likely to be involved in just one incompatibility . In cases with increased p , however , sufficient genetic differences could accumulate that ( K-1 ) p ≥ 1 ( i . e . , loci are , on average , involved in one or more incompatibilities ) . After this “saturation” point , each new genetic difference will continue to generate ( K-1 ) p new incompatible interactions , but the number of loci involved in hybrid incompatibility increases just by one . Note that many loci could be involved in more than one incompatibility before the system reaches saturation so that the deceleration of the snowball need not be abrupt [35] . We illustrate the phenomenon of saturation by simulation of the accumulation of pairwise incompatibilities ( Fig 5 ) . These simulations ( carried out in R; [23] ) simply count the number of pairwise incompatibilities between two lineages as substitutions fix , allowing each new substitution to be incompatible with any of the existing ones at that time ( with probability p ) . As Fig 5 shows , incompatibilities ( i . e . , interactions that affect hybrid fitness ) in saturated systems continue to snowball while the number of individual loci contributing to hybrid incompatibility starts to increase linearly , breaking the relationship between sterility-affecting loci and DMIs . Our simulations are consistent with recent analytical results that also show that , as p increases , the number of incompatibility loci tends to underestimate the number of pairwise incompatibilities [35] . Saturation of the gene network for pollen fitness would be consistent with our observation that several pollen-sterile ILs showed multiple interactions ( which suggests they are involved in a higher-order DMI or multiple pairwise ones ) and with the previously observed lack of snowball for this phenotype ( [20] . Interpreted under a framework of saturation , these two observations suggest many loci underlying pollen fertility are already involved in at least one incompatible interaction between species , so that the accumulation of new interactions does not result in the increase of unique sterility loci at a similar rate . In other words , pollen sterility might be close to saturation for hybrid incompatibility loci ( the nodes in our network; Fig 4 ) , even while the DMIs in which they are involved ( the edges in our network; Fig 4 ) continue to snowball among incompatibility loci . If pollen fertility is indeed saturated , it follows that pollen-affecting DMIs may be accumulating in tomatoes faster than previously thought ( i . e . , [20] ) because the number of incompatibility QTL underestimates the number of DMIs . This inference is similar to that of recent simulation work on the accumulation of DMIs in an RNA-folding model , which showed that the number of pairwise DMIs will not snowball in systems where these interactions are converted into higher-order DMIs ( when new substitutions result in the involvement of a new locus in an existing DMI; [35] ) . Interestingly , this recent theoretical work also showed that the rate of accumulation of pairwise DMIs slows down as the probability of incompatibility increases [35] . This implies that the high connectivity of the pollen network could be responsible for the observed lack of snowball even if the network is not fully saturated . Properties of the genome-wide gene network , such as modularity and degree distribution , could also contribute to our proposed saturation of pollen sterility . For instance , highly modular genomes can allow for high p within a module ( e . g . , among the genes controlling pollen fertility ) even when there is a low genome-wide mean probability of incompatibility . Previous estimates of genome-wide p in tomatoes ( ~10−9; [20] ) are smaller than those estimated for Drosophila ( 10−7–10−8 , [3] ) , and much smaller than would seem to be required for saturation . However , our inference of saturation in the pollen network implies large p among pollen loci only , which need not be reflected in p genome-wide . A skewed degree distribution could also contribute to an apparent saturation by having highly connected loci involved in multiple DMIs while most loci are not involved in an incompatibility . In other words , pollen sterility could be primarily controlled by highly connected genes ( i . e . , be network ‘hubs’ ) , which have a much higher probability of interacting than the genome-wide average . On the other hand , evolutionary processes could have accelerated the rate of divergence specifically at loci involved in fertility traits ( that is , have elevated K at pollen loci only , rather than across the whole genome ) . For reproductive traits , this coevolution could arise from antagonistic male-female interactions during reproduction ( e . g . , [36 , 37] ) . In contrast to these inferences from our pollen data , our observations imply that seed sterility loci in these species might be more likely to engage in DMIs that involve simple pairwise interactions ( rather than higher-order epistasis ) , and that this phenotype is further from saturation . That is , each additional seed sterility-causing substitution is more likely to originate a new DMI with a unique interacting partner , rather than engaging in fitness-affecting interactions with existing sterility loci . In this case , seed sterility loci themselves will still appear to snowball with divergence between lineages; indeed , seed sterility loci have been shown to snowball between Solanum species [20] . While our conjecture regarding the saturation of different sterility phenotypes needs to be confirmed with further work , a more general implication of our results is that QTL mapping ( or any count of individual hybrid incompatibility loci ) might not always provide a suitable estimate of the number and complexity of accumulating DMIs . Despite the considerable additional experimental burden this implies , future empirical tests of the snowball effect should aim to assess both the number of apparent incompatibility loci and evidence for non-independence among these loci . Moreover , we note that the high frequency of interactions with antagonistic effects on sterility also has implications for the pattern of accumulation of reproductive isolation phenotypes ( rather than loci or DMIs ) between these species . The amount of reproductive isolation is expected to snowball if DMIs are independent and have , on average , similar fitness effect sizes . If , however , antagonistic epistasis among sterility loci is common , the accumulation of reproductive isolation is expected to be less-than-linear , because new DMIs will tend to contribute ‘diminishing returns’ on fitness . This lack of a snowball specifically in incompatibility phenotypes has been previously noted [38 , 39] , and pervasive phenotypic antagonism could contribute to explaining this common observation [35] . Finally , in contrast to its phenotypic consequences , the mechanistic basis of pervasive genetic antagonism is unclear . This pattern of diminishing deleterious effects is consistent with observations in E . coli [40] and yeast [15] , but the underlying basis is unresolved . For Drosophila , Yamamoto et al . [41] proposed that similar patterns of antagonism might be expected among loci that are involved in stabilizing selection within populations; however , this is unlikely to be relevant here both because our interacting loci are not normally segregating within populations , and because fertility traits are less likely to be subject to stabilizing selection at some intermediate phenotypic optimum ( compared to other phenotypic traits ) . Alternatively , less-than-additive fitness effects can result when different mutations each have deleterious individual effects within the same developmental pathway , but their pairwise combination suppresses or ameliorates these deleterious effects ( as has been observed , for example , with double deletion strains in yeast; [15] ) . Our proposal that , in saturated systems , new substitutions are likely to interact with loci already involved in hybrid incompatibility , is consistent with sterility-affecting mutations accumulating in a tightly-connected network such as a specific developmental pathway . Of course , additional empirical work is needed to evaluate this possibility . Moreover , the frequency of antagonism and the potential for phenotypic diminishing returns are likely to be dependent on the properties of the diverging gene network . Further theoretical work is needed to determine how specific network properties , such as its connectivity and modularity , can affect the patterns of accumulation of incompatibilities and reproductive isolation .
|
A characteristic feature of new species is their inability to produce fertile or viable hybrids with other lineages . This post-zygotic reproductive isolation is caused by dysfunctional interactions between genes that have newly evolved changes in the diverging lineages . Whether these interactions occur between pairs of divergent alleles , or involve more complex networks of genes , can have strong effects on how rapidly reproductive isolation—and therefore new species—evolve . The complexity of these interactions , however , is poorly understood in empirical systems . We examined the fertility of hybrids that carried one or two chromosomal regions from a close relative , finding that hybrids with two of these heterospecific regions were frequently less sterile than would be expected from the joint fitness of hybrids that have the same regions singly . This ‘less-than-additive’ effect on hybrid sterility was widespread ( observed in 20% of pairwise combinations ) , and especially pronounced for male sterility . We infer that genes contributing to male sterility form a more tightly connected network than previously thought , implying that reproductive isolation is evolving by incremental dysfunction of complex interactions rather than by independent pairwise incompatibilities . We use simulations to illustrate these expected patterns of accumulation of reproductive isolation when it involves highly interconnected gene networks .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"pollen",
"epistasis",
"plant",
"anatomy",
"plant",
"science",
"fitness",
"epistasis",
"cell",
"biology",
"quantitative",
"trait",
"loci",
"chromosome",
"biology",
"heredity",
"genetics",
"biology",
"and",
"life",
"sciences",
"introgression",
"evolutionary",
"biology",
"seeds",
"evolutionary",
"processes",
"genetic",
"loci",
"chromosomes"
] |
2017
|
Pervasive antagonistic interactions among hybrid incompatibility loci
|
We are using the fungus Neurospora crassa as a model organism to study the circadian system of eukaryotes . Although the FRQ/WCC feedback loop is said to be central to the circadian system in Neurospora , rhythms can still be seen under many conditions in FRQ-less ( frq knockout ) strains . To try to identify components of the FRQ-less oscillator ( FLO ) , we carried out a mutagenesis screen in a FRQ-less strain and selected colonies with altered conidiation ( spore-formation ) rhythms . A mutation we named UV90 affects rhythmicity in both FRQ-less and FRQ-sufficient strains . The UV90 mutation affects FRQ-less rhythms in two conditions: the free-running long-period rhythm in choline-depleted chol-1 strains becomes arrhythmic , and the heat-entrained rhythm in the frq10 knockout is severely altered . In a FRQ-sufficient background , the UV90 mutation causes damping of the free-running conidiation rhythm , reduction of the amplitude of the FRQ protein rhythm , and increased phase-resetting responses to both light and heat pulses , consistent with a decreased amplitude of the circadian oscillator . The UV90 mutation also has small but significant effects on the period of the conidiation rhythm and on growth rate . The wild-type UV90 gene product appears to be required for a functional FLO and for sustained , high-amplitude rhythms in FRQ-sufficient conditions . The UV90 gene product may therefore be a good candidate for a component of the FRQ-less oscillator . These results support a model of the Neurospora circadian system in which the FRQ/WCC feedback loop mutually interacts with a single FLO in an integrated circadian system .
Circadian rhythms are approximately 24 h cycles of behavior , physiology , etc . that are driven by an endogenous biological clock . They are found at all levels of eukaryotic life , and in some prokaryotes . Molecular models for the endogenous oscillators that drive these rhythms in eukaryotes are based on rhythmic transcription of a small number of “clock genes” , translation of the rhythmic transcripts into rhythmic levels of clock proteins , and negative regulation of the clock genes by their own proteins . These models have been called transcription/translation feedback loops ( TTFL ) . Although a great deal of evidence has accumulated to support the concept of negative and positive feedback loops regulating the expression of these clock genes , there are many indications that the TTFL model is inadequate as a complete clock mechanism ( reviewed in [1] ) . In recent mammalian examples , it was found that cycling of mammalian clock proteins CLOCK and CRY is not required for rhythmicity in fibroblasts [2]; large reductions in overall transcription rate and levels of clock proteins do not eliminate circadian oscillations in mouse fibroblasts [3]; and rhythmic cAMP signaling is required to sustain rhythmic transcription in mammalian cells [4] . These and other findings are not compatible with the canonical TTFL models . In the prokaryotic cyanobacterium Synechococcus elongatus , a post-translational oscillator has been shown to be the core pacemaker . It can operate in vitro in the absence of any transcription or translation , but in vivo a TTFL is coupled to it as a slave oscillator [5] , [6] . The architecture of the cyanobacterial clock provides certain advantages that may be relevant to eukaryotic clock systems as well [6] . Progress in understanding circadian oscillators in eukaryotes will depend on identifying clock components outside of the TTFL [7] , [8] . The filamentous fungus Neurospora crassa is a model organism that has provided many insights into the molecular basis of circadian rhythmicity [9]–[11] . Asexual spore formation ( conidiation ) is controlled by the circadian clock in Neurospora , and rhythmic spore formation can be easily monitored during growth as a pattern of thick conidiation “bands” alternating with thin growth in “interbands” as the fungal mycelium advances across a solid agar surface . The current model for the Neurospora circadian oscillator ( called the FRQ/WCC TTFL ) consists of interlocked negative and positive feedback loops involving the clock genes frq , wc-1 and wc-2 [9] , in which a complex of the WC-1 and WC-2 proteins ( WCC ) activates transcription of frq and is in turn negatively regulated by FRQ protein . However , there have been many reports of rhythmicity with periods in the circadian range that can be seen in strains with null mutations in frq or wc genes . These include conidiation rhythms [12]–[22] and molecular/biochemical rhythms [23]–[26] . These rhythms are said by some to be driven by many different “FRQ-less oscillators” ( FLOs ) that may be completely separate from the FRQ/WCC feedback loop [26]–[28] . An alternative view sees the FRQ/WCC TTFL as a component of a larger circadian system in which the FRQ/WCC TTFL interacts with a single FLO [29] . Neurospora provides a unique eukaryotic system in which we can directly access clock components outside of the TTFL by assaying rhythmicity in FRQ-less strains in which the TTFL is not functioning . We are using this system to identify components of FLO by searching for mutations affecting rhythmicity in FRQ-less strains . Such mutations can also provide insight into the interactions between the FRQ/WCC TTFL and the FLO . The new mutation we report here disrupts two FRQ-less rhythms and also affects the FRQ/WCC TTFL , suggesting that a single FLO may interact with the FRQ/WCC TTFL in a circadian architecture similar to that of cyanobacteria .
To isolate mutations affecting the FLO , we conducted a mutagenesis screen to identify mutations affecting conidiation rhythms in a frq knockout background . We used UV light to mutagenize uninucleate microconidial spores of a strain carrying both the frq10 knockout allele and chol-1 . Strains carrying the chol-1 mutation express conidiation rhythms in FRQ-less strains on choline-deficient media , and these rhythms are therefore an example of a FLO [16] . On high choline , chol-1 strains are identical to chol+ [16] . The parent strain for mutagenesis was both frq10 and chol-1 and was therefore rhythmic without choline but arrhythmic on high choline ( Figure 1A , upper 4 tubes ) . We analyzed 600 colonies derived from mutagenized spores on individual race tubes and found one mutant that was arrhythmic both with and without choline ( Figure 1A , lower 4 tubes ) . The mutation was named UV90 , as it was the 90th spore analyzed after UV mutagenesis . This strain , genotype csp-1; chol-1 rasbd; frq10; UV90 , was backcrossed to rasbd and the phenotypes of the progeny were analyzed . The chol-1 phenotype was assayed by comparing growth rates on choline-supplemented and choline-deficient media . The conidiation banding phenotypes were easily identified among the progeny and these phenotypes were consistently identified by two individuals using blind-coded tubes or photographs . Periods and growth rates of all strains described below are listed in Table 1 , and conidiation phenotypes are presented in Figure 1 . The chol-1 progeny were identified by slow growth on low choline medium and rapid growth on high choline ( Table 1 ) . Among the chol-1 progeny , the frq10; UV90 double mutant phenotype was identified by its similarity to the mutant parent: short aerial hyphae with very even conidiation on high choline ( Figure 1A , tubes 5 and 6 ) and heavier conidiation and disrupted rhythmicity on low choline ( Figure 1A , tubes 7 and 8 ) . Densitometry traces of individual chol-1; frq10 progeny are presented in Figure 2 . The chol+ progeny were identified by rapid growth on both high and low choline ( Table 1 ) . Among the chol+ progeny , the frq10; UV90 double mutants ( Figure 1B ) were identified as having a phenotype on choline-free medium that was identical to the chol-1; frq10; UV90 parent strain on high choline ( Figure 1A , tubes 5 and 6 ) . When the frq10 mutation was crossed out of the original UV90 strain , the effects of the UV90 mutation on the circadian system when the FRQ/WCC oscillator is functional could be observed . The chol-1; frq+ progeny ( Figure 1C ) were identified by the presence of rhythmicity on high choline ( Figure 1C , tubes 1 , 2 , 5 and 6 ) . Progeny in which the long-period rhythm on low choline was abolished ( Figure 1C , tubes 7 and 8 ) were identified as UV90 . Densitometry traces of individual chol-1; frq+ progeny are presented in Figure 3 . These progeny also showed a low-amplitude rhythm on high choline ( Figure 1C , tubes 5 and 6 ) . The chol+; frq+ progeny ( Figure 1D ) were identified as indistinfguishable on choline-deficient medium from the phenotypes of the chol-1 progeny on high choline ( Figure 1C , tubes 1 , 2 , 5 and 6 ) . The UV90 progeny were identified by the damped conidiation rhythm ( Figure 1D , tubes 3 and 4 ) . The low amplitude rhythm of frq+ UV90 strains was due to an increase in average levels of conidiation , as shown by densitometry ( Figure 4A ) . One of the csp-1; rasbd; UV90 progeny was backcrossed to the Mauriceville wild-type to remove the rasbd mutation . The ras+ progeny were identified by growth rates more rapid than the rasbd strains ( Table 1 ) . Although these progeny ( Figure 1E ) produced very poor banding rhythms , as expected , a UV90 phenotype was noticeable , increasing conidiation levels and producing nearly constant conidiation ( Figure 1E , tubes 3 and 4 , and Figure 4B ) . We conclude that the UV90 mutation severely compromises the function of the FLO in chol-1 , completely disrupting conidiation rhythmicity in the frq10 knockout background in choline-deficient conditions . UV90 also damps the amplitude of the conidiation rhythm in frq+ strains in the presence of a functional FRQ/WCC TTFL . The UV90 phenotype was found to segregate in backcrosses with the expected 1∶1 ratio for segregation of a single-gene mutation . In the backcross of the original csp-1; chol-1 rasbd; frq10; UV90 mutant to rasbd , out of 88 csp-1; rasbd progeny assayed , 44 were classified as UV90 and 44 as UV90+ based on banding phenotypes on high and low choline . In the cross to Mauriceville of a putative csp-1; rasbd; UV90 strain , out of 40 csp-1; rasbd progeny assayed , 21 were classified as UV90 and 19 were classified as UV90+ based on banding phenotype . Among the chol-1 progeny of the backcross of the original mutant to rasbd , the two UV90 banding phenotypes segregated together . Out of 45 csp-1; chol-1 rasbd progeny tested , 39 gave clear phenotypes on choline-deficient medium and 6 were unclear . Of those 39 clear phenotypes , 20 produced bands on low choline similar to UV90+ and all had high-amplitude banding rhythms on high choline; 19 were arrhythmic on low choline and all produced low-amplitude bands on high choline similar to UV90 . The UV90 mutation slightly decreased the growth rates in most cases ( Table 1 ) . This property also co-segregated with the UV90 banding phenotypes . Figure 5 presents growth rate data for frq+ progeny grown on high and low choline media , and it can be seen that progeny with a UV90+ banding phenotype , both chol+ and chol-1 , clustered together at higher growth rates than UV90 progeny . The UV90 mutation mapped to the right arm of linkage group ( LG ) VI . In the backcross of the original UV90 strain to rasbd , the UV90 phenotype was unlinked to either frq ( LG VIIR ) or chol-1 ( LG IVR ) : out of 88 progeny tested , 50 . 0% were recombinant with frq and 48 . 9% were recombinant with chol-1 . In a cross to the multiply-marked alcoy tester strain [30] , recombination frequencies were: 51 . 9% with cot-1 ( LG IVL/VR ) , 40 . 7% with al-1 ( LG IR/IIL ) and 16 . 7% with ylo-1 ( LG IIIR/VIR ) . These results suggested that UV90 is linked to either LG III or LG VI . The original UV90 strain was crossed to the Mauriceville wild-type and mapped using cleaved amplified polymorphic sequence ( CAPS ) markers and bulked segregant analysis according to the method of Jin et al . [31] . The UV90 phenotype was unlinked to a CAPS marker on LG III ( 3–52-EcoRI in Jin et al . ) and was linked to CAPS markers 6–39-HaeII on the left of the centromere on LG VI , and 6–68-MspI on the right of the centromere . Recombination rates between UV90 and 6–39-HaeII were roughly 5–10% ( estimated from the relative band intensities of the cleaved PCR fragments from the bulked segregants ) and recombination between UV90 and 6–68-MspI was roughly 0–5% , suggesting that UV90 maps to the right arm of LG VI , in a gene-rich region of the chromosome . One of the defining characteristics of a circadian rhythm is that the period of the rhythm does not change appreciably when the organism is maintained at different constant temperatures; this property of “temperature compensation” distinguishes circadian clocks from simple chemical reactions that increase in rate with an increase in temperature . We assayed the conidiation rhythm of our new mutant strain at different constant growth temperatures to determine the effects of the UV90 mutation on period and temperature compensation of the rhythm . All strains were wild-type for frq . Both chol+ and chol-1 strains were assayed on high choline to repair the defect in chol-1 . As shown in Figure 6A , the UV90 mutation significantly increased the period by several hours at some temperatures , and altered the response of the period to temperature in comparison to the UV90+ strains . The UV90 mutation also decreased the growth rate by small but significant amounts at temperatures above 19°C ( Figure 6B ) . These results suggest that the UV90 mutation may have minor effects on metabolism affecting growth , and minor but significant effects on the temperature compensation of the circadian clock in the presence of a functional TTFL . It has been shown that FRQ-less strains can be entrained to repeated pulses of high temperature and behave as if they contain a functional heat-entrainable oscillator [15] , . This heat-entrainable oscillator is a second example of a FLO . We used 2-hour pulses of 32°C on cultures growing at 22°C to entrain the conidiation rhythm to various T-cycles ( where T = total number of hours in the cycle ) ( Figure 7 ) . Changes in peak timing and peak shape in different T-cycles are indicative of entrainment of an underlying oscillator [19] , [32] . In the presence of wild-type frq ( Figure 7A–7D ) , the UV90 mutation had small but significant effects on the shape of the entrained peaks . In the FRQ-less background ( Figure 7E–7H ) , the UV90 mutation had a dramatic effect on the peak timing , shifting the major peak to a much earlier time . We conclude that the UV90 mutation has a greater effect on entrainment behavior in the absence of functional FRQ ( in the frq10 strain ) than in the presence of functional FRQ , and therefore is likely to primarily affect the FLO rather than the FRQ/WCC feedback loop . The damping of the amplitude of the conidiation rhythm seen in UV90 mutant strains with functional FRQ ( Figure 1D ) could be due to either an effect on the output from the circadian oscillator , such as continuous activation of the conidiation developmental pathway , or an effect on the amplitude of the circadian oscillator itself . To distinguish between these two possibilities , the amplitude of the oscillator was probed with phase-resetting stimuli . According to oscillator theory [33] , [34] , an oscillator with a small amplitude will respond to a particular stimulus with a larger phase resetting response than will an oscillator with a large amplitude . A similar effect should be seen with any type of stimulus , regardless of the input pathway it uses [33] . We used both light pulses and high temperature pulses to probe the amplitude of the oscillator in the UV90 mutant strains in the frq+ chol+ background . Data were plotted in both the traditional phase response curve ( PRC ) format , plotting phase shifts against the circadian time of the pulse , and in the phase transition curve ( PTC ) format [35] , plotting new phase against old phase . The PTC format is preferred , as it does not introduce artifactual breaks in the data where large phase delays meet large phase advances [36] . As seen in Figure 8 , the PRCs for the UV90 mutant showed larger phase shifts than the wild-type for both light and heat stimuli . The PTCs indicate that under these conditions , the wild-type produced type 1 PTCs and the UV90 mutant produced type 0 PTCs . A type 1 PTC is indicative of a weaker response than a type 0 PTC [34]–[36] . Because the same stimuli produced a weak response from wild-type and a stronger response from the UV90 mutant , these results are consistent with a smaller oscillator amplitude for the UV90 mutant [36] . We assayed the relative levels of FRQ protein in the UV90 mutant strain and the UV90 wild-type to determine whether the UV90 mutation affects the expression of FRQ . Three circadian cycles were assayed every three hours , from 3 to 72 h in DD ( constant darkness ) , and levels of FRQ protein were directly compared between the two strains . The pattern of FRQ expression in the wild-type ( Figure 9A ) was very similar to the pattern we previously reported for cultures on solid agar medium across two cycles from 24 to 69 h in DD [21] . The phosphorylation pattern of FRQ was similar in mutant and wild-type ( Figure 9B ) . Expression of FRQ in the UV90 mutant was lower than wild-type in the first cycle and rapidly damped out such that expression was much lower in the second and third cycles ( Figure 9A ) . Mean expression levels from three replicate experiments indicated that the level of FRQ expression in UV90 was significantly lower than in UV90+ ( Figure 9C ) . We conclude that the UV90 mutation affects the functioning of the TTFL by damping the rhythm of FRQ protein . It is interesting to note that , although the levels of FRQ protein in the UV90 mutant were significantly lower than UV90 wild-type ( Figure 9 ) , there was very little effect on the period of the conidiation rhythm at the temperature used in this experiment ( 22°C , Figure 6; see also Table 1 ) .
The UV90 mutation was found to affect two FRQ-less rhythms: the free-running rhythm under low-choline conditions ( Figure 1A and 1C ) and the heat-entrained rhythm in choline-sufficient conditions ( Figure 7E–7H ) . Some prd mutations also affect more than one FRQ-less rhythm: The prd-1 mutation affects FRQ-less rhythms in the cel [37] and chol-1 [29] mutant backgrounds , in geraniol-supplemented cultures [17] , and under heat entrained conditions [29] , and prd-2 affects the last three rhythms but has not been assayed for effects on cel . As with UV90 , it is not known whether these prd mutations affect other FRQ-less rhythms that have not yet been assayed . Each of the products of the UV90 , prd-1 and prd-2 genes may have multiple molecular targets and multiple functions , affecting several independent FLOs . Alternatively , if we assume the simplest model , that each gene product has one major function and one major target , then these results are consistent with the hypothesis that all FRQ-less rhythms are driven by a single FRQ-less oscillator , or FLO [29] . The various FRQ-less rhythms that have been reported appear in some cases to have different characteristics , and this may be seen to support the hypothesis of multiple independent FLOs . For example , temperature compensation is reported to be defective in some cases [14] , [18] , and in others the period is compensated within a temperature range but not outside of that range [16] , [22] . This is similar to the observations of differences in temperature compensation between period-affecting frq mutants and other clock-affecting mutants in the presence of a functional FRQ/WCC TTFL , in which the genetic background can affect temperature compensation in frq+ strains [38] . FRQ-less rhythms have been reported from several different laboratories , and the assay conditions , media , and genetic backgrounds of the strains differ . It is possible that under the widely varying conditions and genetic backgrounds used to assay the various FRQ-less rhythms , the properties observed may be affected by those conditions . The same suite of properties has not been assayed for all FRQ-less rhythms , and it is often impossible to directly compare them . For example , the molecular rhythm in DAG [25] has not been characterized beyond the observation that it persists in frq-less strains; the nitrate reductase rhythm [23] is known to persist in frq and wc null mutants and in both constant dark and constant light; but in neither case has temperature compensation been assayed nor is it known what molecular mechanisms drive these rhythms . In the absence of further details about the characteristics of the various FRQ-less rhythms , the most parsimonious assumption is that a single oscillator can drive all observed FRQ-less rhythms and this single FLO is affected by multiple gene products such as UV90 , prd-1 and prd-2 . In the absence of a functional FRQ/WCC TTFL , the FLO may continue to oscillate with a low amplitude and with alterations of properties such as temperature compensation and period stability; the various conditions under which FRQ-less rhythms can be observed may act on different parameters of the FLO to increase its amplitude and allow FRQ-less rhythms with differing properties to be expressed . Two lines of evidence suggest that the UV90 mutation primarily affects the FLO rather than the FRQ/WCC feedback loop . First , the free-running conidiation rhythm in low-choline frq10 cultures is completely abolished by UV90 while a low-amplitude rhythm persists in the presence of functional frq ( Figure 1 ) . Second , entrainment to heat pulses is affected more strongly in the frq10 background than in frq+ ( Figure 7 ) . However , the UV90 mutation does have significant effects in the presence of functional frq . The amplitude of the free-running conidiation rhythm is dampened ( Figure 1D ) , the temperature compensation of the free-running conidiation rhythm is affected ( Figure 6 ) , the amplitude of the oscillator as assayed by phase resetting is reduced ( Figure 8 ) and the expression of FRQ protein is dampened ( Figure 9 ) . The UV90 gene product may have several molecular functions and may affect the FRQ/WCC TTFL independently of its effects on the FLO , but we prefer the simpler explanation stated previously that UV90 has a single major target . In this case , these findings lend support to the hypothesis that the FRQ/WCC feedback loop and the FLO interact with each other . The long-period conidiation rhythm in the chol-1 strain grown on choline-deficient medium persists in the absence of the FRQ/WCC TTFL , in either frq knockout or wc mutant strains [16] . This defines it as a FRQ-less rhythm driven by a FLO , but the long-period conidiation rhythm is also seen in frq+ strains . This raises the question: What is happening to the FRQ/WCC TTFL when frq+ cultures on low choline are producing long-period conidiation rhythms driven by the FLO ? Shi et al . [28] found short-period rhythms of frq promoter activity in chol-1 cultures with long-period conidiation rhythms , and concluded that the FLO in chol-1 is “in no way connected to the circadian system” [28] . These authors did not use mathematical techniques to look for relative coordination or frequency demultiplication between the short-period molecular rhythm and the long-period conidiation rhythm and so might have missed evidence of interactions . We have previously reported evidence for an influence of the FRQ/WCC TTFL on the FLO in chol-1: a short-period frq1 mutation [39] and a wc-2 mutation [16] can significantly alter the long period in chol-1 , and introducing the frq10 mutation into chol-1 can lengthen the long period ( Table 1 and [29] ) . If our suggestion that the UV90 mutation primarily affects the FLO is correct , then our results now provide evidence for an influence of the FLO on the FRQ/WCC TTFL . The UV90 gene product is required for normal functioning of the FLO and to sustain the amplitude of the FRQ/WCC TTFL . Evidence that prd mutations can affect rhythmicity in both frq+ and frq10 backgrounds [29] suggests interactions between these two oscillators . The complete system maybe more complex , with multiple targets for the UV90 and prd gene products , and multiple oscillators , but the current data set can be explained with this two-oscillator model . Further experiments will be needed clarify the functions of the gene products in question , identify their targets , and describe the mechanism of the FLO ( s ) . We conclude that the simplest model for the architecture of the circadian system of Neurospora [29] is a single FLO that mutually interacts with , and is required to support , the FRQ/WCC TTFL . The fungal circadian system may thus be conceptually similar to the cyanobacterial system , in which a post-translational core oscillator interacts with a TTFL [6] . The coupling of a TTFL with a post-transcriptional oscillator is an emerging theme in circadian biology [40] , with examples coming from plants [41] , [42] and animals [8] , [43] as well as the cyanobacteria . Mathematical modeling demonstrates that such a system of coupled oscillators could be more robust to noise and changes in growth rate than either oscillator alone [6] , [44] . Neurospora has long been a fruitful model system for elucidating mechanisms of circadian rhythmicity common to many organisms , and therefore our results should encourage the search for circadian system components outside of the TTFL loops in other eukaryotic organisms .
All strains carried the ras-1bd ( formerly bd ) and csp-1 mutations , as previously described [15] , [29] . The frq10 mutation is a knockout created by targeted gene disruption [12] . The chol-1 mutation requires choline for normal growth as previously described [16] , [45] . Multiple mutant strains were created by standard crossing methods , and were backcrossed to the rasbd strain to verify genotypes by segregation of the expected phenotypes . Cultures were grown on maltose/arginine ( MA ) medium containing Vogel's salts , 0 . 5% maltose , 0 . 01% arginine , 2% agar , and either high choline ( 100 µM ) or no choline supplementation . On 100 µM choline , the chol-1 strains are not distinguishable from chol+ strains [39] . Cultures were initially grown in constant light on agar plates , and small plugs of mycelium were transferred to race tubes ( for rhythm assays ) or to cellophane-covered Petri plates ( for biochemical assays ) as previously described [21] , [29] before transfer to constant darkness ( DD ) at 22°C . The parent strain for mutagenesis was genotype csp-1; chol-1 rasbd; frq10 . Uninucleate microconidia were prepared by the method of Maheshwari [46] , [47] , using iodoacetate to induce microconidiation . Microconidia were suspended in water and exposed to ultraviolet light . Survival rates ( compared to untreated microconidia ) ranged between 10% to 80% in various experiments . Treated microconidia were plated on a sorbose-containing medium to induce colonial growth and individual colonies were tested for rhythmic phenotypes on race tubes . All strains carried the wild-type frq+ allele ( in addition to ras-1bd and csp-1 ) and either chol+ or chol-1 , and UV90+ or UV90 . The four genotypes used were: ( 1 ) csp-1; rasbd , ( 2 ) csp-1; rasbd; UV90 , ( 3 ) csp-1; chol-1 rasbd , and ( 4 ) csp-1; chol-1 rasbd; UV90 . Cultures were grown on race tubes containing MA medium with 100 µM choline . The period of the conidiation rhythm and the growth rate were assayed for each tube by linear regression as previously described [29] , [39] . At 19°C the UV90 strains produced only two or three conidiation bands before the rhythm damped out , therefore only the second period on each tube was used for all strains at this temperature . The number of tubes averaged for means ranged between 6 and 12 . Means were compared using the two-tailed Student's t-test for samples with equal variances . UV90 strains were compared to the corresponding UV90+ strains , i . e . , chol+ UV90 was compared to chol+ UV90+ , and chol-1 UV90 was compared to chol-1 UV90+ . All strains carried chol-1 ( in addition to ras-1bd and csp-1 ) and either frq+ or frq10 , and UV90+ or UV90 . The four genotypes used were: ( 1 ) csp-1; chol-1 rasbd , ( 2 ) csp-1; chol-1 rasbd; UV90 , ( 3 ) csp-1; chol-1 rasbd; frq10 , and ( 4 ) csp-1; chol-1 rasbd; frq10; UV90 . Cultures were grown in DD at 22°C on race tubes containing MA medium with 100 µM choline to repair the chol-1 defect . 2-hour heat pulses to 32°C were delivered as previously described [15] , [29] and density traces were collected and analysed as previously described [15] , [29] . Average density traces were calculated from the last two complete cycles in each experiment , averaging two cycles per tube and six tubes per set for N = 12 . Standard errors were calculated for each average pixel value . Average density profiles were normalized by setting the minimum and maximum pixel values to 0 and 1 , and confidence intervals were plotted as ± one S . E . M . All strains carried the wild-type alleles of chol+ and frq+ ( in addition to ras-1bd and csp-1 ) . The two genotypes used were: ( 1 ) csp-1; rasbd , and ( 2 ) csp-1; rasbd; UV90 . Cultures were grown in DD at 22°C on race tubes containing MA medium . For light resetting , groups of 4 tubes were exposed to a 2 min pulse of cool-white fluorescent light at 20–24 µmol/m2/s at 3 h intervals between 24 to 39 h in DD . For heat pulse resetting , groups were exposed to a 15 min pulse of 37°C at 3 h intervals between 22 to 40 h in DD . After growth had finished at 22°C , the positions of the bands that formed after the pulses were used to calculate the phase ( in circadian time ) of each race tube relative to the average phase of the un-pulsed control group . Circadian time ( CT ) was calculated using the periods determined from the un-pulsed controls , defining one period ( approx . 22 h ) as equal to 24 circadian hours . One circadian hour is therefore 1/24th of a period . CT 12 is defined as the light-to-dark transition . Phase shifts were calculated as the difference in circadian hours between the pulsed tubes and the average phase of the un-pulsed controls , and were plotted against the CT of the pulses to generate phase response curves . To generate phase transition curves , the new phases of the pulsed tubes ( in CT ) were plotted against the CT of the pulses , defined as the phases ( in CT ) of the controls ( “old phase” ) at the time of the pulses . The method was essentially as previously described [21] . Cultures of the rasbd; csp-1; chol-1 and rasbd; csp-1; chol-1 UV90 strains were grown at 22°C on top of cellophane on MA medium with 100 µM choline in 150 mm Petri plates . Cultures were initially grown in constant light , then transferred to DD , and harvested after 72 hours of total growth . Circadian phase ( reported as hours in darkness ) was varied by varying the time at which cultures were transferred from light to dark . Plates were transferred every three hours , and samples were collected from 3 to 72 h in DD . Cultures were harvested , protein was extracted , and FRQ was detected as previously described [21] . The primary antibodies were generously supplied by M . Merrow [48] and M . Brunner . Samples were divided into three sets for electrophoresis and blotting: 3–27 , 27–48 , and 48–72 h in DD . Both UV90+ and UV90 samples were run on the same gel to directly compare expression levels between strains . FRQ was normalized against total protein by staining blots with Coomassie Blue after immunodetection . To plot one complete time series from 3 to 72 h , normalized values were adjusted to equalize the repeated samples between sets ( 27 and 48 h ) . To calculate mean ratios , three time points were chosen near the peak values in each cycle . Three independent experiments were carried out at each time point , running UV90+ and UV90 samples on the same gel for direct comparison of FRQ levels . The relative level of FRQ in UV90 was divided by the relative level in UV90+ to calculate the ratio . A ratio of 1 . 00 indicates that FRQ levels are the same in the UV90+ and UV90 samples . The means and S . E . M . s of the three independent ratios were calculated . One-tailed one sample t-tests were used to test the null hypothesis that the means are not less than 1 . 00 , that is , that UV90 FRQ levels are not lower than UV90+ levels .
|
All eukaryotes ( including humans ) , and some bacteria , have evolved internal biological clocks that control activity and physiology in a daily ( circadian ) cycle . The molecular oscillators that drive these circadian rhythms are said to depend on rhythmic expression and feedback regulation of a small set of “clock genes . ” However , there is increasing evidence that there is more to the story than these well-studied feedback loops . In the fungus Neurospora crassa , rhythms can still be seen in mutants that are missing one of the clock genes , frq . There is currently a controversy as to whether there are many different frq-less oscillators and whether they interact with the frq clock . To identify the molecular mechanism that drives these frq-less rhythms , we started with a frq-less strain and mutagenized it to look for genes that affect the frq-less rhythms . We found a new mutation that not only disrupted two frq-less rhythms but also affected the rhythm when the frq gene is present . Our results suggest there is only one frq-less oscillator , and it interacts with the frq clock . Our new mutation may identify a gene that is critical to both oscillators . We suggest that a similar clock architecture may be common to all organisms .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"developmental",
"biology",
"model",
"organisms",
"genetic",
"mutation",
"neurospora",
"crassa",
"genetics",
"yeast",
"and",
"fungal",
"models",
"biology",
"microbial",
"growth",
"and",
"development",
"microbiology",
"genetics",
"and",
"genomics"
] |
2011
|
A New Mutation Affecting FRQ-Less Rhythms in the Circadian System of Neurospora crassa
|
The positions of nucleosomes in eukaryotic genomes determine which parts of the DNA sequence are readily accessible for regulatory proteins and which are not . Genome-wide maps of nucleosome positions have revealed a salient pattern around transcription start sites , involving a nucleosome-free region ( NFR ) flanked by a pronounced periodic pattern in the average nucleosome density . While the periodic pattern clearly reflects well-positioned nucleosomes , the positioning mechanism is less clear . A recent experimental study by Mavrich et al . argued that the pattern observed in Saccharomyces cerevisiae is qualitatively consistent with a “barrier nucleosome model , ” in which the oscillatory pattern is created by the statistical positioning mechanism of Kornberg and Stryer . On the other hand , there is clear evidence for intrinsic sequence preferences of nucleosomes , and it is unclear to what extent these sequence preferences affect the observed pattern . To test the barrier nucleosome model , we quantitatively analyze yeast nucleosome positioning data both up- and downstream from NFRs . Our analysis is based on the Tonks model of statistical physics which quantifies the interplay between the excluded-volume interaction of nucleosomes and their positional entropy . We find that although the typical patterns on the two sides of the NFR are different , they are both quantitatively described by the same physical model with the same parameters , but different boundary conditions . The inferred boundary conditions suggest that the first nucleosome downstream from the NFR ( the +1 nucleosome ) is typically directly positioned while the first nucleosome upstream is statistically positioned via a nucleosome-repelling DNA region . These boundary conditions , which can be locally encoded into the genome sequence , significantly shape the statistical distribution of nucleosomes over a range of up to ∼1 , 000 bp to each side .
The long DNA molecules of eukaryotic genomes are packaged into a compact structure with the help of histone proteins [1] . The fundamental unit of this structure , a nucleosome , comprises almost 150 base pairs ( bp ) of DNA wrapped around a histone octamer [2] , [3] . Individual nucleosomes are typically linked by 15–70 bp of free DNA into a “beads on a string” conformation , the primary and most stable structural level of chromatin . While packaging renders the genome compact , it also makes up to 80% of the DNA inaccessible for protein-binding at any given time [4] , potentially hindering the molecular processing of genetic information . In principle , accessibility might be attained dynamically , since mechanisms are known for spontaneous unwrapping [5] , [6] and diffusive sliding of nucleosomes [7] , as well as active remodeling [8] . However , numerous recent studies indicate that nature's solution to the accessibility issue is based , at least in part , on the widespread use of nucleosome positioning [4] , [9]–[14] . Nucleosome positioning essentially amounts to the opposite strategy of constraining the mobility of nucleosomes , rendering a selected set of DNA sites constantly accessible . Recent experiments measuring the distribution of nucleosomes across the genomes of several model organisms have robustly identified three salient features [11]: ( i ) A significant fraction of nucleosomes appears rather well positioned . In other words , the nucleosome positions determined from a large ensemble of cells do not average out to a constant density , but display many pronounced peaks . ( ii ) Typically , genes have a nucleosome-free region ( NFR ) upstream of their transcription start site ( TSS ) . That is , when genes are aligned at the TSS and with the direction of transcription to the right , the average nucleosome density exhibits a clear dip , about one nucleosome wide , to the left of the TSS . ( iii ) Downstream of the TSS , the gene-averaged nucleosome density displays strong oscillations , with an amplitude that decays with the distance from the TSS . Furthermore , biochemical experiments have firmly established that the DNA-binding affinity of histones depends on the DNA sequence , largely due to the intrinsic sequence-dependence in the biophysical properties of DNA , such as its bendedness and bendability [15] . Hence , a genomic free energy landscape for nucleosome positioning can be programmed into the genome sequence by appropriate placement of nucleosome attracting and repelling sequence motifs . Indeed , bioinformatic and biophysical approaches that parameterize sequence-encoded effects on nucleosome positioning have been remarkably successful in modeling and predicting the large-scale genomic nucleosome occupancy [16]–[20] , which has led to the notion of a genomic code for nucleosome positions [16] . Yet , the causes of the three above salient features are not yet disentangled . In particular , a recent study on nucleosome positioning in Saccharomyces cerevisiae [10] argued that the oscillatory pattern in the average nucleosome organization downstream of the TSS is qualitatively consistent with the statistical positioning mechanism proposed by Kornberg and Stryer [21] . With this mechanism , most nucleosomes are not individually positioned , but a non-random relative arrangement arises collectively , from statistical correlations induced by the interaction between neighboring nucleosomes . The phase of such a statistical arrangement relative to the DNA is determined by “barriers” on the genome , i . e . , local disturbances of the “nucleosome gas” . A disturbance is created regardless of whether the local effect on nucleosomes is attracting or repelling , e . g . , by sequences that attract or repell nucleosomes or by other bound proteins [22] , [23] . According to this scenario , termed the ‘barrier nucleosome model’ [10] , [24] , sequence-encoded positioning is required only for barrier creation , whereas nucleosomes adjacent to the barriers are positioned “for free” , i . e . , primarily via statistical correlations and with DNA sequence playing only a minor role . However , while the observed oscillatory pattern downstream of the TSS is reminiscent of the pattern calculated by Kornberg and Stryer [21] , there should be a similar pattern upstream of the TSS if statistical positioning is indeed the dominant force , since barriers act to both sides . Also , can the observed pattern be quantitatively explained by statistical positioning ? Finally , does the precise shape of the pattern permit conclusions on the nature of the barrier , e . g . , whether it is caused by an attractive or repelling effect on nucleosomes ? Here , we address these quantitative questions using the yeast data of Mavrich et al . [10] and a quantitative description of statistical positioning , which is essentially the same as in the work of Kornberg and Stryer [21] and equivalent to the ( much older ) “Tonks gas” model from statistical physics [17] , [25]–[28] .
Kornberg noted early on [24] that a nonrandom quasi-periodic nucleosome pattern arises already from the interplay of two basic biophysical constraints , ( i ) the constraint that the same DNA segment cannot simultaneously be incorporated into two nucleosomes , and ( ii ) the constraint that nucleosomes cannot form at ‘barrier’ genome locations , e . g . , those already occupied by other proteins such as sequence-specific transcription factors . The significance of the first constraint is that the exclusion between nucleosomes creates correlations in their statistical distribution along the DNA . Theoretically [26] , these correlations are revealed by a decaying oscillatory pattern in the two-particle distribution function , , which measures the probability to find , in the ensemble of all admissible nucleosome configurations , a nucleosome at a given location and another one at a distance from it . In other words , the knowledge of the position of one nucleosome leads to a partial knowledge of the nucleosome positions in the vicinity ( however , this two-particle distribution function is difficult to measure directly in experiments ) . The significance of the second constraint is that barriers in the “nucleosome gas” pin down the phase of the correlations , such that even the average nucleosome density displays a decaying oscillation as a function of the distance from the barrier [21] . Such barriers can be created by a variety of mechanisms; in particular , barriers can also be directly encoded in the DNA sequence , e . g . , via poly ( dA∶dT ) -tracts that are energetically unfavorable to incorporate into the nucleosome structure [29] . Similarly , “road block” nucleosomes that are particularly well-positioned will form a barrier for the surrounding nucleosomes . Here , we treat the average nucleosome density as a quantitative experimental feature that can be assayed for clues about the nature of these barriers and , more generally , about the extent to which statistical positioning is reflected in the nucleosome organization in vivo . This analysis must be based on a quantitative description of statistical positioning . In statistical physics , the interplay between interaction and entropy of particles in a one-dimensional configuration space has long been quantified in simple models for gas/liquid systems [25] , [26] , [30] . The classic quantitative study of statistical positioning , by Kornberg and Stryer [21] , is also consistent with this general framework . The simplest model is the ‘Tonks gas’ [25] where particles with a fixed size and a mean density interact only via hard-core repulsion that makes them impenetrable . For this model , the explicit analytical expression for the average particle density at a distance ( in bp ) from a perfect barrier is [26] ( 1 ) where denotes the Heaviside step function . This average density is related to the above-mentioned two-particle distribution function in an infinite system via . Eq . ( 1 ) produces the decaying oscillatory pattern that is characteristic for statistical positioning , see Fig . S1 for an illustration and ‘Materials and Methods’ for a self-contained derivation and a brief discussion of the physical mechanism underlying the density oscillation . The wavelength of the oscillatory pattern and the characteristic length over which its amplitude decays are both determined by the two physical parameters of the model , i . e . , the particle size and the average particle density . Note that the expression ( 1 ) holds only for a perfect barrier; more general situations will be considered below . As Eq . ( 1 ) describes a nontrivial effect that arises only from properties which the “nucleosome gas” shares with any other one-dimensional gas of impenetrable particles , it can be regarded as a ‘null model’ , i . e . , a quantitative reference that helps to identify relevant effects beyond the universal features for systems of this class . With this goal in mind , we wanted to compare Eq . ( 1 ) to patterns extracted from experiments . To extract the consensus distribution of nucleosomes around the NFR at the 5′ end of genes , previous studies aligned the genes at their TSS and averaged the nucleosome distributions over all genes [11] . This procedure is not suitable for our quantitative analysis , since the TSS cannot be mapped to a feature in the nucleosome gas . Instead , we used the positions of the NFR-flanking nucleosomes as reference points for our alignments , which permits a quantitative comparison of the averaged pattern with the nucleosome gas model ( see below ) . In addition to the appropriate choice of reference point for the alignment , it was important to process the experimental data in a way such that it became directly comparable to the physical density . Many studies determine nucleosome positions using a procedure of the following type [11]: First , the nucleosomal DNA is extracted from an ensemble of cells using micrococcal nuclease ( MNase ) . The genomic positions of these DNA fragments are then located using hybridization or sequencing approaches . Usually this raw data is further processed with hidden Markov models ( e . g . , [9] ) or peak detection algorithms ( e . g . , [31] ) , in order to infer the typical or putative nucleosome positions . These typical nucleosome positions are then used for subsequent analysis of nucleosome organization , including the consensus distribution around NFRs . However , such averages over typical nucleosome positions do not correspond to a physical observable . For qualitative analysis , the data processing algorithms are useful filters to enhance and highlight positioning effects . However , the use of a single , typical position for a nucleosome eliminates any cell-to-cell variation in the position . For our quantitative analysis , we had to use the undistorted raw data instead ( i . e . , the density of DNA reads along the genome for the sequencing approach ) , which is the best available experimental proxy to the physical density , see ‘Materials and Methods’ for details . Note that our observable , the nucleosome density , is distinct from the other frequently used observable , the nucleosome occupancy , which measures the probability to find a specified base pair covered by a nucleosome . Fig . 1 summarizes the nature of the data from a physics perspective . As illustrated in Fig . 1A , the extracted nucleosomal DNA originates from many cells with nucleosome positions that generally differ from cell to cell . The experimentally observed read density corresponds to the histogram shown in the bottom of Fig . 1A . This histogram would be directly comparable to the theoretical density for a nucleosome gas , if ( i ) the average over the different cells is equivalent to the thermal average , ( ii ) a DNA read identifies a nucleosome position uniquely and precisely , and ( iii ) the average number of reads per nucleosome is known and its fluctuations due to the random sampling are negligible . None of these conditions is entirely satisfied . Clearly , the relevant question ( discussed in ‘Materials and Methods’ ) is how much this affects the physical interpretation of the data . Since the average number of reads per nucleosome is in fact unknown , it is already clear that one cannot readily convert the read density to an absolute nucleosome density , i . e . , the experimental proxy to is not normalized . Fig . 1B illustrates the second averaging procedure , which is akin to a “disorder average” in statistical physics , in that it involves averaging over an ensemble of different systems rather than an ensemble of different states of the same system . Clearly , each gene is intrinsically different and could display a distinct pattern of nucleosome organization . However , as illustrated in the bottom of Fig . 1B , the common pattern that emerges by aligning the genes by the position of their +1 nucleosome ( the first downstream from the NFR ) exposes the generic features in a large set of genes . For individual genes , this pattern is obscured by the noise due to the limited statistics of the data . We performed our analysis on the data of Mavrich et al . [10] . The red dots in Fig . 2 display the average read density when the genes are aligned to the +1 nucleosome , with the direction of transcription from left to right . Our definition of the +1 nucleosome position is the most likely position of the first nucleosome downstream from the TSS based on the list of TSSs and nucleosomes by Mavrich et al . [10]; see ‘Materials and Methods’ for details . On a qualitative level , the pattern of Fig . 2 ( red dots ) closely resembles the consensus pattern from previous studies ( see , e . g . , Fig . 2 in Ref . [11] ) . In particular , both display the same salient features , i . e . , the pronounced downstream oscillations , the slow decay to a constant density , the nucleosome-free region , and the weak upstream oscillations . However , on a quantitative level , the patterns are significantly different , and only the pattern of Fig . 2 is suitable for quantitative comparison with a physical model . Our analysis leading to Fig . 2 did not include a correction for the known sequence bias of the MNase enzyme [32] , [33] . However , Fig . S2 compares the pattern of Fig . 2 with the result of an alternative analysis that also incorporates a correction for the MNase bias , and suggests that the MNase bias does not significantly affect the pattern; see ‘Materials and Methods’ for details . Another concern is that the entire set of genes contains a significant fraction where the gene ends within the 2000 bp downstream range plotted in Fig . 2 , see Fig . S3A . Therefore , we repeated our analysis on the subset of long genes with a size of more than 2000 bp in length . Fig . S3C shows that the resulting pattern is quantitatively very similar to that of Fig . 2 . Taken together , these results indicate that the pattern of Fig . 2 ( red dots ) represents a robust quantitative signature of the nucleosome organization near transcription start sites in yeast . To interpret the extracted pattern within the physical model described above , we performed a nonlinear least-squares fit to Eq . ( 1 ) , as described in ‘Materials and Methods’ . We kept the width of the nucleosomes fixed at the value bp suggested by the crystal structure [3] , and hence the only fit parameters were the mean nucleosome density and the global normalization factor for the data ( see above ) . The best fit is displayed as a gray line in Fig . 2A . To judge the quality of the agreement , it is useful to recall that the experimental pattern is basically described by five quantitative characteristics: the period of the oscillation , the length scale over which the oscillation decays , the asymptotic value of the density , and the amplitudes of the peaks and valleys in the density , above and below the asymptotic line . Given only two fitting parameters , the overall quantitative agreement between the physical model and the biological data is therefore remarkably good . Fig . S4 shows the corresponding fit to only the set of long genes , with a similar result . In both cases , the most apparent deviation between the model and the data is in the shapes and the amplitudes of the first two peaks , associated with the +1 and +2 nucleosome . We wanted to test whether this is solely a consequence of the fact that Eq . ( 1 ) assumes a perfectly positioned +1 nucleosome , while the data displays a small residual positional variability for the +1 nucleosome . We therefore convoluted the model density , Eq . ( 1 ) , with the shape of the +1 peak in the data ( see ‘Materials and Methods’ for details ) . The corresponding fit of this convoluted density to the data is shown in Fig . 2B ( gray line ) . By construction , the shape of the +1 peak now matches , but we note that the deviation in the +2 peak disappeared as well , suggesting that the finite positional variability of the +1 nucleosome is indeed sufficient to explain most of the deviation between the physical model and the biological data . Before discussing the obtained parameter values and the robustness of the fitting procedure , we address the immediate question that emerges from the above results: On the one hand , the agreement between model and data is consistent with the hypothesis that most of the nucleosomes downstream of the +1 nucleosome are statistically positioned . On the other hand , the statistical positioning mechanism has no intrinsic bias to a particular direction , i . e . , the pattern upstream of the NFR should be described as well by a viable physical model . However , the upstream consensus pattern reported in previous studies displays much less pronounced oscillations than on the downstream side [4] , [10] . To test whether this is simply a consequence of the gene-to-gene variation in the distance between the −1 nucleosome and the TSS , which should smear out the averaged pattern , we analyzed the statistical distribution of these distances and realigned all genes by the position of their −1 nucleosome . The −1 position is defined here by the first nucleosome upstream from the TSS , see ‘Materials and Methods’ . Fig . 3A displays the statistics of the nucleosome positions relative to the TSS , as derived from the nucleosome map of Ref . [10] . While +1 nucleosomes are restricted to a region about bp downstream from the TSS [31] , [34] , the −1 nucleosome position is considerably more disperse . Accordingly , the distance between the +1 and −1 nucleosomes , i . e . , the gap size , also has a wide distribution , see Fig . 3B . This distribution indeed smears out an oscillatory upstream pattern , which is uncovered by an alignment to the −1 nucleosome position that eliminates the gap size variation , see Fig . 4A ( blue dots ) . However , while this upstream pattern does display regular oscillations , the comparison to the superimposed downstream pattern from Fig . 2 demonstrates that these two patterns are significantly different . A possible concern with this upstream pattern is the frequent occurrence of another NFR closely upstream of the −1 nucleosome , either at the start of a divergently transcribed neighboring gene or at the 3′ end of a gene transcribed in the same direction ( 3′ NFRs are analyzed further below ) . To address this concern , we selected only the subset of genes with no gene start or end within 1000 bp upstream of the TSS and compared their averaged pattern to that for all genes . Fig . S3D shows that these two patterns are quantitatively very similar ( and clearly different from the downstream pattern ) , suggesting that the adjacent NFRs located at various distances have no significant effect on the average upstream pattern . The difference in the up- and downstream pattern might be an indication of positioning mechanisms beyond statistical positioning . Alternatively , this difference might be due to an intrinsic asymmetry of the NFRs , caused by different molecular determinants for the up- and downstream NFR boundary . Such an asymmetry would lead to a different boundary condition for the nucleosome gas on the two sides of the NFR . To illustrate the possible effect of the boundary condition on the pattern in the nucleosome gas , Fig . 4B shows the patterns for a range of boundary conditions together in a 3D plot . Here , the different boundary conditions are parameterized by an energy scale , , which measures the strength and the sign of the local effective free energy for nucleosome binding: Positive ( towards the front ) correspond to a nucleosome repellent region , i . e . , nucleosomes at positions to the left of the origin receive an energetic penalty . In contrast , negative ( towards the back ) correspond to an attractive positioning potential that is localized to a narrow region , the width of which is chosen here to roughly correspond to the finite peak width of the +1 nucleosome in the data . Note that all of the patterns contained in the 3D plot of Fig . 4B are qualitatively similar , irrespective of the value of . However , they are different on a quantitative level , and we next exploit this difference , using the experimental pattern as a quantitative signature , to infer the type of the boundary condition that is effectively implemented in vivo . In particular , it is instructive to contrast the case of a perfectly repulsive barrier ( ) with a perfect attractive positioning potential ( ) . Our above analysis of the downstream pattern in Fig . 2 was based on the latter case , i . e . , we assumed that most +1 nucleosomes are directly kept at particular positions on the genome through the action of specific molecular forces . We found that this assumption is compatible with the data . Given that the upstream pattern does not comply with this direct positioning scenario , we hypothesized that most −1 nucleosomes are instead indirectly ( statistically ) positioned by a repulsive barrier located at the upstream edge of the NFR . Fig . 4C displays the upstream pattern ( blue dots ) together with the model prediction assuming a perfectly repulsive barrier ( gray line ) . Note that this prediction is obtained with the same values for and normalization factor as inferred from the fit to the downstream pattern , i . e . , it does not involve parameter fitting , see ‘Materials and Methods’ for details . The agreement is surprisingly good , consistent with the interpretation that the positioning of most nucleosomes in the vicinity of the TSS is induced by a NFR that is intrinsically asymmetric: Our quantitative comparison suggests that the upstream boundary of the NFR is typically determined by repulsion rather than direct positioning of the −1 nucleosome . To put these observations on a systematic basis , we performed simultaneous fits on both sides of the TSS , for all combinations of boundary conditions and compared the results quantitatively on the basis of the mean square deviation per data point , see Fig . S5 , Table S3 , and ‘Materials and Methods’ . The results corroborate that the experimental pattern is best explained by the scenario where the +1 nucleosome is directly positioned whereas the −1 nucleosome is statistically positioned by a repellent region , as illustrated in Fig . 4D . The second best fit is obtained by the scenario where both the −1 and the +1 nucleosome are statistically positioned . As Fig . S5 shows , both patterns are quantitatively well explained with a single average nucleosome density bp for both up- and downstream of the TSS . Indeed , we find no clear evidence in the data that the average density of nucleosomes is different in intergenic and genic regions ( see ‘Materials and Methods’ ) , contrary to some observations made in other studies . We robustly obtained density values in the range of one nucleosome per 172 to 180 bp , described above and independent of the detailed choice of the fitting method . These values are slightly ( but consistently ) larger than the “nucleosome mode” of bp inferred by Mavrich et al . [10] by determining the typical peak to peak distance in the experimental pattern . Finally , it is interesting to note that NFRs have also been reported at the 3′ end of genes , although their biological significance is obscure [10] , [35] . In order to see to what extent our findings can be generalized to this class of NFRs , we also extracted the average up- and downstream pattern for 3′ NFRs by aligning to the respective flanking nucleosomes . Fig . S6 shows these patterns; see caption for details . We observe that on neither side the pattern displays the strong features associated with the direct positioning scenario . Instead , both 3′ patterns resemble the 5′ upstream pattern , which is superimposed for comparison in Fig . S6 . This suggests that the 3′ NFR is typically only a repulsive region , and hence less structured than the typical 5′ NFR .
The recent genome-scale identification of nucleosome positions revealed that a large fraction of nucleosomes are non-randomly positioned , that a large fraction of genes have a nucleosome-free region ( NFR ) at their promoters , and that the NFRs are flanked by salient oscillatory patterns in the nucleosome density [9] , [10] . Here , we performed a quantitative analysis of the average up- and downstream patterns , to reveal hidden information about factors that affect nucleosome positioning in promoter regions . To this end , we reanalyzed previously published yeast data [10] in a physical way . We found that the up- and downstream patterns differ significantly , but both are quantitatively consistent with a minimal model where nucleosome positioning is effected only from the location of the NFRs , but radiates over a range of up to bp to each side via the statistical positioning mechanism . Within this model , the difference in the average up- and downstream patterns is explained as an intrinsic asymmetry of the NFRs , which leads to different boundary conditions for the “nucleosome gas” on the two sides , see Fig . 4 . In contrast , we found no evidence of such an asymmetry for 3′ NFRs at the end of genes . That statistical positioning in the vicinity of barriers is a mechanism capable of producing a non-random nucleosome arrangement has long been established theoretically [21] and experimentally [22] . Statistical positioning of nucleosomes around promoter regions has been proposed several years ago [9] , while testing of this hypothesis has started only very recently [10] , [36]–[38] . The first study [10] presented qualitative evidence for statistical positioning , but was limited by its approach relying on consensus nucleosome positions and TSS alignments . However , that study also performed a thorough statistical analysis of the DNA sequence around promoters , and found that sequence elements known to be involved in nucleosome positioning ( dinucleotide patterns and poly ( dA∶dT ) stretches ) are concentrated to the NFR and the positions of the −1 and +1 nucleosomes , and are significantly less frequent up- and downstream from this region . This finding is consistent with our conclusions drawn from the quantitative analysis of the nucleosome patterns . Additionally , our analysis suggests that the sequence elements around the position of the −1 nucleosome are either not sufficiently widespread or not sufficiently effective to directly position the −1 nucleosome in the average pattern . This is not unlikely given that other mechanisms than direct sequence specificity are needed to obtain the precise positioning of the +1 nucleosome in vivo [39] . Two additional studies on statistical positioning in genic regions appeared after our work was completed [36] , [37] . These studies did not consider alignments to TSSs or +1 nucleosomes , but instead ranked genes by the distance between their first and last nucleosome , revealing a striking organization of the local minima in the nucleosome occupancy . This organization was found to be consistent with a Tonks gas that is constrained by repelling barriers from both sides . This analysis , with its focus on the genic regions and the positions of the minima , is complementary to ours , which focused on the quantitative shape of the average density , in particular also in the upstream intergenic region , and analyzed the difference between the up- and downstream pattern . Taken together , our and the existing studies of statistical positioning support the view that long-range correlations in nucleosome positions produced by localized features in the effective free energy landscape for nucleosome binding are an important determinant of the genome-wide nucleosome organization . Indeed , for yeast , where TSSs are typically spaced bp apart ( Fig . S3B ) , statistical positioning from features encoded only at the TSSs is sufficient to obtain non-random positioning for most nucleosomes . The physical origin of statistical positioning is an interplay between the mutual exclusion and the positional entropy of nucleosomes . While this mechanism does not “glue” nucleosomes to specific locations on the DNA , it does effect that , on average , nucleosomes favor certain positions over others . It can therefore make specific ( binding ) sites on the DNA more ( or less ) accessible for proteins . Moreover , it may also cause a bias for mutation processes , thereby creating a position-dependent mutation rate [40] and possibly long-range DNA sequence correlations . The approach taken in the present study may be classified as a “reverse approach” , which starts from the observed distribution of nucleosomes along the genome and ultimately seeks to determine from it the underlying free energy landscape for nucleosome binding ( see ‘Materials and Methods’ for a discussion of the assumptions leading to the concept of an effective free energy landscape ) . Here , this approach has led to the typical form of local features in the landscape that is depicted in Fig . 4D . Note that by construction , our approach has two important limitations: First , it cannot pinpoint the molecular mechanisms responsible for creating the features in the effective free energy landscape . For instance , our findings are equally compatible with sequence-determined depletion like in the HIS3-PET56 promoter [23] , chromatin remodeler induced nucleosome organization like in the POT1 promoter in its repressed state [41] , or with varying promoter architecture in response to transcriptional perturbation [35] , [42] . Disentangling the molecular mechanisms on a genomic scale , requires the use of the complementary “forward approaches” based on bioinformatic methods ( see , e . g . , [16] , [18]–[20] ) or biophysical modeling ( e . g . , [17] , [43] ) to predict nucleosome positions from sequence . Second , since reverse approaches rely on good statistics , our study is presently limited to the study of average patterns , obtained from a large number of different genes . Of course , many genes could have additional features in their free energy landscape at various positions . Again , these features could be directly encoded , by the intrinsic specificity of the DNA-histone interaction [15] , [16] , [44] , or in trans , via competition with other specific DNA-binding proteins , biochemical histone modifications [12] , [45] , or chromatin remodeling [8] . Such additional features do not necessarily affect the average pattern . However , our study firmly establishes the simple physical model of a Tonks gas with “programmable” boundary conditions as an excellent quantitative ‘null model’ for nucleosome positioning , which can be used as a reference point to identify specific positioning effects as deviations from it . Such a reverse approach on a gene-by-gene basis will likely be very fruitful once data with sufficient statistics and precision becomes available .
The data of Mavrich et al . [10] is the basis for our analysis . Mavrich et al . extracted nucleosomal DNA from yeast cells and sequenced the DNA stretches obtaining a list of reads which they aligned to the Saccharomyces cerevisiae genome . Nearly perfect alignments resulted in a list of reads with start and end coordinates on the Watson or Crick strand , which we obtained from the authors . Assuming a nucleosome width of bp we merged the information from both strands and assigned to each read the putative location of the midpoint of the original nucleosomal DNA sequence ( see “Supplementary Information” of Ref . [31] ) . Originally , some reads were aligned to multiple positions on the genome giving them an artificially high weight . Therefore , we counted the number of alignments for each read ( number of reads with same read identifier ) and weighted the reads by the reciprocal number of their occurrence . For example , if alignment to the yeast genome resulted in 5 hits , each alignment was weighted by a factor . The frequency of reads vs . location on the genome defines the read density map serving as our proxy for nucleosome density and is denoted by below . A small region of the read density map is sketched in Fig . 1A and Fig . S2A . Our list of start and end sites of genes is based on the list of transcribed regions and open reading frames as reported in “Supplementary Research Data” of Ref . [10] ( file Supplementary_Table_S2 . xls ) . We combined the start sites of transcribed regions ( class: pol II , subclass: mRNA ) and the end sites of open reading frames ( class pol II , subclass: ORF ) with same ‘feature ID’ to one ‘gene’ with a total of 4792 genes . See Figs . S3A and B for statistics of length of the genes and distances between TSSs . We used alignments of the read density map to the positions of nucleosomes surrounding the nucleosome free region ( NFR ) at the TSS for a quantitative test of statistical positioning . Since the read density map we used for our analysis does not allow direct annotation of individual nucleosomes , we had to employ the list of identified nucleosomes from the “Supplementary Research Data” of Ref . [10] ( file Supplementary_Table_S1 . xls ) . We used the definition of the +1 nucleosome as the first nucleosome at or downstream from the transcription start site ( TSS ) while the −1 nucleosome is defined as the first nucleosome upstream from the TSS . The probability distributions of the nucleosome's distance to the TSS are peaked at some distance from the TSS ( Fig . 3A ) such that a slightly different definition of the nucleosome has no significant effect on the results . Next , we aligned the read density map to the position of these nucleosomes and averaged ( Fig . 2 and Fig . 4A ) . To test the influence of gene starts and ends close to the nucleosomes of interest , we additionally created alignments using only genes larger than 2000 bp ( Fig . S3C ) and using those genes without gene start or end sites within 1000 bp upstream from the TSS ( Fig . S3D ) . The read density map we derived does not include any correction for sequence bias of micrococcal nuclease ( MNase ) . To test for the importance of such a correction , we performed an alternative analysis towards a nucleosome density around the nucleosomes . To that end , we exploited the list of nucleosomes as identified by Mavrich et al . : Based on the reads aligned to the yeast genome , these authors identified individual nucleosomes using a peak detection algorithm after correcting for MNase bias ( see “Supplementary Information” of Ref . [31] ) . The emerging list of nucleosomes also includes the standard deviation ( measure of fuzziness ) for each nucleosome ( “Supplementary Research Data” of Ref . [10] , file Supplementary_Table_S1 . xls ) . Interpreting the nucleosome's standard deviation as a cell to cell variation ( instead of an experimental error ) , we represented each nucleosome with assigned standard deviation larger than by a Gaussian distribution with standard deviation given by the nucleosome's standard deviation . This results in an alternative proxy for the nucleosome density as sketched in Fig . S2B . Both proxies for nucleosome density , the one based on the raw data ( reads ) and the one based on processed data ( individual nucleosomes ) , significantly differ locally ( compare Figs . S2A and S2B ) . The corresponding alignments to the +1 and −1 nucleosomes , however , are pratically identical ( Fig . S2C ) having accounted for genome-wide normalization ( 997655 reads correspond to 52918 nucleosomes ) . This indicates that MNase bias correction as performed by Mavrich et al . is not essential for our analysis . As a side-remark note that the proxy at first sight should represent nucleosome density without any further normalization . However , repeating parts of our fitting analysis ( see below ) with instead of revealed that a fit to the Tonks gas model is only possible if we allow for a normalization factor significantly different from unity ( ) , suggesting that the proxy underestimates the number of nucleosomes . A possible explanation is that up to 20 percent of the nucleosomes were missed by the peak detection filter applied by Mavrich et al . . This explanation appears likely , since not all of the yeast nucleosomes are well positioned , i . e . , a significant portion of nucleosomes will not lead to a clear peak in the distribution , given the average over many cells that is taken in the experiment . To systematically compare the quantitative model to the nucleosome alignments of the read density ( i . e . , our proxy for nucleosome density ) , we performed least squares fits using the function ( 2 ) where is a normalization factor , tests for a possible horizontal offset in the data , and the function from Eq . ( 1 ) contains the parameters and . In all our fits , the nucleosome width was kept fixed at bp . We used the offset parameter also to distinguish between the two opposite boundary conditions considered for our fits: As explained above , , corresponds to the direct positioning scenario where the first nucleosome is a fixed barrier for the neighboring nucleosomes , while a shift by one nucleosome width corresponds to the statistical positioning scenario where the boundary is not a nucleosome , but another repellent feature on the genome . ( In the latter case , the different genes should in principle be aligned to the location of the boundary , but since this is not possible , our alignment to the first nucleosome is the best alternative . ) For each of our fits , one of these two scenarios is imposed by choice of the starting value for , since each scenario corresponds to a deep “basin” in the least-squares score function . As can be seen from the Tables in the Supporting Material , each best-fit value for either clearly corresponds to the direct positioning scenario , , or to the indirect positioning scenario , . We performed fits to −1 nucleosome alignment data in the same way as for +1 nucleosome alignment data , except that we mirrored the data at the origin . For the fits , we used the data in a range from 200 to 2000 bp downstream from the +1 nucleosome and upstream from the −1 nucleosome , respectively . Altering the fitting range to 200–1200 bp had no significant effect on the results . To ensure best possible parameter estimates , we performed each fit 300 times from a wide range of starting parameters . Best fits are shown in Figs . 2A , S4 , and 3C ( where a peak at has been added where applicable to indicate the directly positioned nucleosome , i . e . , for the case ) . The corresponding parameter estimates are displayed in Table S1 where denotes the squared deviation per data point between data and model . The parameter estimates from fits to the +1 nucleosome alignment are robust against variations in details of the fitting procedure: ( i ) Fitting to the average over all genes yields almost the same parameter estimates as a fit to an average where only genes larger than 2000 bp are considered . For the latter , nucleosome density is estimated slightly higher due to the slightly further ranging oscillations ( Fig . S3C ) , but it does not significantly differ from the estimate obtained from the alignment including all genes ( see Figs . 2A and S4 , Table S1 ) . ( ii ) Randomly partitioning the set of 4792 genes over which the average is performed into four subsets and repeating the fitting analysis yielded almost identical results , see Table S2 . ( iii ) To account for the effect of the residual cell-to-cell variation in the position of the +1 nucleosome , we also performed a fit using a ‘convoluted Tonks model’ , where Eq . ( 2 ) ( with ) was convoluted with a probability distribution function corresponding to the experimental nucleosome density in the range of bp around zero . The first peak downstream from the +1 nucleosome , corresponding to the +2 nucleosome , is much better characterized by this fit ( compare Figs . 2A and B ) suggesting that cell-to-cell variation of the +1 nucleosome's position is reflected in cell-to-cell variations of the downstream nucleosomes . Yet , parameter estimates are very similar to those obtained from the fit without convolution ( Table S1 ) indicating that including this effect is not essential when fitting the Tonks gas model to the data in the range of 200 to 2000 bp as we do everywhere else in this study . For the fit to just the −1 nucleosome alignment , we used the parameter estimates for nucleosome density and normalization obtained from the fit to the +1 nucleosome alignment . Thus , the only remaining fit parameter here was the offset ( Table S1 ) , which was started at values . In order to systematically test alternative scenarios ( e . g . , direct positioning of the −1 nucleosome and indirect positioning of the +1 nucleosome ) , we performed simultaneous fits to both the +1 and −1 alignment data for each of the four possible boundary conditions . Fits were carried out analogously to the procedure described above , but with the and parameters constrained to take the same values on both sides . Fig . S5 displays the results , Table S3 shows the parameter estimates . Regarding the mean squared deviation per data point , scenarios C and D are similar , while scenarios A and B are less probable . In both eligible scenarios , the −1 nucleosome is indirectly positioned . In the best fit scenario C the +1 nucleosome is directly positioned . In our systematic fitting procedure described above , we assumed the same average nucleosome density up- and downstream from the NFR . This must be justified by comparing the average density in intergenic regions to that in genic regions . To estimate their ratio , we used the proxies for nucleosome density described above , i . e . , the read density ( ) and the representation of nucleosomes by Gaussians with appropriate width ( ) . To exclude the influence of the 5′ NFR , which is mostly located within intergenic regions , we excluded the NFR regions . Using proxy we obtained a ratio of 1 . 00 for the density in intergenic to the density in genic regions , whereas a ratio of 0 . 85 resulted from using . We conclude that there is no clear indication of a density bias between intergenic and genic regions ( apart from the existence of NFRs ) . We therefore assumed equal average density up- and downstream from the TSS for the fitting procedure .
|
Within the last five years , knowledge about nucleosome organization on the genome has grown dramatically . To a large extent , this has been achieved by an increasing number of experimental studies determining nucleosome positions at high resolution over entire genomes . Particular attention has been paid to promoter regions , where a canonical pattern has been established: a nucleosome-free region with pronounced adjacent oscillations in the nucleosome density . Here we tested to what extent this pattern may be quantitatively described by a minimal physical model , a one-dimensional gas of impenetrable particles , commonly referred to as the “Tonks gas . ” In this model , density oscillations occur close to a boundary at dense packing . Our systematic quantitative analysis reveals that , in an average over many promoters , a Tonks gas model can indeed account for the nucleosome organization to both sides of the nucleosome-free region , if one allows for different boundary conditions at the two edges . On the downstream side , a single nucleosome is typically directly positioned such that it forms an obstacle for the neighboring nucleosomes , while such a barrier nucleosome is typically missing on the upstream side .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"molecular",
"biology/chromatin",
"structure",
"genetics",
"and",
"genomics/bioinformatics",
"genetics",
"and",
"genomics/chromosome",
"biology"
] |
2010
|
Quantitative Test of the Barrier Nucleosome Model for Statistical Positioning of Nucleosomes Up- and Downstream of Transcription Start Sites
|
Reciprocal chromosomal translocations ( RCTs ) leading to the formation of fusion genes are important drivers of hematological cancers . Although the general requirements for breakage and fusion are fairly well understood , quantitative support for a general mechanism of RCT formation is still lacking . The aim of this paper is to analyze available high-throughput datasets with computational and robust statistical methods , in order to identify genomic hallmarks of translocation partner genes ( TPGs ) . Our results show that fusion genes are generally overexpressed due to increased promoter activity of 5′ TPGs and to more stable 3′-UTR regions of 3′ TPGs . Furthermore , expression profiling of 5′ TPGs and of interaction partners of 3′ TPGs indicates that these features can help to explain tissue specificity of hematological translocations . Analysis of protein domains retained in fusion proteins shows that the co-occurrence of specific domain combinations is non-random and that distinct functional classes of fusion proteins tend to be associated with different components of the gene fusion network . This indicates that the configuration of fusion proteins plays an important role in determining which 5′ and 3′ TPGs will combine in specific fusion genes . It is generally accepted that chromosomal proximity in the nucleus can explain the specific pairing of 5′ and 3′ TPGS and the recurrence of hematological translocations . Using recently available data for chromosomal contact probabilities ( Hi-C ) we show that TPGs are preferentially located in early replicated regions and occupy distinct clusters in the nucleus . However , our data suggest that , in general , nuclear position of TPGs in hematological cancers explains neither TPG pairing nor clinical frequency . Taken together , our results support a model in which genomic features related to regulation of expression and replication timing determine the set of candidate genes more likely to be translocated in hematological tissues , with functional constraints being responsible for specific gene combinations .
Chromosomal translocations are genomic rearrangements in which reciprocal exchange of genetic material between two non-homologous chromosomes results in the formation of novel fusion genes . Some of these fusion genes display oncogenic properties and have a strong impact on cancer progression and prognosis , particularly in hematological malignancies [1] . Therefore , several hundred translocations have been described in hematological cancers , although recent reports support their emerging role also in solid tumors [2] . However , whether the fusion gene resulting from a reciprocal chromosomal translocation ( RCT ) is a driver of tumor progression or just a passenger event is not yet fully understood in all cases [3] . Common and important features of RCTs were reviewed recently [4] , but to the best of our knowledge a thorough and extensive data-mining study has not yet been performed . A fairly comprehensive catalog of genes involved in RCTs is available in public manually-curated databases . Mitelman Database of Chromosome Aberrations and Gene Fusions in Cancer provides extensive documentation of clinical cases , making possible the estimation of clinical frequencies of translocations . TICdb [5] , on the other hand , provides manually curated mapped translocation breakpoints , allowing analysis of the sequences flanking those breakpoints . Several such analyses have been performed over the last years , highlighting the association of various sequence motifs with the presence of double-strand breaks ( DSB ) in some types of translocations . Current consensus about the general requirements for breakage and fusion is that an increased frequency of DSBs at particular genomic locations , together with close spatial proximity of certain loci [6] , [7] , determine the probability of some RCTs and why some combinations of translocation partner genes ( TPGs ) are more likely to occur ( for a review , see [8] ) . However , quantitative support for this as a general mechanism applicable to oncogenic RCTs in general is still lacking . The aim of this paper is to characterize the genomic hallmarks of TPGs using computational and statistical analyses of available clinical and high-throughput datasets . We focus on hematological malignancies , as they constitute the largest set of well-documented RCTs described to date . In this study we have made use of a large collection of data sources . We have used gene expression data for various tissues obtained from publicly available datasets , as well as multiple genomic features from the human genome assembly , to analyze TPG expression and regulation . For the analysis of fusion protein functions we have used InterPro domain features , gene ontology annotations and a recently available protein-protein interaction dataset . We have also used recent high-throughput data on spatial and temporal architecture of the nucleus to search for potential characteristic features of TPGs . While the mechanisms that could potentially lead to RCT formation are well-studied , it is now emerging that oncogene selection during tumorigenesis could be the decisive factor for RCT appearance in tumors [9] . Thus the logic of this study is to try to explain RCT occurrence from the point of view of functional selection . First , we compare various features of TPGs with non-translocated genes and find statistical support for some characteristic properties of TPGs related to expression levels and regulation of transcription . We also address the issue of non-random TPG pairing in RCTs , examining issues like nuclear distance , the types of protein domains retained in fusion proteins or replication time . We try to explain some open questions in the field , such as why RCTs are tissue-specific [10] and why they drive oncogenesis in specific lineages [11] . Finally we address the recurrence of RTCs and their clinical frequencies [1] . Taken together , our data , robustly supported by quantitative methods , provide new insights into the global features shared by TPGs and afford a unified explanation of the mechanisms responsible for the specificity and recurrence of RCTs .
Most hematological 5′ and 3′ TPGs are specifically involved in hematological translocations , with only a few of them translocated in other tissue types ( 10% and 4% for 5′ and 3′ TPGs respectively , according to TICdb ) . In an attempt to explain this fact , we first analyzed the expression of all TPGs in normal tissues . Expression of both 5′ and 3′ TPGs is significantly higher than average gene expression ( all RefSeq genes ) , but only 5′TPGs ( not 3′ TPGs ) are more highly expressed than hematopoietic-specific genes in hematological tissues ( see Figure 1A and supplementary Figure S1A ) . Analysis of gene expression in other tissue types revealed that 5′ TPGs , but not 3′ TPGs , are more highly expressed in tissues of hematopoietic origin than in tissues of epithelial or mesenchymal origin ( see Figure 1B and Figure S1B ) . This is true for around 65% of 5′ TPGs . Likewise , when we compared the expression of 5′ TPGs with their corresponding 3′ TPGs we found that 5′ TPGs generally have higher expression levels ( see Figure 1C ) . This was robustly replicated in data from various tissues of hematopoietic lineage and in different microarray datasets for almost 70% of translocations checked . Moreover , we studied lineage-specificity by analyzing expression of 5′ TPGs involved in translocations reported in lymphoid ( HEM-L , 177 translocations ) and in myeloid ( HEM-M , 201 translocations ) malignancies . Their expression levels in two cell lines of lymphoid and myeloid origin ( GM128 and K562 , respectively ) confirmed a lineage-specific pattern ( only 5′ TPGs in HEM-L show higher expression in GM128 compared to K562 ) , as shown in Figure 1D . In order to reveal possible mechanisms for such expression differences we analyzed a number of genomic features of TPGs using available tracks from UCSC Genome Browser . Analysis of promoter features in several cell lines suggests that the difference in expression of 5′ and 3′ TPGs could be explained by promoter activity . First , trimethylation of H3K4 , which is a marker of active promoters , is higher for 5′ TPGs than for 3′ TPGs in cell lines of hematopoietic origin ( Figure 2A ) . Likewise , we calculated the probability of observing Polymerase II ( Pol2 ) binding in the promoter region of TPGs ( −3 kb/+3 kb ) as the ratio of cell lines in which a Pol2 peak is found . Normalized Pol2 frequency ( ratios were normalized to zero mean and unit standard deviation for all RefSeq genes ) was significantly higher in cell lines of hematopoietic origin than in non-hematopoietic for 5′ TPGs , but not for 3′ TPGs . Moreover , Pol2 binding probability was higher in promoters of 5′ TPGs than in 3′ TPGs in hematopoietic cell lines ( Figure 2B ) . These findings suggest that tissue-specific expression of 5′ TPGs could be explained by more active transcription in hematopoietic cells , regardless of other issues such as mRNA stability . Since 3′ TPGs contribute their 3′-UTR to fusion transcripts , they could also have an impact on expression levels of fusion genes because 3′-UTRs are known to have a regulatory role in mRNA stability and half-life . For instance , it has been shown that genes more tightly regulated have longer 3′-UTRs with more regulatory elements in them [12] . We explored this by comparing several features of 3′-UTRs of TPGs . We observed that 3′ TPGs have significantly shorter 3′-UTRs than 5′ TPGs . Moreover , there were fewer conserved elements and microRNA target sites in 3′-UTRs of 3′ TPGs ( Figure 3 ) . MicroRNAs are known to be major players in post-transcriptional regulation , frequently playing the role of tumor-suppressors by inhibiting expression of proto-oncogenes [13] . All this indicates that 3′ TPGs might also play a role in changing the regulation of expression of oncogenic fusion genes . We have previously shown that the complex network of TPGs involved in RCTs can be decomposed into a simpler network of protein domain combinations [14] , indicating that selection for certain domain combinations is a powerful mechanism driving preferential pairing of TPGs . To further explore this issue , we analyzed the number of domains and protein interaction interfaces ( PIIs ) that are retained or lost in fusion proteins upon translocation ( see Methods for details about extraction of domains and PIIs ) . While proteins encoded by 3′ TPGs retain significantly more domains and PIIs than they lose , the opposite is true for proteins encoded by 5′ TPGs ( Figure 4 , panels A and B ) . In fact , quite a few proteins encoded by 5′ TPGs contained no domains at all , or contained domains that were difficult to associate with oncogenic functions , suggesting that the major role of 5′ TPGs is transcriptional up-regulation of fusion genes rather than contributing specific protein domains to fusion proteins . In order to define this more strictly , we classified protein domains according to their functional ontologies . The details of this manually curated classification procedure are described in Methods and the lists of annotated domains and their classes are given in supplementary Table S1 . Type 1 domain classes include DNA-binding ( D ) , protein interaction ( P ) , kinase ( K ) and histone modification domains ( H ) , whereas type 2 domains comprise other types of domains not involved in oncogenesis and domains with unknown function ( O ) , as well as protein parts without recognizable domains ( N ) . Total numbers of retained domains of each class are presented in Figure 4C , which shows that K and H domains are almost absent in 5′ TPGs , whereas D and P domains are the most frequent in both 5′ and 3′ TPGs . Likewise , Figure 5 shows that the occurrence of type 1 and 2 domains in 5′ TPGs and 3′ TPGs from the same translocation are strongly dependent , with significant under-representation of type 2-type 2 domain co-occurrences ( in agreement with their limited oncogenic potential ) . Using permutations with a Benjamini-Hochberg FDR cutoff of 10% we discovered 13 over- and 8 under-represented domain class pairs in hematological translocations ( Figure 5 ) . Most over-represented pairs of domain classes are in agreement with known mechanisms of oncogenicity of fusion proteins described in the literature ( see Discussion ) . To further explore the non-randomness of domain co-occurrence in translocations for more complex sets than domain pairs , we performed clusterization of fusion protein functional profiles , defined as binary strings where each bit indicates whether a TPG retained at least one domain of a certain class . Clusterization using an expectation-maximization algorithm with cross-validation yielded an estimate of n = 6 clusters , with four clusters ( C0 , C2 , C3 and C4 ) containing fusions with “regulation of transcription” signature and the other two ( C1 and C5 ) having a “kinase” signature . Notably , small clusters C2 and C5 include only true chimeric proteins in which both TPGs have similar functions . Figure 6A shows the resulting cluster profiles and counts . We then overlaid these clusters onto the network of fusion genes ( Figure 6B; for each translocation , edge color indicates the cluster to which it belongs ) . This shows that clusters tend to be restricted to non-overlapping sets of hubs and that events where the same gene belongs to translocations of both “regulation of transcription” and “kinase” clusters are extremely rare . To investigate the role of TPGs containing only type 2 domains we analyzed their predicted impact on the expression of their corresponding fusion genes . As shown above , 5′ TPGS are characterized by higher expression levels whereas 3′ TPGs are characterized by shorter 3′-UTRs . Thus , we computed the relative difference of expression between 5′ TPGs and 3′ TPGs from different translocations based on the presence of type 1 or type 2 domains in 5′ TPGs . Similarly , we calculated the decrease in length of the 3′-UTR of fusion proteins in translocations with 3′ TPGs containing or lacking type 1 domains . In both cases , we found evidence ( Figure 7 and Figure S2 ) that TPGs without type 1 domains can lead to overexpression of the fusion gene ( in the case of 5′ TPGS ) or improved stability of its mRNA ( for 3′ TPGs ) . As mentioned above , expression levels of 3′ TPGs could not explain why most of them are translocated exclusively in hematological tissues . To explore other possible ways in which 3′ TPGs could contribute to tissue-specificity , we identified all protein interaction interfaces ( PII ) in proteins encoded by 3′ TPGs , and extracted data about the proteins that interact with those PIIs . With this information we built a network of interactions between PIIs in 3′ TPGs and their interaction partners , shown in Figure 8A ( the network for 5′ TPGs is shown in supplementary Figure S3 ) . This network has a complex structure with many self-interactions and overlaps between interaction partners of different TPGs , many of which comprise large hubs . We then calculated the expression of these interaction partners in several tissues in the same expression datasets that we had used before . Interestingly , expression levels of interaction partners of 3′ TPGs were , on average , significantly higher in hematopoietic than in epithelial or mesenchymal tissues ( Figure 8B and supplementary Figure S4 ) . Likewise , 84% of 3′ TPG interaction partners are highly expressed in hematopoietic stem cells from bone marrow and 70% to 80% of them are highly expressed in various peripheral blood cells , according to annotations in DAVID database [15] ( GNF_U133A_QUARTILE used , these percentages are significant with Benjamini-corrected P-values less than 0 . 05 ) . Taken together , these results suggest that 3′ TPGs might influence tissue-specificity of translocations indirectly , via tissue-specific expression of the proteins interacting with them . Preferential fusion of certain 5′ and 3′ TPGs in hematological cancers is generally attributed to the spatial proximity of those loci within the nuclear space . This has been demonstrated for a few recurrent translocations [8] , but the relevance of spatial proximity as a general feature capable of driving specific translocations in hematological malignancies has not been assessed on a large scale . We therefore used high-throughput spatial proximity data ( represented as interchromosomal contact frequencies ) available for human lymphoblastoid cell line GM06990 [16] , [17] , in order to check whether known 5′-3′ TPG pairs are closer in the nucleus , on average , than random genes . As shown in Figure 9A , mean contact frequency calculated for 239 TPG pairs ( black line ) is significantly higher than for randomly selected Refseq genes ( light green ) , indicating a preferential position in a contact-enriched zone of the nucleus . However , neither replacing a 5′ or a 3′ TPG with a random Refseq gene ( red and dark green lines ) nor , more importantly , randomly permuting 5′-3′ TPG pairs ( blue line ) , significantly decreased the average distance between paired genes . This suggests that spatial proximity , in itself , does not have a major role in determining the preferential pairing of specific TPGs in hematological neoplasms . A more refined way to test the relevance of spatial proximity for lineage-specificity of translocations is to select translocations reported only in lymphoid or in myeloid malignancies ( HEM-L and HEM-M subsets mentioned above ) and to measure the distance between those loci in cell lines of lymphoid and myeloid lineage . Therefore , we compared distances between TPG pairs in GM06990 cells ( lymphoblastoid origin ) and in K562 cells ( myeloid origin ) , for which Hi-C data were also available . As shown in Figure 9B , we found no lineage-specific trends: TPG pairs from HEM-L and HEM-M translocations were at similar distances in GM06990 and K562 cells . Moreover , we observed very significant correlation of distances between cell lines ( see Figure 9B ) . To gain further insights into the relative positions of TPGs in the nucleus , we performed clusterization of Hi-C distances to obtain clusters corresponding to central and peripheral locations ( supplementary Figure S5 ) . We found that 5′ TPGs are significantly over-represented in the central ( proximal ) cluster of the nucleus ( Figure 9C , supplementary Figure S5C ) . Likewise , 5′ TPGs are , on average , closer to all other nuclear loci ( excluding loci on the same chromosome ) than RefSeq genes , according to their higher mean contact frequency ( supplementary Figure S5E ) . This , in contrast , was not observed for 3′ TPGs , indicating that 5′ TPGs occupy a more central position in the nuclear space whereas 3′ TPGs are more evenly distributed . The overall association of TPGs with regions in the central cluster is shown in supplementary Figure S5D . Replication timing ( RT ) is highly correlated , at the chromosomal level , with the organization of chromosomal domains within territories [18] . Furthermore , the RT of a chromosomal region is associated with mutation rate [19] and somatic copy number alterations are often bounded by early replicating regions [20] . We grouped smoothed RT data for GM06990 cell line into early and late replicating regions based on distribution of all RefSeq genes ( supplementary Figure S6 ) and analyzed RT values of TPG-containing regions . As shown in Figure 9C and supplementary Figure S6B , both 5′ TPGs and 3′ TPGs are significantly over-represented in early-replicating regions . Likewise , TPGs are replicated significantly earlier than all RefSeq genes based on raw RT values for their genomic regions ( supplementary Figure S6D ) , with 5′ and 3′ TPGs having similar timing . To further strengthen this observation , we repeated our analysis on more samples from the ReplicationDomain database . The results obtained for hematological samples were in good agreement with GM06990 cell line , while non-hematological samples only show slight trends ( supplementary Figure S6C ) . Another relevant question about the forces that drive specific combinations of TPGs in hematological malignancies is why some translocations are more frequently found in clinical samples whereas other translocations are never , or rarely , recurrent . It has been previously argued [7] that clinical frequency ( estimated from the number of reported cases in Mitelman database ) is correlated with spatial proximity for rearranged loci , leading the authors to suggest that high-order spatial genome organization , with non-random localization of these loci in relation to the center of nucleus , might increase their spatial proximity . Therefore , we extracted data for all translocations reported in hematological malignancies in Mitelman database and computed the contact frequency of various combinations of genes . First , we found that the average distance between TPGs in rare translocations ( reported only once ) is the same as the distance between TPGs in frequent translocations ( reported more than once ) ( Figure 10A ) . Furthermore , when we compared RT and contact frequency across the whole genome ( excluding their own chromosome ) for TPGs in rare and frequent translocations we did not observe any significant difference , as shown in supplementary Figure S7 ( which shows separately 5′ TPGs and 3′ TPGs ) . Again , this observation suggests that spatial proximity is not sufficient to account for the recurrence of specific translocations , and that additional factors must be considered . We therefore evaluated the role that other TPG properties might play in increasing the frequency of certain translocations . First , we found that expression levels of 5′ TPGs were significantly higher in recurrent translocations ( reported more than once ) than in rare ( Figure 10B ) . Similarly , we explored whether the number of known PIIs in 3′ TPGs might be associated with the frequency of their involvement in translocations . We found that 3′ TPGs reported more than once have significantly more PIIs than rare 3′ TPGs ( Figure 10C ) .
Chromosomal translocations frequently lead to cancer development via promoter substitution and loss of transcriptional control [21] . In agreement with this , we find that TPGs display significantly increased expression levels . In the case of 5′ TPGs , this seems to be the result of increased promoter activity , as shown by the presence of significantly more H3K4me3 marks and pol2 peaks near the TSS and evidence early replication timing . As regards 3′ TPGs , we found that they have shorter 3′-UTRs with less regulatory elements than corresponding 5′ TPGs , thus facilitating the escape of fusion genes from post-transcriptional control . In this regard , it has been previously reported that some chromosomal translocations lead to oncogenic transformation by disrupting microRNA-mediated gene repression [22] and that replacing the 3′-UTR of MLL with 3′-UTRs of its TPGs removes post-transcriptional inhibition of its expression [23] . Such loss of post-transcriptional control could contribute to the oncogenicity of MLL translocations . We propose that a similar mechanism could be important , more generally , to determine the oncogenic potential of translocations in hematological neoplasms . This confirms that genes involved in fusions are generally selected because they cause overexpression of oncogenes , as seen for many mutations [4] , [21] . It has also been shown that co-transcriptional processes can influence genome stability due to transcription-induced R-loops in which activation-induced-deaminases ( AID ) introduce DNA breaks [24] . Indeed , such a mechanism has been invoked to explain the presence of RCTs in B-cell malignancies [25] , [26] . More recently , analysis of the genomes of 21 breast cancer samples has revealed that regions of somatic hypermutation tend to associate with breakpoints of somatic rearrangements [27] , again suggesting that AID enzymes might play a role in co-transcriptional generation of DNA breaks . Thus , transcription might establish a link between gene involvement in RCTs and DSB frequency across the genome . This , however , would not explain why 3′ TPGs with moderate expression levels are preferred over other hematological tissue-specific genes . This could be explained by other mechanisms leading to the generation of DNA breaks which are not directly related to high transcription rates , such as replication . It is now clear that DNA breaks leading to some forms of chromosomal rearrangement can be induced during replication [28] . However , recent studies in yeast show that error-prone DNA due to template switching is associated with late-replicating regions [29] , while we observe that TPGs are over-represented in early-replicating regions . Given that early-replicating regions contain active genes with important cellular functions [30] , this is reminiscent of the master gene hypothesis according to which genes involved in RCTs are chosen because of their functional attributes [31] . In this regard , we have previously suggested that the need to combine certain protein domains imposes selective constraints on fusion proteins , so that those combinations with greater oncogenic potential are more frequently found in tumor cells [14] . Similarly , other authors have found that features such as intrinsic structural disorder of fusion proteins are important for their oncogenic properties [32] . However , the role played by functional selection is usually overlooked in models that try to explain specific combinations of TPGs in oncogenic translocations . Here , we have found that oncogenic translocations retain a significant percentage of domains and protein interaction interfaces of 3′ TPGs . We further used InterPro annotations to classify all protein domains into five broad classes , assuming that an oncogenic fusion protein should contain at least one domain with oncogenic properties . Such classification scheme is supported by the fact that fusion proteins in which both TPGs contain “type 2” domains ( that is , domains not belonging to any of the four major functional categories with oncogenic potential ) are significantly under-represented . Among over-represented domain combinations are the co-ocurrence of protein interaction domains ( many of which are capable of protein oligomerization ) with DNA-binding and kinase domains , as well as co-occurence of two kinase domains , which result in aberrant cellular signaling [33] . Likewise , our data show that co-occurrence of histone modification and protein interaction domains is non-random , which is consistent with the important role played by aberrant chromatin modification in cancer [34] , and in particular in translocations involving MLL [35] , [36] . Interestingly , most over-represented domain co-occurrences involve one type 1 and one type 2 domain , indicating that in most fusion genes one of the TPGs is simply promoting overexpression of the fusion protein . A 5′-TPG with type 2 domains could contribute a strong promoter , whereas a 3′-TPG with a type 2 domain might stabilize the fusion mRNA by contributing its 3′-UTR . Our analysis of expression gain of 5′ TPGs , as well as potential mRNA stability gain of 3′ TPGs , suggests that both mechanisms are generally operative . Finally , to obtain a global view of the relative contributions of the various domain features on non-random fusion partner selection , we performed an unsupervised classification of fusion proteins based on the functional signatures of their 5′ and 3′ TPGs . The superimposition of these functional clusters onto the fusion network reveals that functional profiles of fusion proteins robustly capture the basic structure of the network . Our data also suggest that tissue-specificity of translocations might be explained , at least in part , by tissue-specific expression of TPGs . First , we found that 5′ TPGs involved exclusively in hematological neoplasms are significantly more expressed in hematopoietic tissue compared to epithelial and mesenchymal tissues , which is somehow expected because the oncogenic potential of fusion genes relies on their expression levels in the tissue where the translocation takes place . But when we analyzed 5′ TPGs involved in translocations reported only in lymphoid or in myeloid neoplasms , comparing their expression levels in two cell lines of lymphoid and myeloid origin ( GM128 and K562 , respectively ) , our results confirmed that expression of 5′ TPGs is an important contributor to tissue specificity of RCTs . As for 3′ TPGs , we found that their contribution to tissue specificity is dependent on the expression levels of their interaction partners , which are significantly higher in hematopoietic than in non-hematopoietic tissues according to all gene expression datasets analyzed . It is interesting that some of the genomic features of TPGs that we have found correlated with tissue-specificity of translocations are known to be related to the spatial organization of the genome . Transcriptional activity , for instance , clearly affects the position of genes: inactive genes inside a chromosome territory are constrained in their mobility and thus in their potential to interact with distant loci [37] , so that the ability of actively transcribed regions to interact in trans requires that those regions loop outside of their respective territories [38] . Likewise , replication timing is highly correlated with nuclear position and chromosome architecture [39]–[41] . It is well documented that chromosomes occupy spatially defined territories in the nucleus [42] so that intra-chromosomal contacts are more frequent than inter-chromosomal [16] . This organization is tissue specific with some intra-chromosomal contacts being more frequent in some tissues [6] , [43] . Thus it has been speculated that the recurrence and tissue-specificity of translocations could be explained by the spatial proximity and physical contact frequency of translocated loci [6] , [7] . Our observation that 5′ TPGs are preferentially found in transcriptionally active domains in the center of the nucleus could provide a link between high-order genome organization and the potential occurrence of translocations , because a central position within the nuclear space ( which is known to be tightly linked with transcriptional activity ) increases average contact frequency with other loci [44] , [45] . However , our results suggest that spatial proximity per se is not a decisive factor in determining specific combinations of TPGs and their clinical recurrence . Although it has long been speculated that spatial proximity determines the specific pairing of TPGs in translocations [8] , data supporting this contention as a general mechanisms for all types of translocations is lacking . For instance , Parada et al . [6] used mouse chr5:chr6 and chr12:chr15 chromosome pairs to show that the frequency of translocations is correlated with the frequency of chromosomal contacts . These chromosome pairs provide a useful model , as chromosomes 5 and 6 are known to be translocated in hepatocytes but not in lymphocytes , while the opposite is true for chromosomes 12 and 15 . These authors showed that contacts between chr5:chr6 are significantly more frequent in hepatocytes , while chr12:chr15 contacts are significantly more frequent in lymphocytes . But recent Hi-C data provides a more global and accurate measure of contact frequencies between chromosomal loci in a human lymphoblastoid cell line [16] . Using the distribution of interchromosomal contact frequencies for all pairs of loci between two chromosomes as a measure of their closeness , we found that chr12:chr15 ( which are known to be translocated in human hematological malignancies ) are closer than chr5:chr6 ( not reported to be translocated ) . However , we also observed that chromosomes 5 and 2 , frequently rearranged in translocations involving the ALK gene in anaplastic large cell lymphoma cases [46] , were even farther apart in the nucleus than chromosomes 5 and 6 ( supplementary Figure S8A ) . When this analysis was extended to all translocations and all loci , we found no difference between the average distances for chromosome pairs that are rearranged in hematological malignancies versus non-rearranged pairs ( supplementary Figure S8B ) . Thus , while TPGs are non-randomly distributed relative to the center of the nucleus , we propose that their pairing in specific combinations is mainly driven by other factors such as gene activity , which ( through their association with high-order genomic organization ) lead indirectly to their nuclear closeness ( supplementary Figure S8C ) . In this view , spatial proximity is a necessary pre-requisite for the appearance of a translocation , but it is unlikely to be the only ( or even the most important ) factor to explain the specificity and recurrence of oncogenic translocations [47] . In fact , our data show that the specificity of TPG pairing and the recurrence of specific gene pairs are not directly dependent on spatial proximity . Recent findings in the budding yeast have shown that broken chromosomal ends created by DSBs are able to travel relatively long distances within the nuclear space to search for homologous templates . Miné-Hattab and Rothstein [48] have demonstrated that after induction of a DSB the broken chromosome explores up to 30% of the nuclear volume in diploid cells , about 10-fold larger than the volume to which chromosomes are constrained in the absence of breaks . A similar observation was made by Dion et al [49] in haploid cells , where DSBs stayed for hours searching for a homologous template . These results show that it is possible to join regions which are relatively distant in the nucleus , and are consistent with our contention that spatial proximity is not a strong determinant of translocation frequency or specificity in hematological neoplasms on a global scale . Therefore , the specificity of TPG pairing could be better explained by the preferential positioning of 5′ TPGs in the central zone of the nucleus , which in turn is related to functional features such as gene expression and replication timing . This central location might put these genes within the nuclear distance required to undergo a translocation with several potential 3′ TPGs . Then , selection for fusion proteins with oncogenic potential will dictate which specific gene pairs are eventually found in a particular tissue , and their relative frequency in patient samples . The issues discussed here also have important practical implications . Several studies published over the last few years have demonstrated the ability of next generation sequencing ( NGS ) to identify novel fusion transcripts in cell lines and in samples from patients with hematological and solid cancers [50]–[52] . However , the functional significance of newly identified fusions is not always clear because it is possible that many of these novel RCTs are the result of an increased background of genomic instability , rather than being driver oncogenic events [47] , [51] , [53] . Thus , there is a need for methods that identify which of these tumor specific RCTs are required for establishment and maintenance of the transformed phenotype [52] . The genomic features of 5′ and 3′ TPGs that we have identified in this work might help to develop computational approaches for the prediction of fusion genes that are more likely to have a causal role in the initiation or progression of hematological neoplasms .
Data from translocations , including the Ensembl transcript ID of translocation partner genes and the nucleotide position of breakpoints , were extracted from TICdb v3 . 1 ( http://www . unav . es/genetica/TICdb/ ) . There are some translocations involving the same gene pair with different breakpoint positions , which is redundant for some types of analysis . Therefore in some cases we have used unique TPG instances for expression studies , and unique TPG pairs for analysis of genome organization features . For analysis of the functions of fusion proteins we have used TPG pairs that were unique in terms of InterPro domain composition . We successfully extracted from TICdb data for 770 hematopoietic translocation entries involving 1175 TPGs , comprising 245 TPG pairs with unique gene names ( 117 unique 5′TPG and 161 unique 3′TPGs ) . Translocations and TPGs for which certain data were not available ( e . g . expression or genomic organization ) were not included in the corresponding analysis . For lineage-specific analysis , translocations reported in malignancies of lymphoid ( HEM-L , 177 translocations ) and myeloid ( HEM-M , 201 translocation ) origin were selected . Human gene expression data ( dataset #1 ) for three tissue lineages , epithelial ( EPI ) , hematopoietic ( HEM ) , or mesenchymal ( MES ) , was extracted from Gene Expression Omnibus ( GEO , www . ncbi . nlm . nih . gov/geo/ ) as follows . For each tissue lineage we manually selected four tissues as samples: colon mucosa ( GSE8671 ) , lung epithelium ( GSE30660 ) , mammary epithelium ( GSE25487 ) and pancreatic duct epithelium ( GSE19650 ) for EPI; bone marrow ( GSE32057 ) , peripheral blood mononuclear cells ( GSE11281 ) , spleen ( GSE25550 ) and CD3+ T-cells ( GSE6088 ) for HEM; omental adipose tissue ( GSE3526 ) , meniscal cartilage ( GSE19060 ) , skin fibroblasts ( GSE20538 ) and aortic vascular smooth muscle cells ( GSE11367 ) for MES . Gene expression levels for each of four tissue samples were calculated as the average of up to three randomly selected donor samples available in microarray data , using only data for non-malignant tissues . All microarray data were first downloaded as . CEL files and then jointly normalized by RMA express ( http://rmaexpress . bmbolstad . com/ ) . Analysis was replicated in an independent dataset ( dataset #2 ) using human gene expression atlas [54] as an alternative source of data . For that purpose the whole dataset was downloaded as a normalized matrix file from GEO and then manually grouped by tissue type . Overall microarray data from both the manual dataset and gene atlas were in good agreement and yielded similar results . Thus we present here only graphs generated using dataset #1 , providing graphs for dataset #2 as supplementary information . For analysis of lineage-specific expression patterns of 5′TPGs we used expression data in GM12878 and K562 cell lines from ENCODE project ( GSE26312 ) . All genomic data were extracted and processed using UCSC genome browser ( http://genome . ucsc . edu/ ) and Galaxy web service ( http://main . g2 . bx . psu . edu/ ) based on hg18 human genome assembly . H3K4Me3 ( ENCODE track ) and Pol II binding ( Yale TFBS track ) were used to characterize the promoters of TPGs in available cell lines , extracting −3 kb/+3 kb promoter regions of genes according to the annotated transcription start site ( TSS ) in RefSeq . Polymerase II ( Pol2 ) peak frequency was computed as the proportion of cell lines having a Pol2 peak in the promoter region . To perform an unbiased comparison , Pol2 peak frequency of all genes was normalized to zero mean and unit standard deviation , separately for hematopoietic and non-hematopoietic cell lines . Lengths of 3′UTRs were obtained from RefSeq and conserved elements were obtained from phastConsElements44way . Predicted microRNA target sites were queried using MirDIP web service ( http://ophid . utoronto . ca/mirDIP/ ) with ‘balanced precision’ option . Alternatively , phastConsElements44wayPrimates track and ‘4 of 12 databases’ query option were used for 3′-UTR analysis and yielded similar results ( data not shown ) . For analysis of putative functions of fusion proteins , retained InterPro domains and protein interaction interfaces were extracted from Ensembl Database ( http://www . ensembl . org ) via extensive usage of Ensembl Perl API , and from structurally resolved human interactome data [55] based on their positions in relation to the breakpoints ( according to TICdb ) . Briefly , we collected InterPro and protein interaction interface ( PII ) entries that were located entirely upstream or downstream the position of breakpoint in 5′ or 3′ TPGs , respectively . The list of interaction partners for a given TPG is comprised of all proteins that interact with those PIIs present in the corresponding part of the fusion protein . Annotation and gene ontology ( GO ) terms ( if available ) were extracted from InterPro using BioMart and from data provided in [56] , respectively . Domains were then manually classified into five broad functional categories based on available annotations: K ( kinase ) , H ( histone modification ) , D ( DNA binding ) , P ( protein interaction ) or O ( other/none ) . A complete list of protein domains classified according to this criterion is provided as supplementary Table S1 . As some InterPro domain features are small and are present in multiple copies in some proteins , this could cause biases when trying to analyze domain composition of fusion proteins . Therefore , instead of raw domain counts we used functional profiles , defined as binary strings indicating if a TPG has any domain of a given function . For domain co-occurrence analysis TPG pairs with unique genes and functional profiles were used . For analysis of relative position of TPGs in the nucleus we used normalized contact frequency ( Hi-C ) data from [16] , [17] . Clusterization of Hi-C distance data was performed as described in [17] , to obtain central and peripheral clusters . Additionally , we also analyzed the three clusters ( central , centromere-proximal and centromere-distal ) from the original paper . Replication timing ( RT ) data for the same cell line ( GM06990 ) and several additional samples were downloaded from Replication Domain web service ( http://www . replicationdomain . com/ ) and smoothed using LOESS . Normalized RT values for all RefSeq genes range from −1 . 5 to 1 . 5 and we have considered genes with RT>0 . 5 as early-replicating . Chromosomal positions of central and peripheral , early and late , and TPG-containing loci are presented in supplementary figures S5 and S6 . For analysis of lineage-specific differences in TPG distances we used Hi-C data for GM06990 and K562 cell lines [16] , [17] . Mitelman database ( http://cgap . nci . nih . gov/Chromosomes/Mitelman ) was used to provide estimates for clinical frequency of translocations . For this we counted the number of times that each translocation is reported in the database associated to a hematological malignancy . Data were not subject to any regression model , but to simple statistical testing: translocations were categorized as rare ( reported only once ) and frequent ( reported more than once ) . All permutation tests , Hi-C distance data clusterization and binomial tests were performed in Matlab . All other statistical tests were performed in GraphPad Prism . As all analyses were performed for large and heterogeneous sets of genes , non-parametric tests were used in most cases to ensure robust results and conclusions . Weka machine learning package ( http://www . cs . waikato . ac . nz/ml/weka/ ) was used to cluster translocations according to their functional profiles . All networks were visualized using Cytoscape software ( http://www . cytoscape . org/ ) .
A paper by Engretiz et al [57] was published after acceptance of this work , reporting a similar analysis to the one we have performed here . It arrives at quite different conclusions with regard to the importance of nuclear distance in establishing specific combinations of translocation partners . The new report used chromosomal bands involved in translocations from various types of cancer , whereas we focused on the actual genes involved specifically in hematological translocations . As the data and analysis methods are distinct , future studies may reconcile the two findings .
|
A common genetic lesion leading to hematological cancer is the creation of fusion genes as a result of reciprocal translocations between chromosomes . Such translocations are non-random , in the sense that certain genes are more likely to be fused than others , and they appear to be tissue-specific . Current models tend to explain the non-random nature of chromosomal translocations suggesting that chromosome breaks are favored at certain sites and that the distance between genes in the nucleus determines the probability of their being fused together . In this work we have analyzed several genomic features in a large collection of genes involved in chromosomal translocations in hematological cancers , using robust computational methods . Our findings suggest that nuclear distance is a general pre-requisite but does not determine the specific combinations of genes fused together . We find that genomic features related to transcription and replication , together with constraints derived from the functional domains present in the proteins encoded by fusion genes , better explain which genes participate in specific chromosomal translocations and the tissue types in which they are found . The association of such genomic features with the position occupied by genes in the nucleus explains the apparent causal role attributed to spatial position .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods",
"Note",
"added",
"in",
"proof"
] |
[
"hematologic",
"cancers",
"and",
"related",
"disorders",
"oncology",
"medicine",
"genomics",
"chromosome",
"biology",
"basic",
"cancer",
"research",
"genetics",
"genetics",
"and",
"genomics",
"biology",
"computational",
"biology",
"gene",
"networks",
"hematology"
] |
2012
|
Genomic Hallmarks of Genes Involved in Chromosomal Translocations in Hematological Cancer
|
Although autism has a clear genetic component , the high genetic heterogeneity of the disorder has been a challenge for the identification of causative genes . We used homozygosity analysis to identify probands from nonconsanguineous families that showed evidence of distant shared ancestry , suggesting potentially recessive mutations . Whole-exome sequencing of 16 probands revealed validated homozygous , potentially pathogenic recessive mutations that segregated perfectly with disease in 4/16 families . The candidate genes ( UBE3B , CLTCL1 , NCKAP5L , ZNF18 ) encode proteins involved in proteolysis , GTPase-mediated signaling , cytoskeletal organization , and other pathways . Furthermore , neuronal depolarization regulated the transcription of these genes , suggesting potential activity-dependent roles in neurons . We present a multidimensional strategy for filtering whole-exome sequence data to find candidate recessive mutations in autism , which may have broader applicability to other complex , heterogeneous disorders .
Autism is a neurodevelopmental disorder characterized by impaired communication skills , social behavior abnormalities , and stereotypies , with a prevalence of ∼1/150 children [1] . It is considered to be one of the most highly genetic neuropsychiatric disorders with a heritability of 40–80% [2] , [3] . Family studies show that siblings of autistic children are at a ∼25-fold higher risk to develop autism than the general population [4] , and twin studies show concordance of the autism phenotype in 20–30% of dizygotic twins and ∼60% of monozygotic twins [3] , [4] . Genome-wide linkage and association studies , and candidate gene approaches have identified several susceptibility loci and implicated potential autism genes [5]–[7] . The fact that no single genetic aberration accounts for more than 1% of cases suggests extreme genetic heterogeneity [8] , [9] , posing a major challenge to identifying causative genes . To date genes have been identified on the basis of overlap with other syndromic neurodevelopmental disorders ( e . g . Fragile X syndrome , Angelman syndrome , Rett syndrome ) , chromosomal abnormalities and copy number variation , and as causes for nonsyndromic autism ( e . g . NRXN1 , NLGN3/4X , SHANK3 ) [4] , [10] . In a few cases , autism has been shown to be caused by homozygous recessive mutations due to recent shared ancestry [11] , although the contribution of recessive mutations in outbred populations remains unexplored . Recessive mutations in autism may behave like other rare recessive traits , thus allowing gene mapping using homozygosity analysis . Homozygosity mapping is frequently employed to isolate disease genes in families where the parents are known to be definably related , typically as cousins , which increases the risk for recessive disease [12]–[14] . However homozygous recessive “founder” mutations are also common in patients whose parents share only distant ancestry , common ethnicity , or in some cases no apparent ancestry at all [15] , and population analysis of runs of homozygosity has been used to define genomic loci that may harbor such mutations in diseases characterized by genetic heterogeneity [16]–[18] . Here we surveyed the mutational spectrum in individuals with autism from nonconsanguineous populations who were selected for the high degree of homozygosity in the genome , since high levels of homozygosity suggest distant or cryptic shared ancestry of the parents . We identified several patients with potentially new autism mutations , and found that a surprising number of these mutations occurred in genes that are regulated by neuronal depolarization .
To sort the genetic heterogeneity of autism , we used homozygosity analysis [19] to identify a subset of patients likely to be enriched for recessive mutations . We performed a homozygosity-based analysis of 1000 families ( 5 , 431 individuals ) in the Autism Genetic Research Exchange ( AGRE ) [20] cohort . Though most American families in this cohort are of mixed European ancestry and share no acknowledged near ancestors , we hypothesized that a small proportion of European-American parents share a traceable common ancestor , or may share common ethnic ancestry through both parental lines , which in either case may result in homozygosity for rare recessive founder mutations , as has been demonstrated for a host of known Mendelian recessive diseases [21] . We identified a small subset of “outlier” AGRE families ( <2% of the total ) in which the affected children show runs of homozygosity totaling up to ∼9% of their genome . This low proportion of families with elevated homozygosity is consistent with low reported rates of consanguinity in the AGRE collection . Nonetheless , in the few outlier families , rates of homozygosity are far higher than generally observed in individuals whose parents have no common ancestry ( ≤1 . 6% ) , and overlap or exceed in some cases the predicted range of homozygosity expected in offspring of first cousin parents ( 6 . 25% ) [22] ( Figure 1A ) . The sizes of homozygous blocks in probands from these outlier families ranged from ∼5–19 cM on average ( Figure 1B ) , suggesting ancient shared ancestry in these families compared to larger blocks of homozygosity seen in consanguineous families ( ≥20 cM ) [22] . Since the AGRE dataset provides no specific information about shared ancestry or consanguinity between parents , we explored the level of shared ancestry between parents , by performing tests to estimate relatedness between individuals based on identical-by-state ( IBS ) and identical-by-descent ( IBD ) genotype information [23] , [24] . We find that for 16 families where probands had the largest amount of homozygosity in their genomes , some of the parental pairs were more closely related than average ( Figure S1 ) , but that parental relatedness by itself , as analyzed by these methods , did not always predict the degree of homozygosity in the offspring . We performed whole exome sequencing in 16 AGRE patients , selected because they showed the largest total proportion of their genome homozygous ( ∼1%–9% ) of all patients in the collection . We reasoned that some of the runs of homozygosity would contain homozygous causative mutations . Whole exome sequencing allows for the high-throughput , unbiased survey of all exonic variation in a patient , including any known mutations . Sequencing was performed using the Illumina Genome Analyzer II platform following enrichment of exonic sequences using Agilent's SureSelect Human Exome Kit . We obtained an average coverage of 92% at 20X ( Table S1 ) , and identified an average of 34 , 615 total variants per exome ( Table S2 ) , subsequently filtering them to identify rare , likely deleterious changes . Since we wanted to identify rare private mutations , common variants identified by the 1000 Genomes project and dbSNP130 were filtered out , and remaining variants were subject to an in-house bioinformatics pipeline to annotate variants that may disrupt gene function ( by altering the coding sequence , the splice sites , or truncating the protein ) . On average , 735 variants per exome were potentially pathogenic , and out of these , 39 per genome ( on average ) were homozygous ( Table S2 ) . The availability of whole exome sequence allowed us to test each patient systematically for mutations in known autism genes on the autosomes as well as the X chromosome , and no inherited mutations that were predicted to be damaging in well-documented autism genes were found in the 16 patients . To rule out variants that arose from spontaneous cell line artifacts , somatic mosaic mutations , or sequencing errors , we validated all homozygous variants in all family members using Sequenom technology . Genotyping candidate variants in the 16 probands allowed us to examine inheritance of variants as well as segregation with disease , since many families had multiple affected individuals as well as unaffected siblings ( Figure S2 ) . Variants that did not validate with Sequenom genotyping despite high sequencing depth ( ≥100 ) generally occurred in regions of the genome that were not uniquely mappable . For uniquely mapped variants , the rate of validation correlated well with sequencing depth ( Pearson's correlation = 0 . 532 , P = 0 . 001×10−30 , t-test ) ( Figure S3 ) . Analysis of segregation further permitted us to focus on bona fide inherited mutations as we only considered those variants that were homozygous in the proband ( by whole exome sequencing and Sequenom confirmation ) , heterozygous or absent in unaffected siblings , and transmitted from heterozygous parents . This validation step thus eliminates any possible sequencing errors or somatic mutations that complicate many high-throughput sequencing studies . We overlaid the validated variants with the result of our homozygosity analysis and further focused our attention on that subset of variants that fell within runs of homozygosity shared by affected siblings and absent from unaffected siblings . This allowed us to narrow down the number of candidate variants per exome , and for four families only 1 variant segregated with the disease ( Table 1 , Figure 1C ) . For some families our approach did not yield any candidate recessive variants as expected , since homozygous variants will not necessarily be causative even in some families selected based upon homozygosity . We then further examined the prevalence of candidate homozygous mutations in a control population of ∼700 normal individuals . We were able to exclude homozygous variants based on several criteria including: prevalence in controls , the genes not being expressed in brain , or the genes being mutated in other disorders ( Table S3 ) . Under this variant prioritization model ( Figure 2 ) , candidate autism mutations were identified in four of the 16 probands ( Table 2 , Figure 1C ) , with these candidate disease variants falling within runs of homozygosity shared by affected siblings and absent from unaffected siblings . The candidate mutations identified in this study implicate several candidate genes in autism that encode proteins involved in small GTPase mediated signal transduction , transcriptional regulation , and protein modification processes ( Table 2 ) . Among the mutations we identified is a homozygous c . 144 C>T change that creates an R40C mutation in ubiquitin protein ligase E3B ( UBE3B ) , a member of the E3 ubiquitin-conjugating enzyme family . UBE3B is highly expressed in the brain and may play a role in stress response [25] . The UBE3B R40C mutation identified in AU035204 is predicted to be damaging , was homozygous in both affected children ( monozygotic twins ) , heterozygous in the parents and unaffected sibling ( Figure S2 ) , and was absent in the homozygous state in 1344 control chromosomes . UBE3B is highly conserved across species and belongs to the same family as UBE3A , the gene disrupted in Angelman syndrome , a neurodevelopmental disorder characterized by intellectual disability , movement or balance problems , abnormal behaviors , and speech and language impairment . Recent work has shown that experience-driven neuronal activity induces Ube3a transcription and that it regulates excitatory synapse development and function through targeting the key synaptic molecules Arc and Ephexin5 [26] , [27] . We also narrowed down the candidate genes to only one in AU1328302 . An R125C mutation in CLTCL1 , encoding clathrin heavy chain-like 1 , was homozygous in both affected children , heterozygous in the parents and unaffected sibling , and predicted to be damaging ( Table 2 and Figure S2 ) . CLTCL1 is disrupted in a patient with features of DiGeorge syndrome , including intellectual disability , facial dysmorphia , long slender digits , and genital anomalies [28] . It encodes a member of the clathrin heavy chain family , representing a major structural component of coated pits and vesicles involved in intracellular trafficking , which are important to glutamate receptor turnover . Since resequencing of candidate genes in a larger cohort is an important validation step in evaluation of any candidate gene , we screened a larger independent cohort of whole exome data from 418 autism cases and 371 controls , sequenced as part of the ARRA Autism Sequencing Consortium . DNA from these cases and controls underwent whole exome capture , cloning and sequencing in the same fashion that our 16 cases did at the Broad Institute . For all four genes , we compared the rate of mutations under a recessive model , looking for either homozygous or compound heterozygous mutations in cases versus controls . As a group , the 4 genes showed a higher number of recessive mutations ( homozygous or compound heterozygous ) in cases ( 24/418 , 5 . 7% ) compared to controls ( 11/371 , 3 . 0% ) ( P = 0 . 042 , Fisher's exact test , one-tailed ) . These mutations were all missense changes and were relatively rare , all with allele frequencies of ≤5% ( Table 3 ) . One gene , CLTCL1 , especially stood out compared to the other four genes , having 17 mutations in cases versus 6 mutations in controls ( Table 3 ) . Genes with essential roles in synaptic plasticity have been implicated as an important cause of autism ( e . g . NRXN1 , NLGN3/4X , SHANK2/3 ) [29] , [30] , and since many synaptic plasticity genes are regulated by neuronal depolarization [11] , [31] , we tested the degree to which our autism candidate genes showed expression that could be modulated by neuronal activity . We depolarized mouse cortical neuron cultures and assayed changes in gene expression levels . We found that four out of four of the mouse homologs of our candidate genes are upregulated in response to neuronal activity ( UBE3B , CLTCL1/Cltc , NCKAP5L , and ZNF18/Zkscan6 ) ( Figure 3 ) . This is particularly interesting because in general only about 1000 transcripts , or about 3% of the transcriptome , manifest such depolarization-regulated gene transcription [32] . The upregulation of Ube3b in response to depolarization resembles the activity-dependent transcription of its close paralog Ube3a , which has well-documented roles in synaptic plasticity [26] , [27] . The regulation of expression of several potential recessive autism genes by neuronal depolarization suggests that the candidate genes are likely to be involved in neuronal function and/or development , and mutations in these genes might lead to nervous system dysfunction in the context of autism spectrum disorders ( ASDs ) . In the 12/16 patients for whom we did not identify homozygous candidate mutations , we examined the mutational spectrum under different models of inheritance . Out of an average of 696 rare , heterozygous , and potentially deleterious variants per exome , we identified 67 candidate compound heterozygous changes ( at least two deleterious variants in the same gene ) . Sequenom genotyping validated an average of 27 of these variants , and phasing of the resulting set in trios revealed ∼4 true compound heterozygotes with one allele inherited from each parent . Genotyping of unaffected siblings when available reduced this number to ∼2 variants per individual consistent with fully penetrant , recessive disease ( Table S4 ) . For three patients , we narrowed down the candidates to 1 gene and for 8 patients there were no candidate genes with compound heterozygous variants ( Table S5 ) . Analysis of X-linked mutations did not identify mutations in well-validated X-linked autism genes , though 11/14 male patients carried rare hemizygous X-linked variants , three of which occurred in genes associated with intellectual disability ( ARHGEF6 , AFF2 , and OCRL ) . The first variant in ARHGEF6 , which encodes Rac/Cdc42 guanine nucleotide exchange factor 6 , results in an I444N mutation . The second variant in AFF2 , encoding Fragile X mental retardation 2 , causes a P847A mutation that is predicted to be benign by PolyPhen-2 . The third variant disrupts a splice donor site in OCRL ( oculocerebrorenal syndrome of Lowe gene ) ( Table S6 ) . Splicing mutations in OCRL have been identified in patients with Lowe oculocerebrorenal syndrome [33]–[36] , characterized by hydrophthalmia , cataract , intellectual disability , vitamin D-resistant rickets , amino aciduria , and reduced ammonia production by the kidney . Since patient AU1019301 is not known to exhibit a renal phenotype or any other Lowe syndrome phenotypes , it is unlikely that this mutation is causative of the neurological condition of the patient . Segregation analysis showed that these three X-linked mutations were inherited from heterozygous mothers , confirming that they are not cell line artifacts . Since our study design enriched for families with potential shared inheritance , it does not permit confident determination of the causative nature of these potential compound heterozygous or X-linked mutations , which could only be tested by analysis of additional cases . Our results illustrate both the challenges and the potential of whole exome sequencing in an extremely genetically heterogeneous condition such as autism . Each exome contains large numbers of variants that initially challenge analysis . We present a systematic method to approach whole exome data , by filtering for variants compatible with identity by descent , surveying prevalence in controls , segregation analysis , and incorporating functional information ( Figure 2 ) . Almost all instances in which new genetic syndromes have been identified using whole exome or whole genome sequencing have involved families with recessive disorders generally ( Miller syndrome ) [37] , [38] and/or shared parental ancestry specifically ( WDR62-associated cortical malformations ) [39] , because the analysis of homozygous mutations provides tremendous power to improve “signal to noise” caused by sequencing errors , spontaneous cell line mutations , somatic mutations , etc . Hence , tracing ancestry may be an important tool to define genetic causes in a subset of autism patients . Our study further emphasizes the power of whole exome and whole genome approaches in allowing a complete survey of all potential mutations in the patient genome , and the systematic screening of all major modes of inheritance . Recent studies have confirmed the contribution of de novo point mutations ( 5–20% of cases ) [40] and de novo copy number variants ( 5–10% of cases ) [41] to autism . Our data suggest a potentially important role for recessive mutations in autism . Though our pre-selection of 16 patients for whole exome sequencing , and our limited analysis of whole exome data from >400 cases in the ARRA Autism Sequencing Consortium , does not allow us to calculate the proportion of cases likely attributable to recessive as opposed to other causes ( e . g . de novo , X-linked ) , our data do suggest that a systematic analysis of recessive causes of autism would be worthwhile . Homozygous null mutations appear to be exceedingly rare in autism , while homozygous missense changes were found in several candidate genes ( Table 2 ) , consistent with the possibility that some cases of ASD may reflect hypomorphic mutations in genes that have more severe phenotypes when completely disabled [11] . On the other hand , compound heterozygous recessive mutations could be more common in the outbred families represented by the AGRE . Furthermore , we find that different patients showed candidate mutations in different ASD candidate genes , confirming that recessive autism genes are likely to be highly heterogeneous . On the other hand , several of the genes we identified represent new neuronal depolarization-dependent genes , further supporting a role of defective synaptic transmission and neuronal plasticity in the pathogenesis of ASD . Finally , the approach employed here might be of value to the dissection of other complex traits where extreme genetic heterogeneity is suspected or confirmed . Since many neuropsychiatric conditions - including schizophrenia , intellectual disability , and epilepsy - often ( albeit not exclusively ) arise from loss of gene function , it is reasonable to suppose that recessive loss of gene function may play detectable roles in other conditions . Despite the rich variation in the human exome , our study design and approach to variant prioritization allowed identification of candidate autism genes from a relatively small sample .
Whole exome sequencing was performed on DNA samples from the AGRE collection available at the Broad Institute . All human studies were reviewed and approved by the institutional review board of the Children's Hospital Boston , the Broad Institute , Cambridge , and the local institutions . The analysis was performed using the Illumina 550 SNP genotype data for 1000 families from the AGRE collection . The data was obtained with permission from the AGRE [42] . Runs of homozygosity were calculated using custom scripts , allowing for no more than 2 consecutive heterozygous SNPs in a run and 3 heterozygous calls in every 10 consecutive SNPs . Intervals homozygous for the same haplotype and shared by all affected individuals were used to narrow the locus in each family . We used PLINK [23] to calculate the probability that one allele is shared IBD ( Z1 ) , and we calculated IBS2*_ratio and the percent of informative SNPs as described by Stevens et al . [24] . Briefly , IBS2*_ratio is equal to ( IBS2* ) / ( IBS2*+IBS0 ) , and the percent of informative SNPs is equal to ( IBS0+IBS2* ) / ( IBS0+IBS1+IBS2 ) , where IBS0 is the total number of observations in which two discordant homozygotes are present , and IBS2* results when two concordant heterozygotes are compared between any pair of individuals . Exome enrichment was performed on 3 µg of genomic DNA , using the SureSelect Human Exome Kit ( Agilent Technologies , Inc . , Santa Clara , CA ) , according to the manufacturer's protocol . The kit covers exonic sequences of ∼18 , 500 genes and a total of ∼33 Mb of target territory . The captured , purified and amplified library targeting the exome from each patient was sequenced on the Illumina GA II . Paired-end sequences were obtained at a read length of 72 bp . High-throughput sequence analysis was performed according to a customized bioinformatic pipeline for tracking sequence data , aligning reads , calculating coverage , calling variants , annotating variants with respect to functional effect , filtering out benign variation and flagging candidate rare , pathogenic mutations . Briefly , BWA version 0 . 5 . 7 ( ref . 3 ) was employed to align reads to the human genome ( reference build hg18 ) . Consensus and variant base calls were made with SAMtools version 0 . 1 . 7 ( pileup ) , filtered for quality ( mapping quality >10 for insertions and deletions , and >25 for SNPs ) , and loaded into a MySQL database for storage and further processing , including annotation of the predicted consequences ( noncoding , coding synonymous , coding nonsynonymous or frameshift , splice site ) of each variant using GMCC [43] ( Genomic mutation consequence calculator ) . Candidate mutations were identified by starting with a list of all variants , removing those present either in dbSNP130 or the 1000 Genomes Project database , and selecting for coding nonsynonymous , frameshift or splice site changes . Sequence data were visualized using either the UCSC Genome Browser or the Broad Institute Integrated Genome Viewer . All genomic base positions are presented in reference to the human genome NCBI build 36 ( hg18 ) . The functional effect of the mutation on the protein was assessed using PolyPhen-2 [44] . Sequenom genotyping of variants in the probands and their family members was performed on the iPLEX Gold platform at the Broad Institute . Variants were genotyped in control individuals also using the Sequenom iPLEX Gold assay at the Molecular Genetics Core Facility at Children's Hospital Boston . The controls collection consisted of 704 neurologically normal samples obtained from the Coriell Cell Repositories ( Camden , NJ; 584 Caucasian samples ) , or available in our lab ( 80 Saudi and 40 Bedouin samples ) . We screened whole exome sequencing data from a total of 789 exomes ( 418 autism cases and 371 controls ) that were sequenced at the Broad Institute ( as described above ) as part of a case-control study by the ARRA Autism Sequencing Consortium . Recessive mutations ( homozygous and compound heterozygous ) were counted in cases and in controls and a Fisher's exact test was used to determine whether the number of mutations in cases was significantly different than the number in controls . Samples in this study are of European ancestry from the AGRE collection , the Autism Sequencing Consortium ( ASC ) , and the National Institute of Mental Health ( NIMH ) . E16 . 5 C57B6 mouse embryo cortices were dissected and then dissociated in 1× Hank's Balanced Salt Solution ( HBSS ) , 20 mg/ml trypsin ( Worthington Biochemicals , Lakewood , NJ ) , and 0 . 32 mg/ml L-cysteine ( Sigma , St . Louis , MO ) for 10 minutes . Trypsin treatment was terminated with three two-minute washes in 1× HBSS with 10 mg/ml trypsin inhibitor ( Sigma , St . Louis , MO ) . Trituration of cells was performed with a flame-narrowed Pasteur pipette to fully dissociate cells . Neurons were seeded at an approximate density of 1×106/well on 6-well culture plates . The dishes were pre-coated overnight with poly-ornithine ( 30 µg/mL , Sigma ) in water , washed three times with water , and washed once with Neurobasal Medium ( Life Technologies , Carlsbad , CA ) before use . Neurons were maintained in 2 ml/well Neurobasal Medium containing B27 Supplement ( 2%; Invitrogen , Carlsbad , CA ) , penicillin-streptomycin ( 50 µg/ml penicillin , 50 U/ml streptomycin , Sigma ) and glutamine ( 1 mM , Sigma , St . Louis , MO ) . Neurons were grown in vitro for 7 days . 8 ml of the medium was replaced with 10 ml fresh warm medium on the 4th and 6th days in vitro ( DIV ) . For KCl depolarization of neurons , DIV 6 neurons were quieted overnight in 1 µM TTX and 100 µM APV , and they were incubated for 0 or 6 hours in 55 mM KCl . Total RNA was isolated from cultures using 1 ml Trizol/well according to the manufacturer's instructions ( Invitrogen , Carlsbad , CA ) . Isolated RNA was treated with DNAseI Amplification Grade ( Invitrogen , Carlsbad , CA ) and cDNA library was synthesized by cDNA High Capacity cDNA Reverse Transcription Kit ( Applied Biosystems , Carlsbad , CA ) . The cDNA was the source of input for quantitative PCR , using a Step One Plus Real-Time PCR Instrument and SYBR Green reagents ( Applied Biosystems , Carlsbad , CA ) . The relative expression plot was constructed using concentration values that were normalized to corresponding tubulin concentrations . The whole exome sequence data is available online ( The National Database for Autism Research ( NDAR ) Collection ID: NDARCOL0001918 ) .
|
Autism spectrum disorders are neurodevelopmental disorders that are genetically highly heterogeneous , with no single gene accounting for more than 1% of cases . In order to identify recessive mutations , we selected probands from an outbred population based on abundance of homozygosity in their genomes . We interrogated the entire coding sequences of 16 probands that had evidence of parental shared ancestry and identified four candidate autism genes . Furthermore , the expression of these genes was responsive to neuronal activity . We present a strategy for identifying candidate recessive mutations in genetically complex disorders .
|
[
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] |
[
"genome",
"sequencing",
"genetics",
"biology",
"genomics",
"genetics",
"and",
"genomics"
] |
2012
|
Whole-Exome Sequencing and Homozygosity Analysis Implicate Depolarization-Regulated Neuronal Genes in Autism
|
Despite significant , steady progress in schistosomiasis control in the People's Republic of China over the past 50 years , available data suggest that the disease has re-emerged with several outbreaks of acute infections in the early new century . In response , a new integrated strategy was introduced . This retrospective study was conducted between Jan 2005 and Dec 2012 , to explore the effectiveness of a new integrated control strategy that was implemented by the national control program since 2004 . A total of 1 , 047 acute cases were recorded between 2005 and 2012 , with an annual reduction in prevalence of 97 . 7% . The proportion of imported cases of schistosomiasis was higher in 2011 and 2012 . Nine clusters of acute infections were detected by spatio-temporal analysis between June and November , indicating that the high risk areas located in the lake and marshland regions . This study shows that the new integrated strategy has played a key role in reducing the morbidity of schistosomiasis in the People's Republic of China .
Schistosomiasis japonica , caused by Schistosoma japonicum , is a serious parasitic zoonosis threatening millions of people in the Southeast Asia , including the Peoples' Republic of China ( P . R . China ) [1] . Despite the great achievements made during the last six decades in the control of the disease [2] , it remains a public health concern in P . R . China [3] . Changes in the ecosystem due to environmental degradation and infrastructural development contributed to the resurgence of schistosomiasis in the early 21st century . The National Schistosomiasis Control Program was established already in the early 1950s [3] and the disease was recently given top priority together with HIV/AIDS and tuberculosis [4] . The development of an effective schistosomiasis control program gained in importance and the State Council issued the national medium and long-term strategic work plan in 2004 for schistosomiasis control to be achieved by 2015 . The current strategy is focused on control of infection sources , a shift from the earlier strategy of morbidity control [5] , [6] , in order to strengthen the implementation of integrated measures aiming to reduce the transmission of S . japonicum from cattle and humans to snails [7] . The control strategy includes the following main interventions: ( i ) replacement of bovines and most water buffalo by tractors for agricultural activities [8]; ( ii ) rearing livestock in pens and forbidding them to pasture in marshlands where snail habitats exists [9]; ( iii ) recycling excreta from humans and domestic animals to produce methane for cooking [10]; ( iv ) requiring fishermen to use containers to prevent excreta from being released into the water in Poyang Lake and Yangtze River area [11] , ( v ) improving the environment in high risk areas [12]; and ( vi ) implementing other routine health-control measures , such as snail survey and elimination , regular surveys and treatments , and health education [13] , [14] . This work plan has been boosted by joint efforts from both the central and local governments to produce an effective control program . By the end of 2008 the medium term goal was achieved , i . e . , infection controlled with a prevalence rate less than 5% without any outbreak of acute schistosomiasis in all endemic areas , and achievement of transmission control with a prevalence rate less than 1% both in humans and reservoir hosts in the mountainous areas ( Sichuan and Yunnan provinces ) [15] . In addition , Jiangsu province , located in the lower part of the Yangtze River , attained the status of transmission control in 2010 [11] , [16] . Of the 454 counties earlier reported endemic for acute schistosomiasis , 274 ( 60% ) attained the status of transmission interruption , and 103 ( 23% ) achieve transmission control with the remaining 77 counties ( 17% ) for reaching infection control by the end of 2011 [17] ( Figure 1 ) . Due to the complex nature of schistosomiasis and its sometimes irregular distribution , the disease still remains an important public health concern in P . R . China [15] , [18] . Despite strong progress , control activities remain arduous with efforts on the schistosomiasis control program continuing to ensure total elimination of transmission [11] , [19] . It is now essential to further identify the critical epidemiological factors , monitor potential risk areas and provide technical guidance for surveillance and response during the continuing shift in strategy from morbidity control to focus on transmission . This change requires close oversight of the infection status , not only of humans but also of livestock and the intermediate Oncomelania snail host [20] . Since acute schistosomiasis is a highly sensitive indicator in the process of monitoring and evaluating the progress of Schistosomiasis control [4] , it should be the first step in the ongoing major transformation of the control program [21] . However , we are yet to understand fully the role of this indicator in strategy shift from control to elimination . This retrospective study seeks to explore the effectiveness of a new integrated control strategy that was implemented through the national control program , based on reported cases from 2005 to 2012 in P . R . China .
Data of cases and outbreaks were extracted from web-based National Notifiable Infectious Diseases Reporting Information System ( NIDRIS ) with a timeframe of 2005–2012 . There are 38 notifiable diseases ( including schistosomiasis ) in P . R . China and all cases are compulsorily reported through NIDRIS according to the National Regulation on the Control of Communicable Diseases since 2005 [22] . Therefore , as soon as an acute schistosomiasis case is diagnosed , it will be reported through the NIDRIS system within 24 hours , including information such as gender , occupation , age , residential address , date of accident or diagnosis , type of diagnosis , infection or outbreak site , etc . All reported acute cases were diagnosed in accordance with the document of the National Criterion on the Schistosomiasis Diagnosis ( WS 261-2006 edition ) [23] . Briefly , the acute case is considered due to the patient who has: ( i ) history of water contact in the endemic areas of schistosomiasis from last four weeks to three months , ( ii ) symptoms of fever , hepatomegaly and peripheral eosinophilia , and ( iii ) schistosoma eggs found from faeces . Outbreaks are determined in accordance with the criteria of Response and Management Scheme for Schistosomiasis Outbreaks [8] and confirmed by experts from the National Institute of Parasitic Diseases , Chinese Center for Disease Control and Prevention [24] . An imported case of acute schistosomiasis is recorded when confirmed to have been infected outside the reported region , e . g . county or province . Both active and passive surveillance covers personal and clinical information , including patient's name , age , gender , address , diagnosis , infection locality , the type of water contact , and the type of endemic area which has been classified and defined in the nationwide schistosomiasis survey conducted in 2004 [25] . The active surveillances were mainly carried out in 82 national surveillance sites covering eight provinces , with snail detection in the Spring and Autumn every year and samples for case detection taken annually in the Autumn . The passive surveillance was carried out in all clinics or hospitals at village , township and county levels . A database was established by double input into Microsoft Excel 2003 , and all data in the established database were analyzed using SPSS v17 . 0 [26] . The chi-square test was used to explore associations between the infection status and age , occupation and sex . The space-specific database was extracted from the database ( including imported cases ) with the criterion of ( i ) cases without identifiable infection location , and ( ii ) infected locations without clear coordinate information . Then the secondary space-time database was established with the county center-of-mass coordinate extracted from the national county electronic map ( 1∶250 000 ) in ArcGIS 9 . 3 ( ESRI , Redlands , CA , USA ) . The space-time clustering was performed by Retrospective Space-Time analysis using discrete Poisson model in SaT Scan V9 . 1 . 1 [27] . The time aggregation was specified as 7 days , which was same as the period of weekly report of acute infection with S . japonicum , followed by Monte Carlo reiteration statistics 999 times . ArcGIS 9 . 3 was used to visualize the space-time cluster regions in map .
A total of 1 , 047 cases were reported as acute schistosomiasis through the NIDRIS from 2005 to 2012 . Among these , 850 ( 81 . 18% ) were confirmed by stool examination , 197 ( 18 . 82% ) were clinically diagnosed , and 85 ( 8 . 12% ) recorded as imported cases . After a peak of 1 , 114 cases in 2003 , a significant reduction in the number of acute schistosomiasis cases presented from 2005 to 2012 . A 97 . 7% reduction was recorded as the yearly report showed a decline from 564 to 13 , and the incidence rates per 100 , 000 population ranging from 0 . 331 to 0 . 007 , dropped by 97 . 8% . Although there are two micro-rebounds in 2009 and 2012 , the lowest was only 3 cases in 2011 ( Table 1 ) . The proportion of imported cases increased sharply after 2009 , especially in 2011 and all three reported cases were identified as imported cases ( Table 1 ) . Only five outbreaks were reported in Hubei province and one in Sichuan province in 2005 , but none was recorded in subsequent years . During the period of study , only six outbreaks were reported in 2005 . Two outbreaks were reported in areas where the criterion of transmission interruption was achieved , namely Xide County , Sichuan province , and Qichun County , Hubei province , respectively . Another four outbreaks occurred in the lake and marshland regions of Hubei province . A total of 55 cases were reported in these outbreaks , and most cases reported they got the infection through swimming ( 4/6 , 66 . 67% ) and farming ( 2/6 , 33 . 33% ) . A space-time database with 993 cases was established by space-time clustering analysis . The cases in 2011 were excluded from the database because only two cases remained by the pre-processing criterion . Acute infection from 2005 to 2010 presented the similar time frame clustering , with a significant core of clusters from June to November every year ( P<0 . 01 ) . However , the time frame clustered in a limited frame in 2012 with a few cases ( Table 4 ) . A total of 13 cluster regions in space were detected by SaT Scan . However , with the decrease of the reported cases , four clusters did not present significantly different results in 2008 , 2010 and 2012 , respectively ( Table 5 ) . Most of the cluster regions ( 11/13 ) were found to overlap in the lake and marshland regions along the Yangtze River Basin , and the other two clusters were found in mountain regions . One of the clusters in 2005 showed the highest Log-likelihood ratio ( LLR = 126 . 56 ) within the biggest cluster region around the Dongting Lake and Poyang Lake regions in Hunan , Hubei , Jiangxi , and Anhui provinces , which also shows the biggest cluster region overlapped with other clusters in subsequent years . The second cluster in 2005 also showed the highest relative risk value ( RR = 69 . 54 ) and that only happened in two counties within one week , where the outbreak of Xide County ( Sichuan province ) happened . Then the LLR decreased over the years with smaller cluster regions . In contrast , however , one of the clusters in 2009 showed the highest relative risk value ( RR = 13 . 44 ) , focused in most parts of Anhui province , then followed by the one of cluster in 2010 ( RR = 11 ) focused in parts of Jiangxi , Anhui and parts of Hubei ( Figure 4 ) .
Since a re-emergence of schistosomiasis occurred in the early 21st century in P . R . China , the national schistosomiasis control program has been given serious priority in the health agenda in P . R . China [3] . Consequently , the national medium and long-term strategic work plan ( 2004–2015 ) for schistosomiasis control was issued in 2004 [28] and the disease control measures were strengthened by using the integrated strategy of control of infection sources after the key role of livestock as a major source of infection was recognized [4] . As a result of these integrated measures achievements have been made in recent years , and the national human and livestock incidence was reduced from 3–5% to 1 . 70% and 1 . 38% at administrative village level by the end of 2008 , respectively [29] . The mid-term goal of the national plan for schistosomiasis control was achieved by 2009 , and the criteria of infection control in all endemic areas and transmission control in mountain areas was reached in Sichuan province in 2008 followed by Yunnan province in 2009 [29] . In 2010 Jiangsu province reached the criteria of transmission control ahead of schedule [30] . The latest data suggest that there were a total of 286 , 824 cases of schistosomiasis japonica in 2011 , a reduction of 65 . 98% in comparison with 2003 , and no occurrence of acute schistosomiasis outbreak had being reported since 2006 [17] . The general endemic situation of schistosomiasis in P . R . China has now reached a historically low level . The trend of acute human infections of S . japonicum , an important indicator in monitoring the national program of schistosomiasis control and risk assessment , from 2005 to 2012 , was analyzed in this study . By 2012 acute infections had decreased significantly , with a reduction of 97 . 7% in comparison with that of 2005 . Except for two sporadic- and micro-rebounds in 2009 and 2012 , no serious outbreak was reported since 2006 . According to the data from 1999 to 2004 , three stages were identified after the longitudinal observation on the annual incidence of acute schistosomiasis . The first is that of increasing infection from 1999 to 2003 with a peak of 1 , 114 acute schistosomiasis cases in 2003 [31] . The available data suggests that schistosomiasis had re-emerged , probably due to multiple factors , including increased snail diffusion after major flooding events and somewhat reduced control efforts when the World Bank Loan Project on schistosomiasis control in China ceased [32] , [33] . An estimated 843 , 011 people were infected with S . japonicum during 2003 . Among them were 1 , 114 cases with acute schistosomiasis and 24 , 441 cases suffering from advanced schistosomiasis . The population at risk of infection was estimated around 65 million and there were 74 , 000 infected cattle [1] . The second stage is one of decreasing infection rates from 2004 to 2008 with an 89 . 89% reduction compared with that of 2005 [4] . The priority given to schistosomiasis control after re-defining re-emergence of the disease , led to the implementation of a new comprehensive strategy for schistosomiasis control by the central and local government authorities in the country [5] . By the end of 2008 , all of the endemic areas reached the criterion of infection control , with Sichuan and Yunnan provinces reaching the criterion of transmission control in 2008 and 2009 respectively [29] . The number of reported cases of acute human schistosomiasis also declined annually from 564 to 57 [4] . The third stage can be identified as a relative plateau from 2009 to 2012 with unstable rebounds and the lowest number of cases in 2011 ( n = 3 ) . It suggests that the trend of re-emergence of schistosomiasis during 2000–2003 ceased after implementation of intensive control efforts in a revised strategy initiated late in 2004 . It also indicated that the decrease trend of the number of acute schistosomiasis was mainly attributed to the intensified efforts of the new integrated strategy in the national schistosomiasis control program in recent years [5] , [34] . Although the total number of acute cases decreased significantly in the past 4 years , without an obvious change in the natural and social factors in the endemic areas , it seems that the situation of human acute schistosomiasis has entered an unstable plateau stage with a high risk of infection remaining in traditional endemic areas [35] . In this study a retrospective space-time analysis identified the lake and marshland regions as susceptible to high-risk of disease and interruption areas being at risk of re-emergence . Nine risk clusters in space and time were detected ( P<0 . 01 ) by SaT Scan soft , most of which are located in the lake and marshland regions and the middle and lower Yangtze River basin where the environment is favorable for vector snail survival . The risk cluster regions were similar to the results of observed risk areas reveled by the sentinel mice monitoring carried out in high-risk water regions in 2010 [36] . One explanation for those risks remained is that the major source of infection is the bovines in the large marshland areas , especially water buffaloes that the farmers raise and use for their agricultural practices [37] . The pastures used by the animals cannot be isolated easily . Another potential reason is that effort in schistosomiasis control was reduced after the points of transmission control or interruption had been reached [4] , [35] . Most of the cases are students , farmers and fishermen who got the infection mainly by swimming in water bodies where transmission and other activities were ongoing during the 27th–43rd week every year , and the proportion of imported cases of acute schistosomiasis increased sharply from 2011 , which agrees with earlier investigations [4] . This study provides evidence-based information on the importance of current control activities for controlling the risk of acute infection among school-aged children , farmers and fishermen , and especially for primary school-aged students [38] . More importantly , it should help in further evaluation of the effects of schistosomiasis control strategies and the shift of such strategies from control of sources of infection to surveillance and response . The latter could include the setting up of more surveillance sentinel sites in the different endemic areas , establishing networks of diagnosis reference laboratory and capacity building . Moreover , the surveillance capacity should be further strengthened not only in the endemic areas , but also in the non-endemic or transmission-interrupted areas [39] , due to mass rural-urban migration movements for socio-economic reasons [40] . In conclusion , the results of this study showed the number or incidence of acute schistosomiasis infections decreased significantly from 2005 to 2012 , as shown in the results of this study , indicating that locally ongoing transmission of schistosomiasis has been reduced to a remarkable extent . It also indicates that the integrated strategy of schistosomiasis control has played a key role in reduction of disease burden , both in morbidity-control and infection-control stages . Thus comparatively new cases of schistosomiasis do occur and a potential risk of the acute infection still exists in the lake and marshland regions or transmission-interruption areas . Therefore , there remains a strong need for implementation of intensive surveillance and response activities during the transition stage from control to elimination of schistosomiasis in P . R . China .
|
A retrospective study on the incidence of acute schistosomiasis in the People's Republic of China ( P . R . China ) was performed , in order to assess the new integrated control strategy that was implemented through the national control program from 2005 to 2012 . The lake and marshland regions have been identified as high risk areas as shown by the nine spatio-temporal clusters that we found in the transmission period between June and November each year . When a total of 1 , 047 reported cases of acute schistosomiasis were analyzed , a reduction in prevalence of 97 . 7% between 2005 and 2012 was found . In contrast , imported cases of acute schistosomiasis increased between 2011 and 2012 . These findings support the approach and effectiveness of the new integrated strategy in the reduction of schistosomiasis morbidity .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"ecology",
"and",
"environmental",
"sciences",
"medicine",
"and",
"health",
"sciences",
"infectious",
"disease",
"epidemiology",
"malacology",
"spatial",
"epidemiology",
"tropical",
"diseases",
"parasitic",
"diseases",
"parasitology",
"global",
"health",
"neglected",
"tropical",
"diseases",
"infectious",
"disease",
"control",
"zoology",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"environmental",
"epidemiology",
"epidemiology",
"helminth",
"infections",
"schistosomiasis",
"disease",
"surveillance",
"biology",
"and",
"life",
"sciences"
] |
2014
|
Reduction Patterns of Acute Schistosomiasis in the People's Republic of China
|
The CDTI model is known to have enhanced community participation in planning and resource mobilization toward the control of onchocerciasis . These effects were expected to translate into better individual acceptance of the intervention and hence high Treatment Coverage , leading to a sustainable community-led strategy and reduction in the disease burden . A survey revealed that after 10–12 rounds of treatment , prevalence of onchocerciasis was still high in three drainage basins of South West Cameroon and transmission was going on . We designed a three ( 3 ) -year retrospective ( 2012 , 2013 and 2014 ) , descriptive cross-sectional study to explore the roles of operational challenges in the failure of CDTI to control the disease as expected . We administered 83 semi-structured questionnaires and conducted 12 in-depth interviews with Chiefs of Bureau Health , Chiefs of Centers , CDDs and Community Heads . Descriptive statistics was used to explore indicators of performance which were supported with views from in-depth interviews . We found that community participation was weak; communities were not deciding time and mode of distributions . Only 6 ( 15 . 0% ) of 40 Community Drug Distributors reported they were selected at general community meetings as required . The health service was not able to meet and discuss Community-Directed Treatment with Ivermectin activities with individual communities partly due to transportation challenges; this was mostly done through letters . Funding was reported to be inadequate and not timely . Funds were not available to conduct Community-Self Monitoring after the 2014 Mass Drug Administration . There was inadequate health staff at the frontline health facility levels , and some Chiefs of Center reported that Community-Directed Treatment with Ivermectin work was too much for them . The mean operational Community Drug Distributor-population ratio was 1 Community Drug Distributor per 317 populations ( range: 194–464 , expected is 1:250 ) . Community Drug Distributor attrition rate was 14% ( 2012 ) , 11% ( 2013 ) and 12% ( 2014 ) of total Community Drug Distributors trained in the region . Lack of incentive for Community Drug Distributor was primary reason for Community Drug Distributor attrition . Number of Community Drug Distributors trained together by health area ranged from 14 to 127 ( mean ± SD = 51 ±32 ) with duration of training ranging from 4–7 hours ( mean ± SD = 5 . 05 ± 1 . 09 ) . The trainings were conducted at the health centers . Community Drug Distributors always conducted census during the past three distributions ( Mean ± SD = 2 . 85 ± 0 . 58 ) . Community-Self Monitoring was facing challenge . Several of the community heads , Chiefs of Bureau Health and Chiefs of Center agreed that Community-Self Monitoring was not being carried out effectively due to lack of incentives for monitors in the communities . Inadequate human resource , funding issues and transportation challenges during distribution periods reduced the ability of the health service to thoroughly sensitize communities and supervise CDTI activities . This resulted in weak community understanding , acceptance and participation in the process . CDTI in our study area did not achieve sustainable community-led campaign and this may have led to the reduced impact on Onchocerciasis .
Cameroon health system is coordinated by a Ministry of Public Health headed by a Minister , who works through ten ( 10 ) Regional Delegations of Public Health headed by medical doctors . A region is divided into Health Districts , headed by district medical officers who are also medical doctors; there are eighteen ( 18 ) health districts in the South West Region . A health district is further divided into health Areas . A health area comprises about 5–10 communities that are served by a health center which may have smaller health posts in addition . The health center is usually headed by a senior nurse called Chief of Center . A nurse in charge of a health post is called Chief of Post . Health decisions at this level are taken by a Health Area Committee . This is the level where the health system interacts with communities , also referred to as frontline health facilities . In CDTI , CDDs collect Ivermectin from here and report directly to health personnel at the frontline health facilities . A community or village in Cameroon is headed by a chief or community head . The community is demarcated into what is known as Quarters , comprising of about twenty ( 20 ) houses depending on the size of the community . Each quarter has a head who takes decisions with household heads of all households in the quarter . The quarter heads and community head forms the village council . They oversee the work of CDDs , other developmental activities and resources in the community . Decisions are disseminated by quarter heads through household heads or through an announcer called ‘town crier’ . However , the influence of these traditional rulers varies among the tribes and is generally eroding away very fast . The implementation of CDTI in Cameroon started in 1998 . There are currently 15 CDTI projects in Cameroon ( the third largest after Nigeria and DR Congo ) . The two CDTI projects , South West I & II situated in the South West Region are responsible for reaching some 1 , 367 endemic communities with Ivermectin [8] . They were approved in 1998 and 1999 , and became operational in 1999 and 2000 respectively . The process started in 1999 in communities of the Meme and Mungo river drainages , and in the year 2000 in communities of the Manyu river drainage ( South West Regional Delegation of Health Database ) . South West CDTI Project I covers the Meme and Mungo drainage basins . It achieved 100% geographic coverage from 2001–2014 , except in 2008 ( 97% ) and 2009 ( 98 . 33% ) . Therapeutic coverage rose from 32 . 56% in 2001 to 82 . 83% in 2010 , and has since achieved at least 81% as at 2014 . The South West CDTI Project II , which covers the Manyu drainage basin , achieved a geographic coverage between 95% - 100% from 2001–2014 . Therapeutic coverage rose from 37 . 2% in 2001 to 83 . 7% in 2009 , the project has since achieved closely same coverage as at 2014 . According to ONCHOSIM predictions , the outcome of elimination of onchocerciasis depends on precontrol endemicity level , frequency of MDA and treatment coverage ( TC ) achieved [11] . Given the precontrol hyperendemic ( 60–98% prevalence ) levels in communities in our study area , CDTI projects here must achieve and sustain annual TCs above 80% in order to bring the disease under control in about 18 years [11] . Kim et al . predicted that the elimination scenario for onchocerciasis is feasible by 2028 in some areas but could go beyond 2045 in countries with operational challenges [12] . Certain challenges have been identified with the CDTI process . These include: maintaining timely drug-collection mechanisms; integrating CDTI with existing primary-healthcare services; strengthening local health infrastructure; achieving and maintaining an optimal treatment coverage; establishing and scaling up community self-monitoring; designing and implementing operational research locally; ensuring the adequacy of community-directed distributors; increasing the involvement of local non-governmental developmental or community-based organizations in the programme; achieving financial sustainability; implementing equitable cost-recovery systems; and engaging in effective advocacy among the stakeholders , especially with the affected communities [13] . After twelve ( 12 ) annual MDA with Ivermectin in communities in three drainage basins in the South West Region of Cameroon , a situation analysis of onchocerciasis through entomological and parasitological ( epidemiological ) surveys revealed that prevalence and transmission did not reduce as predicted ( Prevalence of microfilaria: Meme drainage basin– 52 . 7% , Mungo drainage basin– 41 . 0% , Manyu drainage basin– 33 . 0% ) [14 , 11] . This paper explored the possible roles operational challenges could have played in the failure of CDTI to control the disease in these areas , by examining the implementation practices over the past three distributions ( 2012–2014 ) . We carried out this study to document the programmatic factors associated with the limited impact of CDTI to control onchocerciasis in three drainage basins . We also wanted to assess the community perceptions as well as their level of participation in the CDTI process . We hypothesized that stakeholders were not able to fully implement the CDTI protocols and hence communities did not adhere adequately to treatment .
The study was carried out in five ( 5 ) health districts located in three forest drainage basins of South West Region of Cameroon . These drainage basins receive about eight ( 8 ) months of rainfall and hence create long periods of favorable ecological environment for the insect vector ( blackflies ) . About 90% percentage of the road network in the drainage basins were not tarred . The poor road network , coupled with long rainy seasons pose transportation challenges to health supervision teams . The respondents were the South West Regional Coordinator for Neglected Tropical Diseases ( NTDs ) , the Chiefs of Bureau Health ( CBH ) of five ( 5 ) Health Districts ( 1 female and 4 males ) , 12 Chiefs of Centers ( COC ) ( 12 Health Areas; 8 females and 4 males ) , 40 CDDs ( 7 females and 33 males; 30 farmers; 34 married; all had at least primary school leavers’ certificate ) and 24 Community Heads from communities across the five health districts . Once a community was included in the study , its head and CDDs were sampled . The COCs and CBHs of the health areas and health districts of the community also become part of the respondents . Since communities were randomly selected , bias was reduced . A 3-three year retrospective , cross-sectional and descriptive approaches were used to explore possible weak links in the collaboration of the stakeholders and context specific factors that may be acting as implementation barriers to the CDTI process in the these drainage basins . Data was collected with standard , semi-structured questionnaires and guides for in-depth interviews developed with reference to prescribed functions of stakeholders in the CDTI process [4 , 8 , 15] . Twelve ( 12 ) in-depth interviews were held with a community head , CDD , COC , and CBH from each drainage basin and also with the regional coordinator for NTDs . The interviews were held in English , recorded with a 4GB capacity Xgenx digital voice recorder ( GDVR-901 ) and were transcribed into Microsoft word text format . Performance of CDTI implementers were measured as number of training , census , supervision , CSM , reporting and review meetings conducted during the past three distributions ( 2012–2014 ) . Adequate and timely arrival of funds and drugs , adequate health workers , amount of sensitization done , community participation as well as physical challenges to carrying out of the process were also assessed . Data collection was done from May to July , cleaning and analysis was done during August and September , 2015 . Quantitative information was entered into a template created in Epi info version 3 . 5 . 4 . The data was imported into excel and cleaned . It was then exported to SPSS version 20 and analyzed . Descriptive measures were used to explore indicators of good performance of the CDTI process . All statistical differences were considered significant at p < 0 . 05 . The in-depth interviews were transcribed verbatim and translated where necessary into English by a trained transcriber into separate Microsoft word documents , through the familiarization process . The transcripts were imported into Atlas . Ti , transformed into rich text format ( rtf ) and given file names recognizable by Atlas . Ti . The framework approach was used to screen the data following the development of an initial thematic framework . The transcripts were then coded following the emerging themes previously identified to retrieve meaningful views expressed by our informants . Frequently expressed views were used to explain and to compare with observed trends in the quantitative data . Ethical approval was obtained from the Institutional Review Board , Faculty of Health Science , University of Buea . Administrative authorization was obtained from the South West Regional Delegation of Public Health . Administrative authorizations were also obtained from health districts after thorough review of the study protocol . The objectives , importance and ethical provisions of the study were explained to the respondents and informed verbal consent was obtained before the questionnaires and in-depth interviews were administered .
The communities were participating in CDTI mainly by selecting CDDs and individual members were giving cash incentive to CDDs . All the health personnel interviewed stated that the communities were yet to appropriate the CDTI concept . One CBH stated: “APOC and ministry of public health and the rest I can grade them to be good , not very good . Is the community that is bringing weakness in the partnership” . Distribution activities were planned top-down . All 24 community heads reported they did not participate in deciding the period ( time of year ) and the mode of distributing Ivermectin; whether by door-to-door or at a central location . This was confirmed by the statement of another CBH: “I can say we have not really […] involved them in the planning and when you go and plan something like that and come to tell people without involving them , they become reluctant to participate” . The communities were however given the opportunity to choose their CDDs and decide how to motivate them . Out of 40 CDDs sampled , 25 ( 62 . 5% ) reported having been selected at a meeting of community leaders , 6 ( 15 . 0% ) at a meeting of all community members and 3 ( 7 . 5% ) by health workers ( Fig 2 ) . The primary media used in creating awareness of CDTI activities among the communities included TV/Radio announcements at the regional level; banners , posters , badges ( T-shirts and caps for CDTI personnel and key community members ) at district and health area levels; at community levels , it was done through CDDs , town criers and religious groups . Review of the sensitization records showed that over the past three distributions ( years ) , 2 TV announcements ( alongside news scrow during the distribution weeks ) , 131 radio announcements , 9 872 posters and 11 , 300 fliers ( smaller posters ) were used to sensitize communities in our study area which had a population of 546 136 ( Table 1 ) . The posters and fliers were normally pasted at key public places such as health centers or given to CDDs and other key community members . Only three ( 3 ) out of twenty-four ( 24 ) community heads reported having been completely involved in the sensitization of their members ( Fig 3 ) . Instead of the personal visit to the head of communities and their leaders , health areas were informing them about distribution through letters . One community head reported: “I only hear that […] is where they deposit the drugs and other health units go for it . That’s all the information that I know , but for the […] other issue that they will call or they bring the drugs and the community now […] decides on how to distribute it , to me we have never have such information” . The CBHs and COCs also admitted that the communities have not been fully sensitized on their roles . The view of one COC is presented as follows: “I just know , I think that we have not really sensitized them well , we have not really involved them well . They do not know their duties […] . And so I think that the weakness is our own” . One CBH also pointed out: “The second thing I will say is the failure of the health personnel to sensitize the community members to actually understand what their role is as partners” . The scores by the community heads based on their knowledge of communities’ roles as stakeholders in CDTI are shown in Fig 4 .
Since parasitological and entomological surveys were not done in this study , it is difficult to conclude that implementation challenges were responsible for the inability of CDTI to control onchocerciasis in the study area as expected . The recent parasitological and entomological studies conducted in our study area cited vector competence , favorable breeding conditions for the vector as possible factors responsible for the persisting transmission and high prevalence , though this paper did not compare coverage data from other parts of the country , its findings would likely be same for other endemic areas with similar transmission potentials and geographic characteristics . In such terrains , there would probably be a need for an additional intervention like vaccination when it becomes available , or moving from annual to biannual treatment . The current approach of Test and Treat being used in Cameroon may increase the health gains if it is followed comprehensively . Certain critical weaknesses existed in the implementation process of CDTI in our study area . This included weak community participation towards planning of CDTI activities , sensitization of community members , resource mobilization and monitoring of the process . This may have actually led to low adherence to Ivermectin treatment among community members . Also , inadequate staff at the frontline health facilities , funding issues and transportation challenges derailed efforts of the health service towards implementing adequate training , supervision and monitoring of the process . Future studies should combine entomological , epidemiological , as well as data on the programme performance in order to identify and better explain the factors responsible for certain control outcomes . Also , it would be interesting to examine how closely figures for indictors such as therapeutic coverage , reported by CDTI projects reflect what is actually achieved in the populations . The guidelines on engaging and mobilizing communities for CDTI activities must be comprehensively followed . This should reflect a change in perception of health staff of communities as mere beneficiaries of the process to an attitude that regards the importance of the roles of communities . The roles of communities should be clearly communicated to them , and they should be trained to totally assume those roles . Planning of CDTI activities must begin from community level , with increased community participation in the planning , supervision and monitoring of the process . Again , integration of CDTI with other primary health care deliveries must be quickly broadened to cover many interventions as possible . However , this must be done with caution in order not to dilute the CDTI concept .
|
River blindness is caused by a very tiny , thread-like worm . The disease is better controlled when affected communities are included in the planning and carrying out of distribution of Ivermectin used to treat the disease . For a community to be able to prevent people from getting this disease , members must take Ivermectin once or twice a year , continuously for about 20 years . Hence , the organization in charge of controlling river blindness ( African Programme for Onchocerciasis Control–APOC ) decided that when a control programme is started in a community , the community must be involved and assisted to take full charge of the programme so that within 12 years the community can sustain the distribution of Ivermectin for as long as necessary . This community directed strategy prevented river blindness in many communities . However , after 10–12 years of implementation , studies found that river blindness largely persists in communities in three drainage basins in South West Region of Cameroon . This paper discussed the operational challenges that the programme may have faced in these areas .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"engineering",
"and",
"technology",
"transportation",
"tropical",
"diseases",
"geographical",
"locations",
"social",
"sciences",
"census",
"parasitic",
"diseases",
"research",
"design",
"neglected",
"tropical",
"diseases",
"onchocerciasis",
"infectious",
"disease",
"control",
"africa",
"research",
"and",
"analysis",
"methods",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"cameroon",
"economics",
"people",
"and",
"places",
"finance",
"helminth",
"infections",
"survey",
"research"
] |
2017
|
Programmatic factors associated with the limited impact of Community-Directed Treatment with Ivermectin to control Onchocerciasis in three drainage basins of South West Cameroon
|
During pancreatic development , transcription factor cascades gradually commit precursor populations to the different endocrine cell fate pathways . Although mutational analyses have defined the functions of many individual pancreatic transcription factors , the integrative transcription factor networks required to regulate lineage specification , as well as their sites of action , are poorly understood . In this study , we investigated where and how the transcription factors Nkx2 . 2 and Neurod1 genetically interact to differentially regulate endocrine cell specification . In an Nkx2 . 2 null background , we conditionally deleted Neurod1 in the Pdx1+ pancreatic progenitor cells , the Neurog3+ endocrine progenitor cells , or the glucagon+ alpha cells . These studies determined that , in the absence of Nkx2 . 2 activity , removal of Neurod1 from the Pdx1+ or Neurog3+ progenitor populations is sufficient to reestablish the specification of the PP and epsilon cell lineages . Alternatively , in the absence of Nkx2 . 2 , removal of Neurod1 from the Pdx1+ pancreatic progenitor population , but not the Neurog3+ endocrine progenitor cells , restores alpha cell specification . Subsequent in vitro reporter assays demonstrated that Nkx2 . 2 represses Neurod1 in alpha cells . Based on these findings , we conclude that , although Nkx2 . 2 and Neurod1 are both necessary to promote beta cell differentiation , Nkx2 . 2 must repress Neurod1 in a Pdx1+ pancreatic progenitor population to appropriately commit a subset of Neurog3+ endocrine progenitor cells to the alpha cell lineage . These results are consistent with the proposed idea that Neurog3+ endocrine progenitor cells represent a heterogeneous population of unipotent cells , each restricted to a particular endocrine lineage .
The destruction or dysfunction of the insulin-producing beta cells of the pancreas contributes to a family of metabolic diseases known as diabetes mellitus . Given that the specification of the three major cell types in the pancreas , endocrine , exocrine and ductal cells , occurs in the embryo , understanding the normal course of pancreas development will ultimately facilitate the generation of insulin-producing beta cells from alternative cell sources for beta cell replacement therapies [1] , [2] , [3] . Single knockout mouse models have determined the relative importance of many transcription factors in the process of endocrine cell specification and differentiation . Of particular significance , deletion of the basic helix-loop-helix transcription factor Neurogenin3 ( Neurog3; Ngn3 ) results in the loss of the hormone-producing cell types [4] . Subsequent lineage tracing experiments confirm that hormone-expressing endocrine cell types , including alpha cells ( expressing glucagon ) , beta cells ( insulin ) , delta cells ( somatostatin ) , epsilon cells ( ghrelin ) , and PP cells ( pancreatic polypeptide ) , are Neurog3-derived [5] , [6] . A recent study suggested that each Neurog3+ endocrine progenitor cell within the population is destined to become a single hormone+ cell type [7] . The idea that endocrine progenitor cells are unipotent implies that the transcription factor code responsible for the differentiation of each hormone+ cell type may be delineated before endocrine progenitors are specified . In support of this hypothesis , forced expression of factors within the Pdx1+ pancreatic progenitor cells can affect the resulting complement of differentiated endocrine cells [8] , [9] , [10] . Ultimately , the proper timing and location of transcription factor expression and function during pancreas development is essential for the appropriate differentiation of all the hormone-expressing endocrine cells . The homeobox transcription factor Nkx2 . 2 is a particularly interesting pancreatic regulatory protein due to its dynamic expression pattern and cell-specific regulatory activities . Nkx2 . 2 is widely expressed throughout the early undifferentiated pancreatic epithelium , but gradually becomes restricted to beta cells and a large subset of alpha and PP cells [11] , [12] . Despite its early and widespread expression , deletion of Nkx2 . 2 specifically affects later endocrine lineage specification: beta cells do not form , alpha and PP cell numbers are decreased , and there is a significant increase in the ghrelin cell population . Furthermore , while Nkx2 . 2 is expressed in both glucagon+ alpha cells and insulin+ beta cells [13] and the physical interaction of Nkx2 . 2 with the co-repressor Groucho3 ( Grg3; Tle3 ) occurs in both cell types , the recruitment of a repressor complex to the promoter of the homeobox transcription factor Arx occurs in beta , but not alpha cells [14] , presumably due to cell-specific and/or promoter-specific protein interactions . Disruption of the Nkx2 . 2/Grg3 interaction results in the mis-specification of islet cell types and the subsequent trans-differentiation of beta cells into alpha cells [14] . Studies of other developmental systems , including muscle and CNS , have also provided examples of how a single transcription factor can differentially regulate cell specification [15] , [16] , [17] , [18] . Altogether these studies demonstrate that cell-specific transcription factor regulation plays a fundamental role in cell fate determination and the maintenance of cell identity . While single knockout mouse models can uncover the role of a specific factor in the process of cell fate determination [19] , [20] , [21] , compound deletion mutants demonstrate how multiple transcription factors work together to permit or restrict the differentiation of specific lineages . Whereas the deletion of Arx results in the loss of alpha cells and an increase in beta and delta cells [19] , [22] , deletion of Nkx2 . 2 affects all islet cell types in the pancreas except the delta cell population [12] . Interestingly , simultaneous deletion of these two factors revealed for the first time that Nkx2 . 2 was required to repress somatostatin in the ghrelin-expressing epsilon cell lineage [23] , [24] . Furthermore , the simultaneous deletion of Nkx2 . 2 and the beta cell transcription factor Neurod1 identified an unexpected epistatic relationship between these factors that regulates the formation of the non-beta cell types [25] . While deletion of Neurod1 does not affect the formation of alpha or beta cells , alpha cells are reduced late in development and beta cells undergo catastrophic apoptosis by birth [26] . In contrast , the null mutation of Nkx2 . 2 results in a severe reduction in alpha cells , and beta cells are completely absent [12] , [27] . Despite the expression of Nkx2 . 2 and Neurod1 in beta cells [13] , [26] , [28] and the severe phenotypes associated with beta cells in both single knockout mice [12] , [26] , the simultaneous deletion of Neurod1 and Nkx2 . 2 did not alter the beta cell phenotype but rather restored alpha cell and PP cell formation , while simultaneously reducing the ghrelin-expressing epsilon cells , which are over abundant in the Nkx2 . 2 null pancreas [25] . These examples demonstrate that deciphering the complex pancreatic gene regulatory network will provide valuable insight into the cellular processes required to generate each islet cell type , and will facilitate the in vitro differentiation of functional insulin-producing cells for therapeutic purposes . The Nkx2 . 2−/−;Neurod1−/− ( Nkx2 . 2null;Neurod1null ) compound mutant provides a useful model for how two transcription factors coordinately regulate the specification of multiple endocrine cell types . Our study aimed to dissect the cooperative roles of Nkx2 . 2 and Neurod1 , and determine specifically where and how these factors work together to permit endocrine cell formation in the pancreas . The result of this analysis demonstrated that in the absence of Nkx2 . 2 , deletion of Neurod1 in the Pdx1+ pancreatic progenitors resulted in restoration of the alpha , PP and epsilon cells; however , deletion of Neurod1 from the Neurog3+ endocrine progenitor cells restored the PP and epsilon cells , but only a small population of alpha cells . Using in vitro reporter assays we also showed that Nkx2 . 2 repressed Neurod1 in certain cellular contexts . Consistent with the idea that Neurog3+ cells are unipotent [7] , we hypothesize that Nkx2 . 2 must repress Neurod1 in the Pdx1+ pancreatic progenitors early in development to appropriately prime the Neurog3+ endocrine progenitor cells to become alpha cells .
To determine the precise cell type in which the genetic interaction between Nkx2 . 2 and Neurod1 is required for endocrine cell specification , we conditionally removed Neurod1 from different pancreatic cell populations in the absence of Nkx2 . 2 . We generated a pancreas-specific deletion of Neurod1 in the Nkx2 . 2 null background using Pdx1-cre [29] ( Nkx2 . 2−/−;Neurod1flox/flox;Pdx1-cre , denoted as Nkx2 . 2null;Neurod1Δpanc ) . We first confirmed that the single deletion of Neurod1 in the Pdx1+ cells ( Neurod1Δpanc ) phenocopied the Neurod1null mouse ( Figure 1B , 1F , 1J; Figure S1 ) , displaying the expected reduction in insulin and glucagon mRNA levels at P0 ( Figure 1M; Figure S1 ) [26] , [30] . We also demonstrated that when Neurod1 was deleted from Pdx1+ cells in the absence of Nkx2 . 2 , the pancreas phenotype was identical to the Nkx2 . 2null;Neurod1null mouse [25] ( Figure S1 ) . Specifically , all beta cells were absent , alpha and PP cells were restored , and epsilon cells , which were overabundant in the Nkx2 . 2null , were significantly reduced ( Figure 1A–1L; Figure S1 ) . The partial rescue of the epsilon cells is likely due to the inability of Neurod1 deletion to restore the balance between the epsilon and beta cell populations , similar to the Nkx2 . 2null;Neurod1null mice ( Figure 1N; Figure S1; [25] ) . Hormone expression was quantified using real time PCR and cell numbers were determined with morphometric analysis; these analyses confirmed that the observed gene expression and cellular changes were equivalent between the Nkx2 . 2null;Neurod1Δpanc and the Nkx2 . 2null;Neurod1null ( Figure 1M–1O; Figure S1 ) . Moreover , we confirmed that Neurod1 was appropriately deleted in mutants and controls ( Figure 1P ) . These data demonstrate that in an Nkx2 . 2 null background the deletion of Neurod1 in the pancreas progenitors phenocopies the Nkx2 . 2null;Neurod1null . Given that all hormone-producing endocrine cells are Neurog3-derived [4] , [5] , [6] , we hypothesized that the genetic interaction between Nkx2 . 2 and Neurod1 would be required within the Neurog3+ endocrine progenitors to allow for the specification of particular hormone+ cell types . Using the Neurog3-cre allele [31] , we generated an endocrine progenitor cell-specific deletion of Neurod1 in the Nkx2 . 2 null background ( Nkx2 . 2−/−;Neurod1flox/flox;Neurog3-cre , denoted as Nkx2 . 2null;Neurod1Δendo ) , and assessed the pancreatic endocrine cell phenotype . To achieve optimal recombination in the Neurog3-expressing precursor population , we used the BAC-derived Neurog3-cre allele; Cre is highly co-expressed with Neurog3 in the embryonic pancreas and Cre activity is sufficient to lineage-label all pancreatic endocrine cells in the islet [31] . Importantly , despite the short half-life of Neurog3 protein , we can detect Cre activity in approximately 75% of Neurog3-expressing cells ( Figure S2B ) . Similar to the Nkx2 . 2null;Neurod1Δpanc and Nkx2 . 2null;Neurod1null mice , we observed rescue of PP cells ( Figure 2A , 2B ) , and a large reduction of ghrelin+ epsilon cells in the Nkx2 . 2null;Neurod1Δendo compared with the Nkx2 . 2null mice ( Figure 2C–2H ) . As seen in the Nkx2 . 2null;Neurod1Δpanc and Nkx2 . 2null;Neurod1null mice , there was no rescue of the insulin-producing beta cell population ( Figure 2C–2F; Figure S3 ) . Given this similar phenotype between the Nkx2 . 2null;Neurod1null , Nkx2 . 2null;Neurod1Δpanc and Nkx2 . 2null;Neurod1Δendo we conclude that the genetic interaction between Nkx2 . 2 and Neurod1 is required in the Neurog3+ cells to permit specification of the PP and epsilon cell populations . Changes in the beta , PP and epsilon cell populations were identical when Neurod1 was deleted from either the pancreatic or endocrine progenitors in the absence of Nkx2 . 2 . However , in contrast to the Nkx2 . 2null;Neurod1Δpanc and the Nkx2 . 2null;Neurod1null , the glucagon-expressing alpha cell population was only minimally restored in the Nkx2 . 2null;Neurod1Δendo ( Figure 3A–3D ) . Morphometric analysis ( Figure 3E ) and real time PCR for glucagon expression ( Figure 3F ) confirmed this observation . We also established that the partial rescue was not due to incomplete deletion of Neurod1 by Neurog3-cre , as Neurod1 was reduced at an early stage of Neurog3 expression; becoming almost undetectable in the mutant pancreata by P0 ( Figure 3G; Figure S4 ) . Taken together , these data suggest that the genetic interaction between Nkx2 . 2 and Neurod1 in Pdx1+ progenitors , prior to Neurog3+ endocrine progenitor formation , is required for complete alpha cell formation . Data from the Nkx2 . 2null;Neurod1Δpanc and Nkx2 . 2null;Neurod1Δendo clearly demonstrate that Neurod1 must be deleted from the Pdx1+ progenitor population and not the Neurog3+ endocrine progenitor population to allow for complete rescue of alpha cell formation . Furthermore , the simultaneous loss of Nkx2 . 2 and Neurod1 was able to rescue even the earliest glucagon-expressing cell population; the number of glucagon-expressing cells was equivalent between the Nkx2 . 2null;Neurod1null and wildtype littermate controls at E10 . 5 ( Figure 4A–4D; data not shown ) , Interestingly , the early glucagon-expressing cells are known to express low levels of Pdx1 ( Figure S5; [24] ) . To determine whether the alpha cell restoration was due to deletion of Neurod1 specifically from this glucagon+ ( Pdx1low ) population in the absence of Nkx2 . 2 , we deleted Neurod1 in the glucagon-expressing cells using Glu-cre [32] ( Figure S2C , S2D ) . In the Nkx2 . 2−/−;Neurod1flox/flox;Glu-cre ( denoted as Nkx2 . 2null;Neurod1Δalpha ) , the complement of all hormone-expressing cells in the pancreas was phenotypically identical to the Nkx2 . 2null , as determined by immunofluorescent analysis of islet cell markers ( Figure 5A–5L; data not shown ) and real time PCR for quantitative hormone expression ( Figure 5M–5O; Figure S6 ) . These results suggest that restoration of alpha cells requires the deletion of Neurod1 in Pdx1+ progenitors that have not yet committed to the glucagon-expressing lineage . We hypothesize that Nkx2 . 2 represses Neurod1 in the Pdx1+ cells to give rise to Neurog3+ endocrine progenitor cells that are primed to differentiate into the alpha cell fate . Since Neurod1 is a downstream target of Neurog3 [33] , [34] and the Neurod1 single knockout phenotype does not manifest until the end of gestation [26] , it was surprising that manipulation of Neurod1 within the Neurog3+ endocrine progenitors was not sufficient to rescue the alpha cell fate in the Nkx2 . 2 null background . To begin to reconcile these unexpected results , we re-examined when and where Neurod1 was expressed during pancreatic development . It was previously reported that Neurod1 is expressed at E9 . 5 in the earliest islet precursors , and is often co-expressed with glucagon [26] . Using the Neurod1 null mouse , which has a LacZ insertion into the Neurod1 locus [35] , we confirmed the presence of Pdx1+/Neurod1 ( beta-gal+ ) cells and glucagon+/Neurod1 ( beta-gal+ ) cells in the earliest pancreatic domain ( Figure 6A; Figure S7A ) ; however , not all glucagon+ cells were Neurod1+ ( Figure 6A , 6E ) . Consistent with previous reports [28] , this pattern was also evident at E13 . 5 ( Figure 6B , 6E ) during the stage of pancreas development marked by a major wave of endocrine cell differentiation referred to as the “secondary transition” [36] . Neurod1 is expressed throughout the epithelial cord region , overlapping extensively with the Neurog3+ precursor cells ( Figure S7B , S7C ) . We used expression of the Neurod1:LacZ allele to identify Neurod1 ( beta-gal+ ) cells that co-expressed Neurog3 at E9 . 5 ( Figure 6C ) and at E13 . 5 ( Figure 6D ) . Interestingly , the overlap of Neurog3 and Neurod1 was not exclusive at either age , and a subset of Neurog3+ cells did not express Neurod1 ( Figure 6F ) . We also detected Neurod1 ( beta-gal+ ) expression in a small population of Sox9low cells ( Figure S7D–S7F ) , indicating that Neurod1 expression can be found in cells that are transitioning into Neurog3 precursor cells [37] . Taken together these expression analyses identified heterogeneous populations of Neurog3+ cells and glucagon+ cells based on their expression of Neurod1 , and may suggest that the presence or absence of Neurod1 could influence downstream cell fate decisions . Our cumulative data suggest that Nkx2 . 2 may function to repress Neurod1 in a subset of Pdx1+ pancreatic progenitor cells to promote specification of the alpha cell fate . We had previously determined that Nkx2 . 2 directly activates the Neurod1 promoter in beta cells , which is consistent with the beta cell phenotypes of the single and double knockout mice [12] , [26] , [28] ( Figure 7A ) . To determine whether Nkx2 . 2 could also repress Neurod1 expression in other ( non-beta ) cell contexts , we analyzed the effect of Nkx2 . 2 on Neurod1 expression in alpha cells in vitro . Utilizing previously described Neurod1 promoter deletion constructs [28] we determined that Nkx2 . 2 repressed the Neurod1 promoter in alphaTC1 cells , which express Nkx2 . 2 [28] ( Figure 7 ) . Specifically , the repressive activity of Nkx2 . 2 mapped to the proximal region of the Neurod1 promoter , which is retained in the NDΔ2 promoter construct ( Figure 7B ) . We also determined that , similar to Nkx2 . 2-dependent activation of the Neurod1 promoter in beta cells , Nkx2 . 2 repression required the presence of at least one of the three Nkx2 . 2 binding sites; deletion of either region containing these consensus elements ( promoter constructs NDΔ3 , NDΔ4 ) resulted in a loss of Nkx2 . 2 repression ( Figure 7B ) . To begin to understand how Nkx2 . 2 mediates differential cell context-specific regulatory activities through the same set of promoter elements , we assessed the ability of Nkx2 . 2 to recruit specific cofactors and/or modified histones to the Neurod1 promoter in alpha versus beta cell lines . We previously demonstrated that Nkx2 . 2 preferentially recruits Grg3 and a large co-repressor complex to the inactive Arx promoter in beta cells , but this complex was not present on the same promoter region in alpha cells , where Arx was actively transcribed [14] . Surprisingly , neither Grg3 nor HDAC1 were recruited to the Neurod1 promoter in either alpha or beta cell lines ( data not shown ) , suggesting that Nkx2 . 2 mediates Neurod1 regulation through an alternative mechanism . Interestingly however , we determined that histone H3K4me3 preferentially occupied the Neurod1 promoter in beta cells , and this differential binding was dependent upon the phosphorylation state of Nkx2 . 2 ( Figure 7C ) . Histone H3K27me3 was not significantly present at the Neurod1 promoter in either alpha or beta cell lines ( Figure 7D ) . These results suggest that while Nkx2 . 2 promotes activation of Neurod1 in beta cells [28] , Nkx2 . 2 appears to prevent the activation of the Neurod1 promoter in alpha cells . This finding is consistent with the idea that Nkx2 . 2 is required to prevent expression of Neurod1 in a subset of Pdx1+ progenitor cells and then maintain this repression in “alpha-cell competent” Neurog3-expressing cells , and subsequently mature alpha cells .
Single deletion mutants have identified the importance of a number of transcription factors for the process of endocrine cell differentiation ( reviewed in [38] ) . Interestingly , very few factors when deleted affect only one islet cell type . Therefore we can deduce that each regulatory protein has multiple roles during development and it is likely that different combinations of these factors must be simultaneously present or absent within the endocrine progenitor cells to permit the specification of alpha , beta , delta , epsilon or PP cells . The generation of compound deletion mutants would assist in deciphering this combinatorial transcription factor code . One such example is the regulatory interaction between Nkx2 . 2 and the alpha cell transcription factor Arx; simultaneous deletion revealed that these factors differentially cooperate to affect the specification of several islet cell lineages [23] , [24] . In this current study , we explore the relative roles of Nkx2 . 2 and the beta cell transcription factor Neurod1 . The single deletion mutants for Nkx2 . 2 or Neurod1 display alterations in several islet cell types [12] , [26]; however , these mutants are noted for their severe beta cell phenotypes . In particular , Nkx2 . 2 and Neurod1 are necessary for beta cell specification and maintenance , respectively [12] , [26] . Interestingly , simultaneous deletion of Nkx2 . 2 and Neurod1 did not affect the respective beta cell phenotypes of the single mutants , but rather identified complex genetic interactions between these factors for the specification of alpha , PP and epsilon cells [25] . In this set of experiments , we have determined the cellular locations of the genetic interactions between Nkx2 . 2 and Neurod1 , and have uncovered a possible mechanism for how these transcription factors contribute to the process of alpha cell specification . Given the increasing number of studies identifying transdifferentiation between alpha cells and beta cells [10] , [14] , [39] , refining our understanding of alpha cell development may provide insight into the unique relationship between alpha and beta cells , and ultimately aid in understanding how beta cells develop in both the normal and diseased state . Knowing that all endocrine cell types are derived from Neurog3-expressing cells [5] , [6] , we hypothesized that the genetic interaction between Nkx2 . 2 and Neurod1 would be required in the Neurog3+ endocrine progenitors to specify islet cell fates . In support of this hypothesis , deletion of Neurod1 from the Neurog3+ endocrine progenitor cells in an Nkx2 . 2 null background ( Nkx2 . 2null;Neurod1Δendo ) was sufficient to rescue the relative ratios of the ghrelin-expressing epsilon cells and pancreatic polypeptide-expressing PP cells when compared to the Nkx2 . 2 null phenotype . This demonstrates that the genetic interaction between Nkx2 . 2 and Neurod1 is required within the Neurog3+ endocrine progenitor population to permit appropriate specification of the PP and epsilon cell populations . In contrast , although alpha cells were completely rescued in the Nkx2 . 2null;Neurod1Δpanc , we observed only a minimal restoration of glucagon+ cells in the Nkx2 . 2null;Neurod1Δendo , suggesting that alpha cell recovery requires the genetic interaction between Nkx2 . 2 and Neurod1 to occur within the Pdx1+ pancreatic progenitors , prior to Neurog3+ endocrine progenitor cell formation . This finding would support the concept proposed by Degraz and Herrera [7] that the Neurog3+ endocrine progenitors represent a heterogeneous population of unipotential cells that are already committed to become a single hormone-producing cell fate . If all Neurog3+ progenitors are indeed unipotent , then how do we explain rescue of the PP and ghrelin cell ratios that resulted from manipulating gene expression after the Neurog3+ cells are formed ? It is possible that there are both unipotential and multipotential endocrine progenitor populations . Alternatively the “pro-PP” or “pro-ghrelin” Neurog3+ populations may retain more plasticity throughout development . The latter explanation is consistent with the findings of Johansson et al . , [9] , which demonstrated that as development proceeds the progenitor cells are less competent to produce alpha cells and instead favor the generation of other endocrine cell types . This would suggest that although the alpha cell fate decision can be made at multiple points during development , the ability to generate alpha cells is most robust in the earliest pancreatic progenitors and becomes restricted over time . Alternatively , it is possible that later born progenitors retain a certain degree of plasticity that accounts for their ability to respond to lineage manipulations after Neurog3+ cell specification has occurred . The inability to rescue alpha cells by simultaneously removing Nkx2 . 2 and Neurod1 from the Neurog3+ precursor population , suggests that the genetic interaction between Nkx2 . 2 and Neurod1 is required in the Pdx1+ progenitor population , prior to acquisition of Neurog3 expression . However , it remains possible that there is a spectrum of Neurog3-cre activity within a Neurog3+ precursor cell , with Cre-based inactivation reaching its peak in the middle or late in the lifespan an individual cell . If this were the case , and the genetic interaction between Nkx2 . 2 and Neurod1 is required only early in the lifespan of a Neurog3+ precursor to rescue alpha cells , then Neurog3-cre activity may occur too late within this population to affect its differentiation potential . Although we are unable to resolve the kinetics of Cre activity in the lifespan of a single cell , we can demonstrate co-expression of Neurog3 , Cre and R26R reporter activity , suggesting that although Neurog3 protein expression is transient , Cre is present and active in most of the Neurog3+ population during the time window when Neurog3 is expressed ( Figure S2B ) . Furthermore , published lineage studies using this Neurog3-cre allele demonstrated that all endocrine cells of the islet , including the glucagon-expressing alpha cells , are labeled by a Cre-dependent R26R:LacZ reporter [31] . This would suggest that even if alpha cells can only be differentiated from “young” Neurog3+ precursors , there is sufficient Cre activity at this earliest stage during the lifespan of a Neurog3+ cell to genetically label the alpha cell population . Our failure to recover alpha cells by deleting Neurod1 in a glucagon-expressing population may also be due to the inefficiency of the Glu-cre allele , especially in Nkx2 . 2null embryos that have a severe reduction in alpha cell numbers . However , we detected similar levels of Glu-cre activity in wildtype and Nkx2 . 2null pancreata , which should have been sufficient to permit any possible alpha cell rescue ( Figure S2C–S2D; see Materials and Methods ) . Although caveats exist with the use of Cre/lox technologies , these are currently the best tools available to assess spatial and temporal protein function . Interestingly , we do observe some rescue of alpha cells in the Nkx2 . 2null; Neurod1Δendo embryos . This could be due to deletion of Neurod1 in a subset of Neurog3+ progenitors that have not yet become restricted in their ability to differentiate into alpha cells . Alternatively , the glucagon-expressing cells recovered in the Nkx2 . 2null;Neurod1Δendo may represent alpha cells that form independent of Neurog3 function; such an alpha cell population has been previously documented [40] , [41] . On the other hand , the recovered alpha cells may actually represent a distinct subpopulation of glucagon-expressing cells that express Neurod1 , which would be consistent with our identification of a subpopulation of glucagon+/Neurod1+ cells . While these explanations are not mutually exclusive , the identification of unique alpha cell markers and the generation of genetic tools utilizing these markers , would be necessary to clarify the existence of subpopulations of alpha cells , as well as the factors involved in the generation of these distinct populations . Our findings also suggest that Nkx2 . 2 must regulate Neurod1 differentially in the Pdx1+ progenitor population in the early pancreatic epithelium in order to initiate the specification of different populations of Neurog3-expressing cells . In particular , the prevention of Neurod1 activation by Nkx2 . 2 would result in alpha cell formation , while the activation of Neurod1 by Nkx2 . 2 results in beta cell formation ( Figure 8 ) . This is compatible with our discovery that not all Neurog3+ cells express Neurod1 , and further supports the idea that the Neurog3+/Nkx2 . 2+/Neurod1+ cells most likely become beta cells , whereas Neurog3+/Nkx2 . 2+/Neurod1− cells would become alpha cells . Ideally , we would test this hypothesis by quantifying the increase in the number of Pdx1+/Neurod1+ pancreas progenitors and/or Neurog3+/Neurod1+ endocrine progenitors expected to be observed in the Nkx2 . 2null pancreas; however , this analysis is confounded by the simultaneous loss of the Neurod1+ pro-beta cell progenitor populations in the Nkx2 . 2null pancreas . Instead , we used an in vitro approach to determine whether it was possible for Nkx2 . 2 to differentially regulate the Neurod1 promoter in different cellular contexts . We had previously demonstrated that Neurod1 is activated by the cooperative binding of Nkx2 . 2 and Neurog3 specifically in beta cells [28] . Given the lack of availability of an appropriate pancreatic progenitor cell line , we reasoned that a genetic interaction between Nkx2 . 2 and Neurod1 that was initiated in a “pro-alpha cell” progenitor would be maintained in the mature alpha cell . We utilized alphaTC1 cells , which express Nkx2 . 2 [28] , to demonstrate that Nkx2 . 2 prevents activation of Neurod1 in alpha cells . Highlighting the complexity of gene regulation , the cell type specific regulation of Neurod1 by Nkx2 . 2 appears to function through a mechanism that is different from Nkx2 . 2 regulation of the Arx gene [14] . This may reflect the mechanism by which Nkx2 . 2 functions as an activator and a repressor in the same cell type and/or the presence or absence of cell-specific co-regulatory proteins . As we gain the molecular tools to study transcriptional and epigenetic mechanisms in purified primary pancreatic cell populations , we hope to elucidate the complex regulatory interactions that are required to form and maintain appropriate islet-cell specific gene expression . While the process of endocrine specification likely requires the concerted action of many factors , our data suggest a mechanism that involves the differential regulation of Neurod1 by Nkx2 . 2 in the Pdx1+ pancreatic progenitor cells to direct the subsequent endocrine progenitors to become specific islet cell types . The generation of tools to identify , separate and analyze different subpopulations of Neurog3+ progenitor cells would conclusively determine whether each hormone+ endocrine cell type is derived from a specific unipotent subpopulation of endocrine progenitor cells , each bearing a unique gene profile . Using the pancreas as a model system , our study has provided a prime example of how lineage decisions are established in the developing epithelium . The cooperative action of multiple transcription factors within the early progenitor cells can dictate the fate of subsequent cell lineages . Altering the regulation or complement of this set of factors within the progenitor populations can ultimately skew cell lineage specification . These data have important implications for the current efforts to generate pancreatic cells in vitro for therapeutic use in diabetic patients . Understanding the cooperative transcription factor code will make it possible to initiate the appropriate program in the Pdx1+ pancreatic progenitor cells necessary to correctly prime the Neurog3+ endocrine progenitor cells and generate pools of functional , single hormone-expressing islet cell types in vitro .
All experiments involving mice were approved by the Columbia University Institutional Animal Care and Use Committee and performed in accordance with the National Institutes of Health guidelines for the care and use of animals . All mouse strains were previously generated , and were bred and maintained on an outbred Black Swiss background ( NTac:NIHBS , Taconic ) . Cell-specific Neurod1 null mice were generated by intercrossing Neurod1tm1Kan ( Neurod1flox/flox; [42] ) and either Tg ( Ipf1-cre ) 1Tuv ( Pdx1-cre; [29] ) , Tg ( Neurog3-cre ) C1Able ( Neurog3-cre; [31] ) , or Glu-cre ( [32] ) mice . Neurod1flox/flox;Pdx1-cre and Neurod1flox/flox;Neurog3-cre mice died postnatal , similar to the Neurod1 null ( data not shown; [30] ) . Certain experiments required the use of either Gt ( ROSA ) 26Sortm9 ( CAG-tdTomato ) Hze ( R26R:Tomato; [43] ) or Gt ( ROSA ) 26Sortm1Sor ( R26R:LacZ; [44] ) reporter alleles . The Pdx1-cre will delete Neurod1 in all pancreatic progenitor cells; however , the Pdx1 expression domain also includes a portion of the stomach and the duodenum [45] , [46] . We and others have previously reported the early and relatively non-mosaic activity of the Pdx1-cre allele ( [29] , [47]; Figure S2A ) . Previous characterization of the Neurog3-cre allele demonstrated almost complete co-expression of Neurog3 and Cre and sufficient Cre activity to lineage label all endocrine cells within an islet [31] . Consistent with this published analysis , quantification of cells co-expressing Neurog3 and the LacZ reporter in a Neurog3-Cre;R26R:LacZ E15 . 5 embryo indicated 74 . 82% Cre efficiency ( 268 Neurog3+ beta− gal+/349 total Neurog3+ cells; calculations were performed as described below ( Figure S2B ) . Similar assessment of the Glu-cre mice demonstrated that the Glu-cre allele is active in approximately 30–35% of alpha cells; notably this degree of activity is unchanged in the Nkx2 . 2null background , despite the overall reduction in alpha cell numbers ( Figure S2C , S2D ) . The heterozygous mice ( Neurod1flox/+;Pdx1-cre ) were crossed to Nkx2-2tm1Jlr knock-in mice [12] to generate compound heterozygotes . Embryos were collected from timed matings between Nkx2 . 2+/−;Neurod1flox/+;Pdx1-cre and Nkx2 . 2+/−;Neurod1flox/flox or Nkx2 . 2+/−;Neurod1flox/+;Neurog3-cre and Nkx2 . 2+/−;Neurod1flox/flox or Nkx2 . 2+/−;Neurod1flox/+;Glu-cre and Nkx2 . 2+/−;Neurod1flox/flox mice . Noon on the day of appearance of a vaginal plug was considered embryonic day ( E ) 0 . 5 . The experimental genotypes of wildtype , Nkx2 . 2−/− ( Nkx2 . 2null ) , Neurod1flox/flox;Pdx1-cre ( Neurod1Δpanc ) , Nkx2 . 2−/−;Neurod1flox/flox;Pdx1-cre ( Nkx2 . 2null;Neurod1Δpanc ) , Neurod1flox/flox;Neurog3-cre ( Neurod1Δendo ) , Nkx2 . 2−/−;Neurod1flox/flox;Neurog3-cre ( Nkx2 . 2null;Neurod1Δendo ) , Neurod1flox/flox;Glu-cre ( Neurod1Δalpha ) , and Nkx2 . 2−/−;Neurod1flox/flox;Glu-cre ( Nkx2 . 2null;Neurod1Δalpha ) were studied . Litters were assessed at postnatal day ( P ) 0 . For expression studies , the Neurod1tm1Jle LacZ knock-in ( Neurod1LacZ/+ or Neurod1null ) [35] was used ( also in combination with the Nkx2 . 2null thereby producing Neurod1null;Nkx2 . 2null double knockout embryos; DKO ) , and embryos were assessed at E9 . 5 , E10 . 5 , E13 . 5 and P0 . All embryo dissections were carried out in cold PBS , using a dissecting microscope ( Leica MZ8 ) . A portion of each embryonic tail or yolk sac was detached from the embryo , digested with proteinase K , and DNA extracted for genotyping purposes . Genotyping was carried out with standard conditions and primers as previously described [12] , [29] , [31] , [32] , [35] , [42] . Pancreas was dissected from each embryo and stored in RNAlater ( Ambion ) until RNA was extracted using the NucleoSpin RNAII Kit ( Clontech ) . Subsequently , cDNA was made with equal amounts of RNA for each sample ( Superscript III Kit , Invitrogen , CA ) . Real time PCR was performed using TaqMan gene expression assays ( Applied Biosystems ) for glucagon ( Mm00801712_m1 ) , ghrelin ( Mm00445450_m1 ) , somatostatin ( Mm00436671_m1 ) , insulin1 ( Mm01950294_s1 ) , insulin2 ( Mm00731595_gH ) , pancreatic polypeptide ( Mm00435889_m1 ) and Neurod1 ( Mm01280117_m1 ) . CyclophilinB was used as a control housekeeping gene , and was assayed using a probe and primer set previously described [25] . A standard two-step real time PCR program was used for all genes assessed , with an annealing temperature of 61°C and 40 cycles of amplification ( CFX96 RealTime System C1000 Thermal Cycler , Biorad ) . All gene expression values were normalized to the internal control gene , cyclophilinB , and relative quantification was performed using a standard curve from embryonic age-matched cDNA . Statistical analyses were conducted with Prism Software ( GraphPad Software , La Jolla , CA ) using both the Mann-Whitney test and the Student t-test . Equivalent results were obtained; t-test results were reported in all Figures . Immunofluorescence was performed according to standard protocols , on E9 . 5 , E10 . 5 , E13 . 5 , E15 . 5 and P0 whole embryos that were embedded in OCT , after fixation with 4% PFA and cryopreservation in 30% sucrose . Transverse frozen sections ( 8 µm ) were cut and mounted on glass slides . Sections were stained with rabbit α-ghrelin ( 1∶800; Phoenix Pharmaceuticals , CA ) , goat α-ghrelin ( 1∶800; Santa Cruz ) , guinea pig α-glucagon ( 1∶1000; Linco/Millipore , MA ) , guinea pig α-insulin ( 1∶1000; Millipore ) , rabbit α-insulin ( 1∶1000; Cell Signaling Technology ) , rabbit α-somatostatin ( 1∶200; Phoenix Pharmaceuticals ) , rabbit α-pancreatic polypeptide ( 1∶200; Zymed ) , rabbit α-amylase ( 1∶1000; Sigma ) , rabbit α-Pdx1 ( 1∶1000; Millipore ) , guinea pig α-Pdx1 ( 1∶500; BCBC ) , rabbit α-Neurog3 ( 1∶500; BCBC ) , goat α-Neurog3 ( 1∶500; BCBC ) , goat α-FoxA ( 1∶1000; Santa Cruz ) , rabbit α-sox9 ( 1∶500; Chemicon ) , and chicken α-beta-galactosidase ( 1∶250; Abcam ) . Donkey α-guinea pig-Cy2 , -Cy3 or -Cy5 , α-rabbit-Cy2 or -Cy3 , α-chicken-Cy3 , and α-goat Cy2 or -Cy5 secondary antibodies were used ( 1∶400 , Jackson ImmunoResearch ) . DAPI ( 1∶1000; Invitrogen ) was applied for 30 minutes following secondary antibody incubation . Images were acquired on a Leica DM5500 or Leica 510 confocal microscope . Morphometric analysis was performed by immunostaining every 10th section throughout each embryo ( N = 3 or 4 for each genotype ) . For quantification of individual hormone-expressing cells at P0 , cell number was assessed versus total pancreas as defined by amylase area . For quantification of hormone-expressing cells at E10 . 5 , cell number was assessed versus total pancreas as defined by Pdx1 area . Pancreas area was calculated using ImagePro software . RNA in situ hybridization was performed on 8 µm sections mounted on glass slides as previously described [25] using an antisense riboprobe transcribed from linearized plasmid . The riboprobe for Neurod1 was generated from the plasmid pCS2:MTmNeuroD1 ( J . Lee ) . RNA in situ hybridization was performed on pancreas tissue sections from Neurod1Δendo and wildtype littermate controls at E10 . 5 and Neurog3-cre;R26RLacZ at E15 . 5 . The Neurod1-2 . 2 kb minimal promoter was fused to the firefly luciferase open reading frame in the pGL3 Basic vector ( Promega ) . The alphaTC1 cells were grown in 12-well plates . The design of all Neurod1 promoter deletion constructs and the transfection conditions were previously described [28] . Firefly luciferase readings were normalized to Renilla luciferase values . A Student t-test was performed to determine significance . Point mutations were made to 3xmyc-tagged Nkx2 . 2 cDNA using the QuickChange II Site Directed Mutagenesis kit ( Agilent Technologies ) with the following primers S-11-A: ( FWD ) CAACACAAAGACGGGGTTTGCTGTCAAGGACATCTTGGAC , ( REV ) GTCCAAGATGTCCTTGACAGCAAACCCCGTCTTTGTGTTG; S-11-D: ( FWD ) CAACACAAAGACGGGGTTTGATGTCAAGGACATCTTGGAC , ( REV ) GTCCAAGATGTCCTTGACATCAAACCCCGTTTTGTGTTG . Wild type or mutated Nkx2 . 2 cDNA encoding a triple myc epitope tag ( 250 ng ) was transfected into betaTC6 or alphaTC1 cells using X-treme gene HP ( Roche ) according to manufacturer's protocol . Chromatin was prepared using the ChIP-IT express kit ( Active Motif ) . Immunoprecipitation protocol was modified from Tuteja et al . [48] . In brief , immunoprecipitation was performed using the isolated chromatin diluted in ChIP dilution buffer with 5 micrograms of either mouse anti-H3K27me3 ( Abcam ) or mouse anti-H3K4me3 ( Abcam ) antibodies while rotating overnight at 4°C . The following day antibody/chromatin complexes were pulled down using ChIP grade protein G magnetic beads ( Cell Signaling ) . After washing , antibody/chromatin complexes were eluted from the beads and allowed to rotate at room temperature for 15 minutes . NaCl ( 5 micromolar ) was added to the eluate and incubated at 65°C overnight . The following day Tris-HCl ( 1 M , pH 7 . 5 ) , EDTA ( 0 . 5 M ) and proteinase K ( 10 mg/mL ) were added and allowed to incubate at 37°C for 1 hour . Samples were then purified using the QIAquick PCR purification kit ( Qiagen ) . Quantitative analysis of ChIP products was performed using SYBR Green fluorescence with primers for Gapdh ( FWD – CTCCACGACATACTCAGCACC; REV – TCAACGGCACAGTCAAGGC ) or Neurod1 ( FWD – AAAGGGTTAATCTCTCCTGCGGGT; REV - CATGCGCCATATGGTCTTCCCGGT ) .
|
Diabetes mellitus is a family of metabolic diseases that can result from either destruction or dysfunction of the insulin-producing beta cells of the pancreas . Recent studies have provided hope that generating insulin-producing cells from alternative cell sources may be a possible treatment for diabetes; this includes the observation that pancreatic glucagon-expressing alpha cells can be converted into beta cells under certain physiological or genetic conditions . Our study focuses on two essential beta cell regulatory factors , Nkx2 . 2 and Neurod1 , and demonstrates how their genetic interactions can promote the development of other hormone-expressing cell types , including alpha cells . We determined that , while Nkx2 . 2 is required to activate Neurod1 to promote beta cell formation , Nkx2 . 2 must prevent expression of Neurod1 to allow alpha cell formation . Furthermore , the inactivation of Neurod1 must occur in the earliest pancreatic progenitors , at a stage in the differentiation process earlier than previously believed . These studies contribute to our understanding of the overlapping gene regulatory networks that specify islet cell types and identify the importance of timing and cellular context for these regulatory interactions . Furthermore , our data have broad implications regarding the manipulation of alpha cells or human pluripotent stem cells to generate insulin-producing beta cells for therapeutic purposes .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"developmental",
"biology",
"histology",
"genetics",
"biology",
"genetics",
"and",
"genomics"
] |
2013
|
Regulation of Neurod1 Contributes to the Lineage Potential of Neurogenin3+ Endocrine Precursor Cells in the Pancreas
|
The ApoE ε4 allele is the most significant genetic risk factor for late-onset Alzheimer disease . The risk conferred by ε4 , however , differs across populations , with populations of African ancestry showing lower ε4 risk compared to those of European or Asian ancestry . The cause of this heterogeneity in risk effect is currently unknown; it may be due to environmental or cultural factors correlated with ancestry , or it may be due to genetic variation local to the ApoE region that differs among populations . Exploring these hypotheses may lead to novel , population-specific therapeutics and risk predictions . To test these hypotheses , we analyzed ApoE genotypes and genome-wide array data in individuals from African American and Puerto Rican populations . A total of 1 , 766 African American and 220 Puerto Rican individuals with late-onset Alzheimer disease , and 3 , 730 African American and 169 Puerto Rican cognitively healthy individuals ( > 65 years ) participated in the study . We first assessed average ancestry across the genome ( “global” ancestry ) and then tested it for interaction with ApoE genotypes . Next , we assessed the ancestral background of ApoE alleles ( “local” ancestry ) and tested if ancestry local to ApoE influenced Alzheimer disease risk while controlling for global ancestry . Measures of global ancestry showed no interaction with ApoE risk ( Puerto Rican: p-value = 0 . 49; African American: p-value = 0 . 65 ) . Conversely , ancestry local to the ApoE region showed an interaction with the ApoE ε4 allele in both populations ( Puerto Rican: p-value = 0 . 019; African American: p-value = 0 . 005 ) . ApoE ε4 alleles on an African background conferred a lower risk than those with a European ancestral background , regardless of population ( Puerto Rican: OR = 1 . 26 on African background , OR = 4 . 49 on European; African American: OR = 2 . 34 on African background , OR = 3 . 05 on European background ) . Factors contributing to the lower risk effect in the ApoE gene ε4 allele are likely due to ancestry-specific genetic factors near ApoE rather than non-genetic ethnic , cultural , and environmental factors .
Late-onset Alzheimer disease ( LOAD ) is a progressive neurodegenerative disorder characterized by loss of memory and other cognitive functions . It is the most common form of dementia worldwide [1] , with prevalence increasing with age ( e . g . , ~30–40% by 85–89 years ) [2] . The etiology of AD is multifactorial with genetic , and environmental factors all influencing risk . The most significant genetic risk factor for LOAD is the ApoE gene [3 , 4] . Three common ApoE alleles have been identified ( ε2 , ε3 , and ε4 ) . The ε3 allele is the most frequent and is typically considered “neutral” regarding AD risk . The ApoE ε4 allele both increases the risk and decreases the age-of-onset of developing AD [4] . Conversely , the ε2 allele is protective against AD [4 , 5] . Although ApoE is an AD risk factor in nearly all populations , the risk of AD for ε4 carriers differs among racial/ethnic groups [6] . The strongest reported risk for ε4 allele is in East-Asian populations ( ε3/ε4 odds ratio OR: 3 . 1–5 . 6; ε4/ε4 OR: 11 . 8–33 . 1 ) [6 , 7] followed by non-Hispanic Whites ( NHW ) ( ε3/ε4 odds ratio [OR]: 3 . 2; ε4/ε4 OR: 14 . 9 ) [6 , 8–10] with a considerably lower risk to develop AD for an ε4 carrier in African-Ancestry populations , such as African Americans ( AA ) and Caribbean Hispanics ( CHI ) . Studies in African-ancestry cohorts consistently reported significant association between ApoE ε4 homozygosity and AD , but showed inconsistent results for ε4 heterozygote allele individuals ( ε3/ε4 OR:1 . 1–2 . 2; ε4/ε4 OR: 2 . 2–5 . 7 ) [6 , 8–13] . The reason for this heterogeneous risk effect of ApoE is currently unknown . This disparity in risk may be due to ethnic-related environmental factors that vary across populations , such as diet and lifestyle activities , or the difference may be due to population-specific genetic factors . Exceptions include studies among the Wadi Ara and American Indian populations , but these studies may suffer from low power due to small sample sizes [14–16] . Ancestral methods examining both global ( GA ) and local ( LA ) ancestry can be used to explore these different hypotheses . GA refers to an individual’s average ancestry across his/her entire genome while LA refers to the ancestral background of a particular ( i . e . , “local” ) chromosomal region within an individual genome ( Fig 1 ) . GA is predominantly correlated with ethnic , cultural , and environmental factors that are related to broader definitions of race and ethnicity [17–20] . Conversely , LA is often correlated with ancestry-specific genetic factors that are located in or near the genomic region in question [21 , 22] . As such , an understanding of LA around the ApoE region may help inform how we interpret the race/ethnicity differences observed in ε4 risk . Specifically , if cultural and environmental effects play a major role in ApoE heterogeneity , we would expect GA to interact with ε4 to influence AD risk . There will also be GA and allele interaction if there is epistasis with alleles on other choromsomes that have different frequencies between ancestral populations . However , if genetic modifiers or protective factors local to the ApoE region ( e . g . , cis-acting enhancers , eQTL , etc . ) play a major role in ApoE ε4 heterogeneity , we would expect LA to interact with ε4 to influence AD risk . Admixed populations , due to their ancestral heterogeneity , often show complex patterns of GA and LA , enabling us to test these hypotheses . As such , we utilized two admixed populations ( CHI from Puerto Rico ( PR ) , and AA ) to assess the relationship between ApoE ε4 risk and patterns of GA and LA . PR individuals commonly have European ( EU ) , African ( AF ) and Amerindian ( AI ) ancestors , while AA individuals often have both EU and AF ancestors . To test the hypothesis that the population-specific risk is due to ethnic-related environmental factors that vary across populations , we compared those ApoE ε4 carriers who inherited most of their chromosomes from AF ancestors to those who inherited most of their chromosomes from their EU ancestors by using GA . If there are additional genomic loci outside of the ApoE gene contributing to the population risk difference , then individuals with the highest GA load of EU ( or AF ) ancestry would match the EU ( or AF ) population risk . Alternatively , to test the hypothesis that the disparity in risk may be due to genetic modifiers or protective factors local to ApoE , we compared the LAs in the admixed populations with those of the corresponding ancestral population ( e . g . , if one inherited his/her ApoE LA from the EU ancestors , his/her risk for AD would be similar to the EU population risk ) . Our results strongly suggest that an ancestry-specific region surrounding the ApoE gene is contributing to the lower risk of AD in AA and PR ε4 carriers , supporting the hypothesis that the “protective” effect is due to the ancestry-specific genetic factors around the ApoE genomic region .
First , we performed two genotype-based regression tests to assess global ancestry and local ancestry interaction with ApoE genotype ( see Methods for details ) . Results showed that the LA by ApoE interaction term ( dose of AF ancestry by dose of ε4 allele; LAxApoE ) was significantly different from 0 in both PR and AA populations ( PR: likelihood ratio test ( LRT ) , p-value = 0 . 019; AA: LRT , p-value = 0 . 005 ) . The effect size of the interaction term was negatively correlated with AD ( PR: OR = 0 . 2 ( CI: 0 . 05–0 . 76 ) ; AA: OR = 0 . 75 ( CI: 0 . 61–0 . 91 ) ) . This was in contrast to the GA by ApoE interaction term ( GAxApoE ) , which was not significant in either PR or AA ( PR: LRT , p = 0 . 49; AA: LRT , p-value = 0 . 65 ) . Since we identified a significant interaction , we performed a haplotype-based regression test to assess the effect size of ancestry-specific alleles ( see Methods for details ) . We found that the effect size of the ε4 risk allele was significant across the ancestral haplotypes , even while accounting for correlations with GA ( Table 1 ) . In the PR dataset , the ε4 alleles on an EU ancestral background were significantly associated with AD ( p-value = 3 . 7e-05; OR = 4 . 49 ) compared to ε3 alleles from an EU ancestral background . However , ε4 vs ε3 showed no significant effect on the AF LA background ( p-value = 0 . 67; OR = 1 . 26 ) . Similarly , in the AA dataset , the ε4 haplotypes of EU ancestry showed a stronger risk effect ( OR = 3 . 05; p-value = 4 . 9e-17 ) than those in the AA dataset of AF ancestry ( OR = 2 . 34; p-value = 9 . 2e-45 ) . We tested the difference between the effect sizes of ancestral backgrounds by using t-test for means . Test results showed that effect sizes between the ancestral backgrounds are different with nominal significance in both populations ( PR: p-value = 0 . 059; AA: p-value = 0 . 068 ) . It is of note that these models all include GA as covariates , indicating that the effects seen are independent of the GA . In the subgroup of individuals with homozygote ε4 and ε3 alleles , results showed that ε4 haplotypes of EU ancestry have a stronger risk effect ( OR = 18 . 44 ( CI: 9 . 6–35 . 6 ) ; p-value = 3 . 5e-18 ) than those with AF ancestry ( OR = 6 . 48 ( CI: 3 . 4–12 . 5 ) ; p-value = 4 . 3e-63 ) . The t-test of means showed that effect sizes of EU and AF backgrounds are significantly different ( p-value = 0 . 003 ) . Since we observed that AF ancestral background surrounding the ApoE gene is contributing to the lower risk of AD , we examined the genetic region surrounding ApoE by using 1000 Genome sequence data from three populations of the Utah Residents with Northern and Western European Ancestry ( CEU ) , Japanese in Tokyo ( JPT ) , and Yoruba in Ibadan ( YRI ) . We identified 43 variants using Pearson’s chi-square test between the CEU vs . YRI and JPT vs . YRI populations , which were significant following the Bonferroni correction for multiple comparisons . Table 2 shows the list of 15 most significant variants with the Bonferroni corrected p-values less than 1 ×10−5 . The whole list of variants is shown in the S2 Table . Fig 2 demonstrates Bonferroni corrected p-values for the pairwise comparisons between CEU and YRI , and JPT and YRI populations . The primary CEU and JPT peaks align , and lie within the strongest Topologically Associated Domain ( TAD ) containing the ApoE gene . None of the significantly different variants were in the protein-coding DNA in the defined region around the ApoE gene . It is noteworthy that just 6 variants in sequence data comparison showed significant difference ( with the lowest p-value = 0 . 0052 ) between the CEU and JPT and all of them were found out of the TAD region containing the ApoE .
These findings strongly support our hypothesis that genetic modifiers local to the ApoE region influence the risk of the ε4 allele , showing a weaker risk effect on the AF ancestral background and stronger effect on the EU ancestral background . There was no evidence that overall ancestry ( GA ) has an effect on the heterogeneity of ApoE ε4 risk within the populations , which we used as a surrogate for non-genetic cultural/ethnic differences . Additionally , we observed a stronger risk effect on the EU ε4 haplotypes ( or conversely , a protective effect on AF ε4 haplotypes ) . This effect was especially pronounced in an analysis of ε4 homozygotes against ε3 homozygotes , a result consistent with previous reports on ApoE risk across populations [6 , 8–13] . The overlapping of the subTAD ( ~50kb ) region and the peaks of the allele frequency differences between the CEU , JPT and YRI support the hypothesis that the variant ( s ) modifying ε4 risk are most likely to lie in this region . The significant differences found in non-protein-coding DNA , suggests the protective effect is due to a regulatory difference between the local ancestries . This would also suggest that possible a modifier ( s ) would affect ApoE expression itself and supports the hypothesis that the genomic region surrounding ApoE with AF background reduces the risk for ε4 carriers and is evidence that genetic factors may be underlying the discrepancy in ε4 allele risk effect across populations . It should be noted that this study was not well-powered to test AI background influence on ε4 risk allele . Further research is needed to study populations with higher AI ancestral background , such as Peruvian , Mexican , and Central American populations , to understand the correlation between the AI ancestry and ApoE . Similarly , limitations in sample size prevented us from assessing effects in ε2 carriers . Our findings suggest that the ApoE region from AF populations may contain protective factors that help mitigate the effect of the ε4 allele . In particular , comprehensive analysis of the ApoE region and testing for protective loci may reveal previously unappreciated biological pathways and provide translational opportunities . Research that focuses on locating protective variants represents a complementary approach to accelerating the identification of more effective targets for drug development . This , in turn , will lead to better treatments , and help reduce health disparities .
All AA cases and controls selected for genotyping were obtained from the John P . Hussman Institute for Human Genomics ( HIHG ) at the University of Miami Miller School of Medicine ( Miami , FL ) , North Carolina A&T State University ( Greensboro , NC ) , Case Western Reserve University ( Cleveland , OH ) , and the Alzheimer’s Disease Genetic Consortium ( ADGC ) . Samples were collected as described previously [23 , 24] . The AA dataset contained 1 , 766 AD cases ( 69 . 8% female , mean age at onset ( AAO ) 77 . 6 years [SD 8 . 2] ) and 3 , 730 cognitively healthy controls ( 72 . 0% female , mean age-of-examination ( AOE ) 76 . 5 years [SD ( 8 . 3 ) ] ) . PR individuals were ascertained as a part of the Puerto Rico Alzheimer Disease and Related Disorders Initiative study . Ascertainment was focused in metropolitan areas of New York , North Carolina , Miami , and Puerto Rico . Participants were recruited and enrolled after they ( or a proxy ) provided written informed consent and with approval by the relevant institutional review boards . For the PR cohort , 220 unrelated cases ( 69 . 6% female , mean AAO 75 . 1 years [SD 9 . 7] ) and 169 unrelated cognitively intact controls ( 66 . 4% female , mean AOE 73 . 6 years [SD 7 . 1] ) were ascertained . For both AA and PR datasets , cases were defined as individuals with AD with AAO>65 years of age; controls were defined as individuals with no evidence of cognitive problems and AOE>65 years of age . All participants were evaluated to determine case or control status based on the National Institute of Neurological and Communicative Disorders and Stroke—Alzheimer’s Disease and Related Disorders Association , criteria [25 , 26] . Individuals with known or suspected dementia were evaluated using the LOAD study reference [27] . Individuals who were deemed to be cognitively normal were screened with the Mini-Mental State Examination [28] or the Modified Mini-Mental State [29] . The participants were classified as AA and PR based on self-report , and the GWAS analysis confirmed these data . Genome-wide single-nucleotide polymorphism ( SNP ) genotyping was processed on three different platforms: Expanded Multi-Ethnic Genotyping Array , Illumina 1Mduo ( v3 ) and the Global Screening Array ( Illumina , San Diego , CA , USA ) . ApoE genotyping was performed as in Saunders et al . [30] . Quality control analyses were performed using the PLINK software , v . 2 . [31] . The samples with a call rate less than 90% and with excess or insufficient heterozygosity ( +/- 3 standard deviations ) were excluded . Sex concordance was checked using X chromosome data . To eliminate duplicate and related samples , relatedness among the samples was estimated by using identity by descent ( IBD ) . SNPs with minor allele frequencies less than 0 . 01 and SNPs available in samples with the call rate less than 97% , or those not in Hardy-Weinberg equilibrium ( p<1 . e-5 ) , were eliminated from further analysis [32] . Further details of the QC analysis can be found in the Supplement ( S1 Table ) . To explore the reasons for the differences in ε4 allele risk between the populations we first assessed the genetic ancestry ( LA and GA ) , and then tested the effect of LA and GA on the ε4 allele by building three logistic regression models . To assess the LA , we phased our datasets independently applying the SHAPEIT tool ver . 2 [33] using 1000 Genomes Phase 3 reference panel [34] with default settings . We defined a region around the ApoE that was broad enough ( chr19: 44 , 000 , 000–46 , 000 , 000 ) to include potential enhancers , topological associated domains , etc . while narrow enough to ensure contiguous LA blocks for most individuals in the study . After selecting the ApoE region , we used RFMix [35] , discriminative modeling approach , to infer LA at loci across the genome . We ran RFMix with the TrioPhased option and a minimum node size of 5 . We used Human Genome Diversity Project ( HGDP ) data as the reference panel; two for AA ( EU , and AF ) , and three for PR ( EU , AF , and AI ) . Then , we eliminated samples with ancestral break points across the 2Mb window ( N = 892 ) and labeled each admixture block using the RFMix estimates . As a result , we obtained haplotype data with three LA states ( AF , EU , AI ) in PRs and two ( AF , EU ) in AAs . Afterwards , we defined haplotypes according to LA states and ApoE variants . S1 Fig illustrates the defining of LA at the ApoE gene and S2 Table shows the number of e3 and e4 alleles along AF and EU local ancestry in each population for cases and controls . Next , we assessed GA by performing principal components analysis ( PCA ) using the Eigenstrat program [36] . The AA and PR datasets were combined with reference panels ( using HGDP reference panels ) representing diverse ancestries: EU and AF for AA , and EU , AF and AI for PR . To assess the effects of GA and LA on ε4 risk we used three logistic regression-based models . The first model utilized a genotype-based test to assess GA interaction with ApoE genotype . This model evaluated the role of GA and factors strongly correlated with GA ( e . g . , ethnic-related environmental factors ) on ApoE risk variation among populations . The second model utilized a genotype-based approach to assess LA interaction with ApoE genotype . In this model , we examined the role of genetic modifiers or protective factors local to ApoE in risk variation . The third model utilized a haplotype-based approach to assess the effect sizes of ancestry-specific alleles ( e . g . , ε4 and ε3 alleles on the AF background ) while accounting for correlations with GA . Statistical analyses were performed using the “GLM2” [37] and “GEE” [38] packages available in R computing environment . To define the potential genetic factors modifying the ApoE effect size we assessed the sequence differences between the ancestral backgrounds among the ε4 haplotypes . First , using the 1000 genomes database , we obtained genomic DNA sequence data from three populations of the CEU , JPT , and YRI . Secondly , we extracted the ε4 haplotypes across the defined LA block of 2 mB . In addition to EU , we tested Japanese haplotypes because ε4 allele in East Asian populations has a high-risk effect as well [6 , 7] . Then , we performed Pearson’s chi-square test using allele frequencies at the region of interest among the populations ( CEU vs . YRI and JPT vs . YRI ) to identify the list of significantly different variants that likely contain the protective variant ( s ) . We assessed the allele frequency difference on ε3 and ε4 haplotypes separately . To make a list of ε4 haplotype-specific alleles with the significantly different frequencies we removed those that showed significant difference also among the ε3 haplotypes . Finally , we performed the Bonferroni correction [39] for the multiple comparisons .
|
The strongest risk gene identified for late-onset Alzheimer disease is ApoE . However , the risk for Alzheimer disease due to ApoE is not consistent across populations . For example , individuals with African ancestry experience less risk from ApoE ε4 than individuals of European or Asian ancestry . The cause of the difference in risk effect is currently unknown . This has led us to ask: What is/are the factor ( s ) contributing to the risk effect variation of ApoE across the populations ? We hypothesized two possibilities for the variability of ApoE risk: 1 ) ethnic-related environmental factors that vary across populations , such as diet and lifestyle activities , or 2 ) a population-specific genetic difference in the ApoE gene , or in its surrounding region . We tested our hypothesis using populations with more than one genetic ancestral background , specifically African Americans and Puerto Ricans . Our study showed that the risk of Alzheimer disease is lower for individuals who inherited the genomic region surrounding the ApoE gene from an African ancestor than it is for risk allele carriers who inherited the region from a European ancestor . These findings suggest that protective genetic variant ( s ) most likely lie ( s ) within the genetic region surrounding the ApoE gene on the African ancestral background .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"european",
"union",
"evolutionary",
"biology",
"neurodegenerative",
"diseases",
"population",
"genetics",
"geographical",
"locations",
"genetic",
"mapping",
"ethnicities",
"african",
"american",
"people",
"population",
"biology",
"alzheimer's",
"disease",
"proteins",
"lipoproteins",
"dementia",
"mental",
"health",
"and",
"psychiatry",
"people",
"and",
"places",
"biochemistry",
"haplotypes",
"african",
"people",
"heredity",
"neurology",
"genetics",
"biology",
"and",
"life",
"sciences",
"population",
"groupings",
"europe",
"apolipoprotein",
"genes"
] |
2018
|
Ancestral origin of ApoE ε4 Alzheimer disease risk in Puerto Rican and African American populations
|
Understanding microbial nutritional requirements is a key challenge in microbiology . Here we leverage the recent availability of thousands of automatically generated genome-scale metabolic models to develop a predictor of microbial minimal medium requirements , which we apply to thousands of species to study the relationship between their nutritional requirements and their ecological and genomic traits . We first show that nutritional requirements are more similar among species that co-habit many ecological niches . We then reveal three fundamental characteristics of microbial fastidiousness ( i . e . , complex and specific nutritional requirements ) : ( 1 ) more fastidious microorganisms tend to be more ecologically limited; ( 2 ) fastidiousness is positively associated with smaller genomes and smaller metabolic networks; and ( 3 ) more fastidious species grow more slowly and have less ability to cooperate with other species than more metabolically versatile organisms . These associations reflect the adaptation of fastidious microorganisms to unique niches with few cohabitating species . They also explain how non-fastidious species inhabit many ecological niches with high abundance rates . Taken together , these results advance our understanding microbial nutrition on a large scale , by presenting new nutrition-related associations that govern the distribution of microorganisms in nature .
Microbial nutrition influences global ecosystems , food production , and human disease . However , most microbes in nature are difficult or impossible to cultivate , and have thus traditionally been nearly impossible to study [1] . Recent cultivation-independent methods such as metagenomic sequencing have changed this , revealing inner mechanics of microbes in samples from whole environments , even if these microbes cannot be individually accessed . This has inspired several thousand studies over the last few years , which have begun to extract order out of previously inscrutable ecosystems [2] , [3] , [4] , [5] . Meanwhile and in parallel , broad insights into microbial lifestyles have begun to emerge through a number of new computational strategies [6] , [7] , [8] , [9] . Particularly notable is Genome-Scale Metabolic Modeling ( GSMM ) , which allows analysis of the full metabolic requirements of microorganisms , but until recently was restricted to a selected group of intensively studied microorganisms [10] , [11] . GSMMs have been used in the past to determine potential nutrient needs of organisms , including attempts to determine minimal nutrient requirements of Haemophilus influenza [12] and the neglected tropical parasite Leishmania major [13] , and critical ( i . e . , impossible-to-replace ) functions in Escherichia coli [14] . Media prediction methods have also been used recently in context of automated model building efforts [15] . A breakthrough in GSMM modeling occurred with the development of model-SEED , a resource of automatically built , functioning GSMMs for thousands of microbes , ranging from the barely to the exhaustively studied ( the SEED: [16] , [17] ) . SEED models have been used to analyze metabolic capacities of dozens of species of the antibiotic-producing phylum Actinobacteria [18] , [19] , and have also been combined with large-scale ecological data to reveal cooperation and competition patterns among microbes , including a previously unobserved phenomenon in which unidirectional loops are formed among sets of organisms that each require a metabolite produced by another [7] . The detection of such patterns requires large scale data and the ability to integrate it in a meaningful fashion – a capability that is only now , through these new tools , becoming available . It is in the context of these new capabilities that we use an expanded set of several thousand SEED models to investigate systems features of the ecological lifestyles of microorganisms , searching for new potential associations that relate central metabolic and ecological variables to the minimal nutritional needs of microbes . We focus on minimal needs because many microorganisms in nature are oligotrophs , i . e . , they cannot survive in rich media [20] , [21] . We first develop a method to predict minimal growth media as a kind of specific nutritional signature for these organisms . This method is akin to the ‘minimal reaction sets’ method of Burgard et al . [22] , which determines the smallest number of metabolic reactions that can be active while still enabling biomass production over a certain target level . We apply a similar methodology but focus on minimizing the number of exchange reactions , which allows us to determine the smallest number of compounds that can be taken up by a cell while still enabling some target production of biomass . We next compare these nutritional signatures across organisms from different ecological niches , and find a striking similarity between patterns of environmental co-growth and minimal nutrient sets . We then explore our minimal nutritional predictions in the context of various genomic and environmental factors , and develop new insights into the relationships between microbial nutritional fastidiousness ( i . e . , the complexity and specificity of nutritional needs of an organism ) , environmental versatility , interspecies cooperation , and environmental co-growth . These insights hint at fundamental patterns that might govern microbial growth , and lay a foundation for future development of model-guided chemically defined growth media .
In order to relate microbial nutrition to ecological distributions and lifestyles , it was first necessary to develop a nutritional predictor that could be universally applied across many organisms . Keeping in mind that many ( perhaps most ) microbes in nature are oligotrophs [23] , we chose to focus on the minimal nutritional needs of microorganisms . Developing a minimal medium predictor in itself is an important basic question that could have a broad and significant impact in microbiology [21] . To develop such a predictor , we leveraged genome-scale metabolic models from SEED and developed a new algorithm , the Minimal ENvironment TOol ( MENTO ) , to predict minimal nutritional requirements for microorganisms . MENTO employs a mixed-integer linear programming ( MILP ) algorithm with a GSMM in order to determine the least number of compounds possible to be in an environment while still enabling growth of a specific microorganism . Production of a nominal level of biomass is enforced during this procedure , which ensures that all of the components needed for cell growth can be produced ( see methods ) . The number of individual compounds in a minimal medium computed by MENTO is a natural measure of nutritional fastidiousness , as a higher number of compounds in the computed environment/medium denotes that an organism is more fastidious , i . e . , that it requires a larger set of specific nutrients to survive ( and hence is a nutritional specialist , whereas less-fastidious organisms , which would typically have small MINENVs , are nutritional generalists ) . We observed that minimal environments computed by MENTO contain two types of metabolites: a unique core of critical metabolites that must be present in any predicted minimal or non-minimal growth media environment for a given microorganism , and then a periphery of ‘replaceable’ metabolites that provide essential elements to the microorganism in the given solution but may not appear in other , equally minimal ( in terms of total number of nutrients ) media solutions . To standardize predicted minimal environments for cross-species analysis , we therefore developed a set of unique minimal environment predictions for each species ( hereafter referred to as MINENVs ) , which contain the number of compounds predicted by MENTO to be the minimal possible , but preferentially include low molecular weight compounds for nutrient sources ( this is achieved by solving a second MILP problem , which reduces the total molecular weight of compounds in an organism's MINENV , while not exceeding the number of compounds predicted by the first MILP optimization in MENTO – see Methods ) . The MINENVs were thus predicted aspiring to find the simplest basic compounds that organisms can be predicted to grow on . All subsequent analyses are based on these unique , simplified MINENVs unless otherwise noted . Overall , MENTO was applied to compute the constitution of the MINENVs and critical metabolites for 2529 microorganisms in SEED , and the results can be seen in Tables S1–S3 , with an analysis of the essentiality of the different components provided in Figure S1 in File S3 . We found these MINENVs to be relatively invariant to small changes in a metabolic model ( single gene removals in E . coli only altered the minenv in 6% of cases , and changed the size of the MINENV in only 4% of cases ) , and so serve as a reasonable yardstick for comparison across species . Media have been developed in the past for culturing a large number of microorganisms in the lab . To test and validate our minimal environment predictions , we reconstructed known compositions of defined media for a subset of microorganisms in our dataset from published media in the Leibniz Institute DSMZ media/strain collection ( see Methods ) . We then compare our predicted minimal environments to these known experimental lab media to determine if MINENVs we predict follow the trends of actual medium compositions . It was important to first understand the degree to which compounds we predicted to occur in MINENVs are common in established lab media , and to understand where possible discrepancies lay . We therefore compared the frequency of appearance of metabolites in the MINENVs we had computed versus the frequency of their appearance across all fully defined media in the DSMZ media database , a set that includes 791 distinct media . Metabolites shared between MINENVs and the DSMZ media were more abundant in both datasets than metabolites that were not shared ( p = 5 . 6e-5 and p = 7 . 6e-8 in ranksum tests on frequencies in MINENVs and DSMZ media , respectively ) . Furthermore , among the 90 metabolites shared between the two datasets ( out of 227 in DSMZ and 372 in the MINENVs ) , the frequency of metabolite appearance across the two spaces correlates significantly ( Spearman rho = 0 . 48 , p = 1 . 8e-6; see Figure 1a ) . To better understand the roles of different metabolites , we manually classified metabolites in the set shared by the DSMZ media and the MINENVs into typical biological and chemical roles ( non-defined media components , such as ‘yeast extract’ were excluded from this analysis ) . Notably , the most prevalent metabolites in both datasets are metal ions , followed by coenzymes ( see Figures 1a and 1b ) . Nucleic acid compounds , such as nucleotides and nucleosides , tend to appear frequently in MINENVs but not in DSMZ media , as do amino acids ( Figure 1a ) , whereas simple alcohols tend to be more prevalent in DSMZ media than in MINENVs . Amino acids also came up as the most differentiating metabolites between organisms living in different ecological environments and lifestyle categories ( see analysis in File S1 ) . These observations point to areas in which SEED metabolic models should be refined , but also areas of potential improvement in developing oligotrophic lab media , such as the increased usage of amino acids . The full results of this analysis are provided in Table S1 in File S2 . There was of course no guarantee that the MINENVs we predict , which are not unique among minimal media , would look like the lab media , which are chemically defined but are not necessarily minimal . Nevertheless , we expected that there would be some trend of similarity of MINENVs and DSMZ media per microorganism . To directly test this , we recalculated MINENVs for each microorganism , this time preferentially choosing compounds that are indeed present in the DSMZ media collection as nutritional sources for that strain ( see methods ) . Comparing the re-calculated MINENVs versus the DSMZ media for all 71 DSMZ lab media on which organisms in our MINENV dataset grow , we found that MINENVs are closer to the proper DSMZ medium per microorganism than expected by chance ( p = 0 . 003 in a non parametric test of # significant p-values across the set of 71; see methods ) . This small but significant trend indicates that the calculated MINENVs do capture some of the DSMZ features . However , more work will be required in the exchange capabilities of the SEED models before they are able to fully and reliably recapitulate known lab media . To assess the predictive potential of the MENTO algorithm , we used MINENV predictions to form minimal lab media for each of 5 commonly grown heterotrophic bacteria: Escherichia coli , Agrobacterium tumefaciens , Bacillus subtilis , Pseudomonas aeruginosa , and Serratia marcescens . We transferred the predictions to lab-realizable media using them as a base , and considering typical metabolite concentrations from M9 minimal medium ( e . g . , see formulation of P . aeruginosa medium in Figure S4a–b in File S3 ) . We then experimentally grew each of these microorganisms in microwell batch cultures in each of the media to assess the computational predictions ( See Figure S4c in File S3 and methods ) . Reassuringly , each microorganism grew on the medium based on its own MINENV prediction . However , we far under-predicted growth of microorganisms on media not designed for them ( only 6 of our 17 negative predictions turned out to be correct ) . This result may partially be a reflection of an intrinsic bias towards nutritional flexibility among heterotrophic microorganisms ( such as the bacteria that we tested ) , but it also may point to trend of higher true nutritional flexibility than our models predict . To check whether inaccuracy in the predictions may have arisen because of errors in the SEED models , we also assessed growth in silico in human curated models of E . coli , B . subtilis , and P . aeruginosa on the five lab media . These models predicted significant growth under many more media conditions than the SEED models ( consistent with what was seen in vitro ) , but displayed less precision , and were not overall better predictors ( see Figure S4c in File S3 ) . However , the higher numbers of positive growth phenotypes among these models supports the hypothesis that the SEED models tend to be more pessimistic predictors of growth , and thus that they may require more improvements before they can reliably predict the selectivity of new media . It is also acknowledged that even in the best studied bacteria , not all nutrient scavenging pathways have been characterized [24] , and the automatic gap-filling of SEED models is focused on biomass production but does not attempt to incorporate all potential carbon sources , as this would often require addition of an unacceptably large number of reactions to fill gaps [16] . Organisms that co-grow in an ecological environment must make do with the same set of nutrients . Therefore , we expected that organisms sharing many ecological niches would have similar minimal nutritional needs . To test this , we mapped microorganisms from the SEED database to operational taxonomic units ( OTUs ) from Greengenes , a database of ecological distributions of microbes . Distributions of microorganisms as reported in Greengenes have been grouped previously into ecological environments ( see [7] , [25] , and methods ) . We binned organism pairs based on similarity of their ecological distributions , and within each ecological distance bin , we assessed the percent of organism pairs whose MINENVs are similar above some threshold ( MINENV and greengenes similarities were calculated by Jaccard metrics; see methods ) . Using these criteria , we found a strong correlation between similarity of MINENVs and similarity of ecological distributions for organism-organism pairs [e . g . , ρ≥0 . 70 , p<1e-14 in Spearman test with MINENV similarity thresholds of 30–70%; see Figure 2 , and methods; if we removed insufficiently sampled ecological distance bins ( i . e . , those with less than 1000 org-org pairs out of the possible ∼3 million ) , we obtained significant correlations even up to MINENV similarity threshold of 99% ( rho = 0 . 54 , p = 1 . 9e-4 ) ] . Having established this trend , we next explored the association between nutritional needs of organisms and co-growth within individual environments . To do this , we formed “aggregated” MINENVs for groups of microorganisms by taking the union of their MINENV metabolites , i . e . , by including in an aggregate MINENV all compounds present in all of the MINENVs of organisms making up the aggregate . We observed that aggregate MINENVs computed for two groups of microorganisms from within the same environment tend to be more similar than aggregate MINENVs computed from two random sets of organisms ( p = 3 . 8e-3 in t-test for lower average distance of MINENVs of within- vs . outside-environment organisms across 43 environments , using half the number of organisms per environment to form groups; see Figure 2b and methods; also re-confirmed using only 5 organisms to form the aggregate MINENVs , as shown in Figure S5 in File S3 ) . Finally , we observed that organisms from within an environment grow in silico on the aggregate media of that environment more often than randomly picked organisms ( see Figure 2c; p = 3 . 9e-2 in t-test for difference in survival using FBA , and p = 8 . 3e-12 for a model-free test using DSMZ media , as shown in Figure 2d and explained in methods ) . Taken together , these results indicate that the minimal nutritional needs of organisms correlate strongly with ecological co-growth , and highlight this property as a potential avenue for developing new growth media for yet uncultivable microorganisms . The size of a given MINENV denotes the minimal number of distinct metabolites an organism needs in order to proliferate . Thus , MINENV size of a given organism can be interpreted as a measure of its nutritional fastidiousness , which is a useful yardstick for comparison with other features of microbial lifestyle . One trend we were interested in investigating is the relationship between fastidious growth requirements and environmental versatility , with the hypothesis that more fastidious microorganisms will display less diversity in the environments they live in than less fastidious ones . Certain strong exemplars of this hypothesis exist , including the obligate parasitic species Mycoplasma , which is known to require extensive nutrition to survive , and the notoriously ubiquitous species Pseudomonas aeruginosa , which can survive on a minimal salts medium using a wide variety of single carbon sources [26] , [27] . As a starting point , we compared the number of compounds present in the MINENVs of 25 Mycoplasma species versus 28 Pseudomonas species in our data set , and observed reassuringly that Pseudomonas species have many fewer MINENV components than Mycoplasma species ( medians = 12 and 32 metabolites , respectively; ranksum test p = 3e-10; see Figure 3a ) . To check if our predicted MINENVs show this trend across a broader range of microorganisms , we next mapped microorganisms in our study to a previously developed set of 5 lifestyle categories , whose rank reflects the breadth of environments each microorganism can live in , with higher rank indicating more breadth ( categories are based on and updated from [28]; see Table S6 in File S2 ) . Against this categorization , we obtained a significant negative correlation vs . the number of compounds in the MINENV of each microorganism , indicating that the size of MINENV corresponds negatively to the breadth of environments an organism lives in , as expected ( Spearman rho = −0 . 41 , p = 1 . 4e-20; see Figure 3b ) . We next compared the number of compounds in the MINENV of each microorganism directly against the number of ecological environments each microorganism lives in . To do this , we turned again to the Greengenes database of ecological distributions of organisms . Consistent with our observations from the lifestyle categorizations , we found that MINENV size negatively correlates ( though to a weaker extent ) with the number of environments that a species is found in ( Spearman rho = −0 . 17 , p = 6e-18; see Figure 3c ) . This upholds the general observation that environmental versatility scales negatively with nutritional fastidiousness . Most microorganisms required 11 to 14 metabolites in their MINENVs ( see Figure 3d ) , but there is a long tail of microorganisms that require many more metabolites . As a rule , we found that microorganisms requiring a very large number of metabolites live in a small number of ecological environments ( p = 4 . 6e-6 in ranksum test of the number of environments for microorganisms with > = or <20 metabolites in their MINENV , Figure 3d ) . Of the 33 microorganisms requiring over 35 metabolites , 14 are species of the genus Borrelia , which are known to have highly specific and fastidious nutritional requirements [29]; 4 have the taxonomic designation Candidatus , since they have not been grown in pure culture; and the rest , which include 9 species of Mycoplasma , are similarly either obligate intracellular bacteria or are known to have highly fastidious nutritional requirements . The microorganism with the very highest number of compounds in its MINENV is Mycoplasma genitalium , which is an obligate parasite with one of the most minimized genomes known . These results strongly suggest that more fastidious microorganisms tend to be more ecologically limited , and that obligate intracellular parasites have highly fastidious nutritional requirements , reflected in their MINENV size . Our observation that MINENV size correlates with the breadth of ecological niches led us to seek other potentially interesting variables that might be associated with nutritional fastidiousness . To do this , we compared MINENV size to several common genomic and metabolic metrics . We found that MINENV size correlates strongly negatively with several key genomic metrics , including number of metabolic reactions , number of metabolic genes , and genome size ( rhos< = −0 . 69 for all three in Spearman test; see Figure 4a–c ) . Fastidiousness hence is associated with small metabolic networks and small genomes . This is plausible , considering that a genome size of at least ∼1 . 75 Mb is required to produce all essential compounds in a cell endogenously [30] . In addition to these genomic measures , we also checked two lifestyle measures published previously by our group: average competitiveness score , which measures the tendency of organisms to compete for resources , and average cooperation score , which measures the tendency of organisms to benefit from co-growth due to nutrient sharing ( see: [7] , and methods ) . We found that MINENV size does not correlate with organism competitiveness , whereas it does correlate negatively with cooperation ( rho = −0 . 41 , p = 7 . 4e-7 in Spearman test; see Figure 4d–e ) . The latter observation suggests that with an increased range of niches available to metabolically-versatile non-fastidious organisms , there is increased chance for mixing between different species , and hence a greater advantage in the ability to share nutrients and be more cooperative . Indeed , the cooperation score correlates more strongly with lifestyle class than with any of the other metrics we tested ( see Figure 4g ) , which supports this hypothesis . In addition to the above correlations , we found that MINENV size correlates significantly negatively with growth rate ( rho = −0 . 35 , p = 4 . 8e-3 in Spearman test; Figure 4f ) . Thus not only are less fastidious organisms more able to share nutrients , but they also grow faster than the more fastidious ones . This suggests a mechanism that might contribute to the current distribution of organisms in niches , in which certain highly metabolically versatile organisms tend to dominate across large numbers of niches . Indeed , an analysis of organism abundances for 1408 organisms across 18 human body sites shows a strong association between the number of environments an organism is found in and the abundance of the organism on those environments ( rho = 0 . 78 , p = 4 . 6e-290 for max abundance vs ubiquitousness across organisms; see Figure 4i and 4j; data taken from HMP 16S data freeze in 2010; [31] ) . The higher growth rates of less fastidious organisms might thus contribute to their dominance even in many environments in which fastidious organisms have found their specific niches . This observation also has broad implications for efforts to culture difficult-to-culture organisms . Aside from the fact that many of these organisms are oligotrophs and thus will require minimal media to grow at all , overly rich media will be more likely to also enable growth of more ubiquitous organisms , which with their higher growth rates will outcompete the more fastidious ones in culture . This phenomenon has been seen quite commonly in microbiology labs ( e . g . , fast-growing Methanosarcina spp will outcompete Methanosaeta in culture [23] ) , and is among other things a common cause of contamination . Developing the most minimal possible medium for a given organism is therefore a worthy goal in any effort to culture difficult-to-culture organisms . The predicted MINENVs may serve as a rational starting point for developing such media .
Using a simple new predictor of minimal media for microorganisms , this study explores potential new associations describing the relation between microbial nutritional needs and other central metabolic attributes on a large scale . To this end we developed MENTO , which estimates potential minimal media for thousands of microorganisms based on analysis of genome-scale metabolic models built automatically by SEED . We tested MINENVs predicted by MENTO using laboratory experiments with new designed media and by comparing it to known lab media , and also showed that the predicted nutritional needs of microorganisms that tend to co-grow in nature are more similar than for those that rarely co-grow . We then used predicted MINENVs to explore how nutritional fastidiousness relates to breadth of ecological niches , genome size , and interspecies cooperation , and identify several interesting associations that characterize their relationship . Next , we showed that fastidious organisms tend to live only in specific niches , and that they grow less abundantly than less fastidious organisms even within those same niches , This may contribute to forming the distributions of organisms among populations that are currently observed . As nutritional yardsticks , MINENVs reveal a number of interesting ecological trends . As a source for developing usable minimal growth media , however , the MENTO method still needs further refinement . Notably , because MENTO determines the lowest number of compounds that can be used to fulfill the nutritional needs of a microorganism , it will sometimes ‘pack’ multiple nutritional needs into a single complex compound , such as including 2 , 3-cyclic CMP as the simultaneous carbon , nitrogen , and phosphorous source for E . coli ( see Figure S4 in File S3 ) . This may yield unrealistically ‘simplified’ media , which are in fact harder for an organism to grow on than if these nutritional needs were filled by multiple compounds . Indeed , we also gain noise directly because of our use of SEED models . Although these are the only models available that are appropriate for ecological analysis across many species , they are draft models that are not as refined as manually built GSMMs , and inevitably add a significant amount of uncertainty to our analyses . Some of this noise could be eliminated in the future by re-doing model gap-filling so as to maximize the parsimony between low confidence parts of models for different species , and ensuring that differences are related to the high-confidence parts of the models , in the way that this has been shown manually for the species P . aeruginosa and P . putida [32] . This work is a large undertaking on its own and obviously beyond the scope of the current project . Because of these caveats , one must be careful in interpreting results gained from SEED models . Topological analyses or discovery of broad trends are more appropriate goals than dissection of the biology of specific organisms for new insights , especially in lower-confidence models ( such as many used in this study , which are based on genomic information and computational gap-filling , with no tuning from phenotypic information ) . In light of this , we have focused in this work on broad trends that are likely insensitive to noise . Despite these shortcomings , a large amount of genomic data can now be obtained for uncultivated microorganisms using single-cell genomics or metagenomics , and MENTO may be a useful tool in designing culture media for them . The fact that in lab tests our new media turned out to be less selective than desired should be acknowledged , and points to the limited accuracy of current automatically-built metabolic models . Addition of trace metabolites , as well as certain uncommon compounds ( e . g . , phenol and asparagines-glycine for B . subtilis ) , were required to get some of the models to display in silico growth on media they grew on in vitro . This indicates a need for careful curation of trace metabolite usage in the SEED , as well as plugging of probable gaps in some of the models . Yet MENTO-predicted media may serve as initial recommended starting points in the search for minimal media , and their overall permissiveness to growth of non-target species ( as shown in Figure S4c in File S3 ) might be mitigated by multiple dilutions as done for Sar11 [33] , or by providing additional selective constraints such as growth temperature , salinity and antibiotic supplementation , which may guarantee the survival of the desired species . We therefore supply the MINENV predicted media as a database for use by the research community ( see Table S2 in File S2 ) . We expect that future studies will elaborate and improve on the results presented in this work , and will hopefully reveal further associations that underlie microbial distributions and growth . Some of the organisms for which we provide MINENVs are currently challenging to culture ( e . g . , some strains of Prochorococcus and Pelagibacter ) , and in addition , methods are being developed for determining full or near-full genomes of unculturable organisms [34] . With sequenced genomes of such organisms , the SEED pipeline can produce metabolic models that we can then analyze using MENTO , and hopefully speed the development of new culture media . Because of the tight connections between oligotrophy , fastidiousness , and development of new highly minimal media , MENTO may thus become a key addition towards this important and timely goal .
We obtained 3286 strain-specific genome-scale metabolic models from SEED , spanning all microorganisms with genomes in the SEED database as of late 2011 . All microorganisms could obtain biomass in silico when grown on rich medium . This list was shortened to 2529 models by taking only those built from the biggest genome that mapped to each taxon ID , as these typically represented the most complete genomes from multiple iterative sequencing/assembly efforts , and would therefore include fewer gap-filled ( and thus low-confidence ) reactions . MENTO is an algorithm for predicting minimal media components , which proceeds in two steps: To check that the ordering of metabolites would not lead to degenerate solutions ( i . e . , choosing one compound of a given molecular weight when another could have been used instead ) , we took the unique MINENVs we calculated from this step and searched for any single compounds from any of them that could be switched out with another single compound of the same molecular weight . Reassuringly , we found that across all of the organisms for which we calculated MINENVs , no compounds could be thus exchanged while still maintaining the minimum required biomass production . Critical metabolites are the metabolites within MINENVs that cannot be exchanged for any other compound under any condition in which a cell can produce biomass . MINENVs and critical metabolites may be defined for different biomass cutoffs . Unless otherwise noted , a ‘nonzero biomass’ cutoff of biomass > = 0 . 1 was used . The first optimization of MENTO is formulated as a mixed integer linear program ( MILP ) as follows:Where: S is the stoichiometric representation of the metabolic model , where the columns represent the reactions in the model . V is a vector that represents the fluxes for all reactions in the model , and fj represents the flux of reaction j in the stoichiometric matrix . fBM represents the flux of the biomass reaction in the metabolic model . Here , in addition to the usual FBA constraints , there is: Identifying a minimal set of exchange reactions and their associated metabolites in a medium then amounts to maximizing the sum of the ϑi variables over all exchange reaction fluxes in V{fi} . From this calculation we get the minimal number of metabolites needed for the microorganism to grow at least with fBM-≥Minimal_biomass . This minimal number of metabolites is what we use as our measure of fastidiousness , and the set of metabolites making up the minimal environment after solving are referred to as ‘non-unique minimal environments’ ( since there could be various sets of compounds that could equivalently be output by the optimization ) . These ‘non-unique minimal environments’ are only briefly analyzed and mentioned , in preference for the MINENVs calculated in the next step . The second optimization in MENTO is formulated as follows:Here we have changed the objective function and added an additional constraint , which limits the number of up-taken external metabolites in the media to the number found in the previous step . The molar_index_of_fi is the index of the exchange metabolite in a vector where all metabolites are sorted by their molar weight in an ascending order . The output of this phase of MENTO is a list of metabolites needed for the microorganism to grow at least with fBM≥Minimal_biomass . We call this list of metabolites a MINENV . Ecological distributions of microorganisms were obtained from the Greengenes database [35] using mappings developed in [25] . Greengenes holds 16S rRNA sequences for microorganisms found in samples reported in hundreds of scientific papers . We used a sample-to-environment mapping developed in [25] , in which keywords in the greengenes sample descriptions were used along with a mapping to the ENVO database [25] to determine an environment corresponding to each sample . The result of these previously developed mappings is a set of rRNA sequences , representing operational taxonomic units ( OTUs ) , distributed among a set of ecological environments in which they were found to occur . Sequences of 16S rRNA for all sample OTUs were mapped to microorganisms from SEED using BLASTP ( with cutoffs of e-value< = 10e-10 and amino acid identity > = 99% ) . Finally , SEED microorganisms mapped to environmental 16S sequences were assigned to the appropriate environments . To investigate artifacts caused by mappings of multiple SEED models to a single 16S sequence , we employed certain filters in mapping species for some of our analyses . Typically , a single microorganisms was chosen to represent each 16S sequence , in order to eliminate redundancy . Media were designed based on minenv predictions , with some compound substitutions to make the media easier to produce or to make them more selective , as shown in Figure S4 in File S3 . Compositions of the 5 media are shown in Table S5 in File S2 . Tests of manual models were done using iMO1056 [27] , iBsu1103 [36] , and iAF1260 [37] for P . aeruginosa , B . subtilis , and E . coli , respectively . A minimal number of extra metabolites ( phosphate , nh4 , mn2 , zn2 , cu2 , ca2 , cl , k , mg2 , cobalt2 , and fe3 ) was added to each model in all conditions in order to allow in silico growth . Bacterial species of interest ( Escherichia coli , Pseudomonas aeruginosa , Agrobacterium tumefaciens , Bacillus subtilis , and Serratia marcescens ) were grown overnight in LB to stationery phase , then washed in sterile NaCl ( 0 . 9% w/v ) solution and diluted ×100 into the required media . Each species was inoculated to each of 5 specific minimal growth media as well as in LB ( positive control ) . The bacteria/media combination was then grown overnight at 37 degrees C with shaking . Two types of growth assays were used: The typical similarity metric used in this study is the Jaccard similarity , which is calculated for two binary vectors thus:Where S1 and S2 represent sets of binary properties of two microorganisms . Jaccard similarities are calculated for a number of properties: notably , MINENVs ( each property in the sets representing presence or absence of a given metabolite in the microorganism's MINENV ) and ecological distribution ( each property representing presence of a microorganism in a given environment ) . When other metrics are used , we describe them . PDF files for the DSMZ media were downloaded from: http://www . dsmz . de/ ? id=441 . Files were manually parsed to extract media components , and components were then manually linked to SEED compounds . The mapping of media to organisms that grow on them was kindly provided by DSMZ . To obtain single scores for each species , we averaged cooperation or competition scores over all pairings with other species ( 118 species total ) from [7] .
|
Understanding microbial nutrition is critical for understanding microbial life , and thus has a major influence in many areas of biology . In recent years , the traditional methods of studying microbial nutrition , which rely on culturing bacteria and assessing their nutritional needs through extensive experiments , have been augmented by the development of genome-scale metabolic models , which enable in-depth analysis and prediction of nutrition for a few well-studied organisms . Recently , a pipeline was developed for generating genome-scale metabolic models automatically ( the SEED ) . Here , we leverage models built from this pipeline in order to develop a novel predictor of microbial minimal medium requirements , which we then apply broadly for thousands of microbes across the tree of life . We first show that nutritional requirements are more similar among microorganisms that co-habit many ecological niches . We then use our medium predictions to examine the fastidiousness of organisms ( i . e . , their need for complex/specific media ) , and suggest an explanation for certain observed features of microbial abundance patterns . This study is one of the first to leverage genome-scale models on a large ( >1000 species ) scale , and sets the potential for a new host of strategies for understanding microbial nutrition and ecology in the future .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"systems",
"biology",
"biochemistry",
"systems",
"ecology",
"computer",
"and",
"information",
"sciences",
"ecology",
"network",
"analysis",
"metabolic",
"pathways",
"biology",
"and",
"life",
"sciences",
"metabolic",
"networks",
"microbiology",
"metabolism",
"microbial",
"ecology"
] |
2014
|
A Novel Nutritional Predictor Links Microbial Fastidiousness with Lowered Ubiquity, Growth Rate, and Cooperativeness
|
We evaluated the fraction of variation in HIV-1 set point viral load attributable to viral or human genetic factors by using joint host/pathogen genetic data from 541 HIV infected individuals . We show that viral genetic diversity explains 29% of the variation in viral load while host factors explain 8 . 4% . Using a joint model including both host and viral effects , we estimate a total of 30% heritability , indicating that most of the host effects are reflected in viral sequence variation .
There are differences in the rate of disease progression among individuals infected with HIV . An easy to measure and reliable correlate of disease progression is the mean log viral load ( HIV RNA copies per ml of plasma ) . The viral load measured during the chronic phase of infection ( referred to as setpoint viral load , spVL ) exhibits large variation in a population . Several studies have been carried out to elucidate whether this variation is primarily driven by host genetics [1–4] , viral genetics [5–9] , or environmental effects [7] . Genome-wide association studies consistently show that amino acid polymorphisms in the peptide binding groove of the HLA-A and HLA–B proteins are associated with the viral load of an individual . Furthermore , variants in the HLA-C and CCR5 genes have also been shown to impact spVL . However , those host factors explain less than 15% of the observed phenotypic variance [4] . In contrast , viral genetic studies and studies of donor-recipient transmission pairs established that 33% of the phenotypic variance is attributable to the transmitted virus itself [5 , 10–13] . HIV is an extremely variable and adaptive organism with a rapid replication time , and high rates of mutation . Within-host evolution of the viral population occurs during the chronic phase of infection in which the pathogen adapts to its host environment . Several studies showed that a major proportion of the viral sequence is under selective pressure in the host environment , and several viral amino acid changes are associated with host genetic variants in the Human Leukocyte Antigen ( HLA ) genes [14 , 15] . Viral strains harbor epitope sequences that can be presented by HLA class I proteins of the infected host , which allows the detection and killing of infected cells . The viral population evades detection through escape mutations that modify the epitope sequence but may incur a fitness cost . Compensatory mutations may follow until the viral population reaches its optimal place in a sequence space constrained by the host immune system [16] . There are two main different approaches to viral heritability estimation in the literature . The first one is based on the regression of phenotypic values in donor-recipient transmission pairs , while the other quantifies the difference between the observed phenotypic variance-covariance structure and the phylogenetic variance-covariance structure . Because our study population did not include donor-recipient data , we used the latter strategy . In particular we used linear mixed models ( LMMs ) to explain inter-patient differences in spVL while taking into account host and viral genetic relatedness . LMMs use the pairwise relatedness of individuals with respect to a large set of features ( rather than the individual data points ) to estimate the fraction of phenotypic variance attributable to those features . Such models have been successfully applied to estimate narrow-sense heritability from genome-wide genotype data [17] . Concurrently , LMMs were proposed to incorporate phylogenetic relatedness between samples in comparative analyses [18] , a technique that was further developed to estimate the viral genetic contribution to spVL [6 , 8] .
To estimate the respective contribution of host and viral genetics to the variation in spontaneous HIV control , we collected paired viral/host genotypes along with spVL measurements from 541 chronically infected individuals enrolled in two prospective cohort studies in Switzerland and in Canada . We estimated the respective contributions of host and viral genetics to spVL by defining two relatedness measures , one with respect to the host genotypes , the other with respect to the viral genotypes , and used these jointly in a linear mixed model . On the host side , we focused on amino acid variations in the HLA-A , B and C genes due to their established associations with HIV control [1] . In particular , we used 33 amino acid polymorphisms selected by L1 regularized regression [19] to represent the genetic relatedness of the host ( S1 Table ) . Principal component analysis based on host genome-wide genotype data confirmed the lack of major population stratification in the host sample . We built three LMMs , one containing human variants , one derived from phylogenetic trees , and one including both host and virus information ( Fig 1 ) . The genetic relatedness matrix created from 33 amino acid polymorphisms of the human class I HLA genes explained 8 . 4% ( SD = 4% ) of the observed variance in spVL . In contrast , 28 . 8% ( SD = 11% ) of phenotypic variation was attributable to the viral phylogenetic tree . Combining the two relatedness matrices in one model yielded a total variance explained of 29 . 9% ( SD = 12% ) , less than the sum of the latter two models . Thus , we show that HLA polymorphisms do not explain additional phenotypic variance beyond viral sequence variation . We next assessed the contribution of viral variants most likely to have an impact on spVL . These included amino acids in known CTL epitopes [20] and those positions whose variation is associated with host polymorphisms [14] ( 82% , 60% and 84% of gag , pol , nef codons respectively , S2 Table ) . We used phylogenetic trees built from those codons to show that viral variation in epitopes or other HLA-associated positions explain 23 . 6% ( SD = 11% ) of phenotypic variance . However , this explained fraction might be overestimated due to linkage disequilibrium on the viral haplotype . We therefore repeated the analysis after randomly picking 70% of variable viral positions , and obtained very similar results . We thus cannot conclude that viral variants in known epitopes contribute disproportionately to variance in spVL . Additional evidence for the existence of substantial linkage disequilibrium on the viral haplotype comes from the analysis of the smaller , complementary set of variable viral positions ( located in non-epitope regions ) , which explained 18 . 5% ( SD = 10% ) of the phenotypic variance . This leads to lower bounds of 11 . 4% and 6 . 3% of variance in spVL explained by variation in epitope and non-epitope regions , respectively , leaving 12 . 2% of variance unresolved due to linkage disequilibrium .
By jointly analyzing host and viral genetic relatedness , we here provide estimates of the total and respective contributions of human and viral genetic variation to HIV control . Our results do not challenge the current consensus estimates of the host or viral contributions to spVL . Nevertheless , our combined analysis demonstrates that human HLA polymorphisms do not explain additional variance in spVL once viral genetic diversity is taken into account . The difference between the variance explained by viral phylogeny and the variance explained by HLA polymorphisms may be attributed to two effects . First , selected viral variants might provide a better surrogate of the impact of the host genotype than the imputed host amino acid variants we used . Rare host genetic factors outside of the major histocompatibility complex region ( e . g . the CCR5 deletion ) , as well as environmental interactions may influence viral fitness , and these effects are not accounted for in our estimate of host heritability . Thus some host effects might be missed from the host partition , while their footprint in the virus is still detected in the viral partition . Second , the difference could partly be due to the effect of viral variation independent of the current host , including transmitted escape mutations , i . e . viral sequence variation carried over from the previous host , rather than induced by the current host . Indeed , a recent study showed that spVL is dependent on the degree of pre-adaptation of the viral strain to the HLA class I genotype of the current host [21] . In particular , an increase in the frequency of pre-existing escape mutations , at the population level , led to higher viral heritability estimates . This indicates that both host and viral estimates of heritability depend on the amount of pre-adaptation in the sample population , which varies based on the level of HLA diversity . It has also been shown that reversion of some fitness reducing escape variants is very slow , potentially allowing for a transitory but measurable effect on viral load at the population level [15 , 22] . A limitation of our study is the fact that study participants were collected from two cohorts . To reduce batch effect , we included a cohort-specific variable in all our models . Still , differences in inclusion criteria , health system , geographical exposure and other factors are very likely to increase environmental variance , thus negatively impacting our heritability estimates . Another potential shortcoming is our implicit assumption of the absence of selection on spVL , which might be incorrect , as suggested by recent studies [23 , 24] , and might thus lead to over- or under-estimation of heritability due to model misspecification . Still , because our estimates are comparable to results obtained in donor-recipient transmission studies and in host-genetic studies , we conclude that they are useful for the purpose of delineating the respective amounts of host and viral contributions to phenotypic variation of HIV spVL . In conclusion , our results suggest that host genetic association studies not taking the virus into account underestimate the population level effect of host genetic variation . Combining host and pathogen data provides additional insight into the genetic determinants of the clinical outcome of HIV infection , which can serve as a model for other chronic infectious diseases .
All participants were HIV-1-infected adults , and written informed consent for genetic testing was obtained from all individuals as part of the original study in which they were enrolled . Ethical approval was obtained from institutional review boards for each of the respective contributing centers . Bulk sequences of the HIV-1 gag , pol and nef genes , human genome-wide genotyping data and viral load measurements were obtained for 541 individuals of Western European ancestry infected with HIV-1 Subtype B , and followed in the Swiss HIV Cohort Study ( SHCS , www . shcs . ch ) or in the HAART Observational Medical Evaluation and Research study in Canada ( HOMER , www . cfenet . ubc . ca/our-work/initiatives/homer ) [14] . Viral sequences data were generated from samples collected two to five years after infection ( for SHCS ) or during chronic infection ( for HOMER ) but prior to the initiation of antiretroviral therapy . Thus , the viral genotypes reflect the result of natural adaptation of the pathogen to the host environment . The viral sequences for 1262 , 2187 and 548 nucleotides of the gag , pol and nef genes were available for at least 80% of samples studied . The analysis was limited to these three genes because sequences of the rest of the retroviral genome were not available for the majority of study samples . Overlapping viral genomic regions were excluded from gag , to avoid duplicated sequences in the analysis . Human DNA samples were genotyped in the context of previous genome-wide association studies . High-resolution HLA class I typing ( 4 digits; HLA-A , HLA-B , and HLA-C ) was imputed from the genome-wide genotyping data as described previously [14] . Set point viral load ( spVL ) was defined as the average of the log10-transformed numbers of HIV-1 RNA copies per ml of plasma obtained in the absence of antiretroviral therapy , excluding VL measured in the first 6 months after seroconversion and during periods of advanced immunosuppression ( i . e . , with <100 CD4+ T cells per ul of blood ) . The distributions of spVL in the two cohorts are shown in S1 Fig . The pairwise genetic relatedness of the dominant viral strains observed in the samples was calculated from phylogenetic trees similarly to [6] . Nucleotide sequences were translated to amino acid sequences , which were in turn aligned with MUSCLE [25] and used to derive the correct codon-aware nucleotide alignment . The phylogenetic tree was built from the aligned nucleotide sequences using RAxML [26] with the following command line: “raxml -w {PATH} -s {PATH} -m GTRCAT -f a -N 30 -k -n {NAME} -T {NUMBER} -x 1234 -p 1234” . Individual sequences were then rooted to the HIV-1 group M ancestral sequence , downloaded from the Los Alamos sequence database . Using an HIV-1 subtype C sequence as outgroup led to similar results . The whole tree was scaled with the inverse of the median height of the branches . We followed the method of Hodcroft et al , to create a relatedness matrix from a phylogenetic tree [6] . The genetic relatedness of two samples in a given phylogenetic tree is the amount of shared ancestry , i . e . the distance from the root of the tree ( excluding the outgroup ) to their most recent common ancestor [27] . We selected 33 amino acid variants with L1-regularized regression ( LASSO ) out of all polymorphisms in the HLA-A , B and C genes and used them to generate a genetic relatedness matrix as described in [17] . Our relatively small sample size made it necessary to use a small subset of selected markers rather than genome-wide variant information to create the genetic relatedness matrix . Doing otherwise would have resulted in very large errors of the estimates . To estimate heritability , we used the gcta software as a generic implementation of the linear mixed model [17] . In such a framework , a multivariate Gaussian distribution models HIV viral load with a variance-covariance matrix consisting of the linear combination of the sample-sample genetic relatedness matrices ( one for the host and one for the virus ) and the identity matrix ( representing sample-specific noise ) . The total heritability estimate is the fraction of variance explained by the genetic relatedness matrices over the total variance . All models included a binary variable indicating cohort as a fixed effect . Variance components were estimated by restricted maximum likelihood .
|
Viral loads of Human Immunodeficiency Virus infections are correlated between the donor and the recipient of the transmission pair . Similarly , human genetic factors may modulate viral load . In this study we estimate the extents to which viral load is heritable either via the viral genotype ( from donor to recipient ) or via the host’s Human Leukocyte Antigen ( HLA ) genotype . We find that a major fraction of inter individual variability is explained by the similarity of the viral genotypes , and that human genetic variation in the HLA region provide little additional explanatory power .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genome-wide",
"association",
"studies",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"evolutionary",
"biology",
"viral",
"transmission",
"and",
"infection",
"pathogens",
"population",
"genetics",
"microbiology",
"retroviruses",
"viruses",
"immunodeficiency",
"viruses",
"rna",
"viruses",
"phylogenetic",
"analysis",
"genome",
"analysis",
"molecular",
"biology",
"techniques",
"population",
"biology",
"microbial",
"genetics",
"viral",
"load",
"research",
"and",
"analysis",
"methods",
"genomics",
"genetic",
"polymorphism",
"medical",
"microbiology",
"hiv",
"microbial",
"pathogens",
"hiv-1",
"molecular",
"biology",
"viral",
"genetics",
"molecular",
"biology",
"assays",
"and",
"analysis",
"techniques",
"virology",
"viral",
"pathogens",
"genetics",
"biology",
"and",
"life",
"sciences",
"computational",
"biology",
"lentivirus",
"organisms",
"human",
"genetics"
] |
2017
|
Estimating the Respective Contributions of Human and Viral Genetic Variation to HIV Control
|
Most pairwise and multiple sequence alignment programs seek alignments with optimal scores . Central to defining such scores is selecting a set of substitution scores for aligned amino acids or nucleotides . For local pairwise alignment , substitution scores are implicitly of log-odds form . We now extend the log-odds formalism to multiple alignments , using Bayesian methods to construct “BILD” ( “Bayesian Integral Log-odds” ) substitution scores from prior distributions describing columns of related letters . This approach has been used previously only to define scores for aligning individual sequences to sequence profiles , but it has much broader applicability . We describe how to calculate BILD scores efficiently , and illustrate their uses in Gibbs sampling optimization procedures , gapped alignment , and the construction of hidden Markov model profiles . BILD scores enable automated selection of optimal motif and domain model widths , and can inform the decision of whether to include a sequence in a multiple alignment , and the selection of insertion and deletion locations . Other applications include the classification of related sequences into subfamilies , and the definition of profile-profile alignment scores . Although a fully realized multiple alignment program must rely upon more than substitution scores , many existing multiple alignment programs can be modified to employ BILD scores . We illustrate how simple BILD score based strategies can enhance the recognition of DNA binding domains , including the Api-AP2 domain in Toxoplasma gondii and Plasmodium falciparum .
Protein and DNA sequence alignment is a fundamental tool of computational molecular biology . It is used for functional prediction , genome annotation , the discovery of functional elements and motifs , homology-based structure prediction and modeling , phylogenetic reconstruction , and in numerous other applications . The effectiveness of alignment programs depends crucially upon the scoring systems they employ to evaluate possible alignments . For pairwise alignments , scores typically are defined as the sum of “substitution scores” for aligning pairs of letters ( amino acids or nucleotides ) , and “gap scores” for aligning letters in one sequence with null characters between letters in the other . Substitution and gap scores may be generalized to multiple alignments , i . e . those involving three or more sequences . Most useful local pairwise alignment algorithms allow gaps and explicitly assign them scores [1]–[4] . However , many local multiple alignment algorithms do not allow gaps , or allow them only implicitly as spacers between distinct ungapped alignment blocks . Indeed the alignments recorded in some protein family databases are explicitly constructed with ungapped alignment blocks separated by variable length spacers [5] , and it has been argued that this formalism corresponds well to the observed relationships imposed by protein structure [6] . Short ungapped blocks are also used in the DNA context , to represent , for example , transcription factor binding sites . Many pairwise substitution scores have been developed for protein [7]–[20] and DNA [21] , [22] sequence comparison , and a statistical theory for substitution scores has been developed for local alignments without gaps [23] , [24] . It is not trivial to generalize pairwise scoring systems to multiple alignments , and the following four principal approaches have been proposed to this long-standing problem: A ) Tree scores . An evolutionary tree can be defined relating the sequences in question , with each sequence residing at one leaf of the tree . By reconstructing letters at the internal nodes of the tree , the score for an aligned column of letters is defined as the sum of pairwise substitution scores for all edges of the tree [25] , [26] . B ) Star scores . As a special case of tree-scores , a single “consensus” letter can be defined for an alignment column . The column score is defined as the sum of pairwise scores for the consensus letter to each letter in the column . The tree in question reduces to a star , with the consensus at the central node . C ) Sum-of-the-Pairs or SP scores . A column score can be constructed as the sum of substitution scores for all pairs of letters in the column [27] , [28] . D ) Entropy scores . Scores can be based on the entropy of the letter frequencies observed in a column [29]; these scores have become particularly popular for DNA alignments . All these approaches are open to refinement , for example by weighting the pairwise scores of the sequences involved . All reasonable substitution scores for pairwise local alignment are implicitly log-odds scores [23] , [30] , which compare the probabilities of aligning two letters under models of relatedness and non-relatedness , and the most popular are explicitly so constructed [7] , [8] , [14] . We argue that multiple alignment column scores should be similarly constructed , based upon explicit target frequency predictions for columns from accurate alignments of related sequences . For this purpose , we propose , the method with the strongest theoretical foundation relies upon the specification of a Bayesian prior , over the space of multinomial distributions for describing alignment columns representing true biological relationships [31] , [32] . We call column scores based on such a formalism “Bayesian Integral Log-odds” or BILD scores . Although these scores are implicit in earlier work , their full generality and utility has not been recognized . They may be calculated efficiently , and may be generalized to allow for the differential weighting of sequences in a multiple alignment . We also consider an alternative approach that allows log-odds column scores to be derived from any pairwise substitution matrix . Given their form , multiple alignment log-odds scores can be used directly to define the proper extent of multiple alignment blocks , and to derive natural scores for profile-profile comparison . We show that they also arise from the perspective of the Minimum Description Length Principle [33] , which allows them to be combined naturally with other information theoretic measures . Other direct applications are specifying when a sequence should be included in a multiple alignment at all , and when an alignment of many related sequences is better split into several alignments each involving fewer sequences . Efficient methods for calculating BILD scores allow them to be incorporated into Gibbs sampling algorithms for ungapped local multiple alignment . Most practical protein applications , however , require provisions for gaps . We describe two methods for extending an ungapped local multiple alignment produced by the Gibbs sampling strategy to a gapped alignment , the first using asymmetric affine gap costs , and the second hidden Markov models . In the latter , column BILD scores inform the construction of position-specific gap costs , and yield gapped alignments in greater conformity with considerations of protein structure . We illustrate the applications of the programs by using them to uncover previously undescribed Api-AP2 domains of Toxoplasma gondii and Plasmodium falciparum . Multiple sequence alignment comprises a diverse set of problems and approaches . Many sophisticated statistical inference techniques have been applied to the multiple alignment problem and to the related problem of phylogenetic reconstruction , e . g . [34]–[37] . It is not our purpose here to develop a new multiple alignment program . Rather , we seek only to argue that the “substitution scores” for multiple alignment columns which lie at the core of most multiple alignment methods can in many cases be improved . Although many statistical alignment methods are Bayesian-based , the BILD scores directly implied by Bayesian reasoning have been heretofore unrecognized .
Log-odds pairwise substitution scores can be written . Here , is the frequency with which residues and correspond in accurate alignments of related sequences , and is the background probability with which residue occurs . The base of the logarithm is arbitrary , and merely defines a scale for the scoring system . We henceforth assume that unless the natural logarithm is specified , all logarithms are base , and the resulting scores are therefore in the units of bits [30] . Note that no target frequencies are uniquely optimal for pairwise sequence alignment , because different are appropriate for comparing sequences diverged by different amounts of evolution [7] , [8] , [13] , [30] . This perception gives rise to families of substitution matrices , such as the PAM [7] , [8] and BLOSUM [14] series for protein comparison . To generalize log-odds scores to multiple alignments , we first develop some notation . We consider the alphabet from which the letters in our sequences are drawn to consist of elements , which for convenience we represent by the numbers 1 through . An ungapped column from a multiple alignment of sequences is a vector , each of whose components through takes on a value in . In essence , the log-odds approach compares two theories , one in which all the letters aligned are related or homologous , and the other in which none are . Each theory implies a probability for observing any given set of data . For the alignment column , we define as the probability of observing the data under the assumption of relatedness , and under the assumption of non-relatedness . Then the log-odds score for this column is defined as ( 1 ) Assuming background probabilities through for the various letters , is given simply by ( 2 ) We will consider one primary strategy for deriving . As with pairwise scores , all sets of multiple alignment column scores with negative expected value are implicitly log-odds scores [23] , [30] . However , unless their values for are explicitly constructed in a sensible way , log-odds scores are unlikely to perform well in the applications suggested below . For alignments of more than two sequences , there are of course other possibilities than for all or none of the sequences to be related . However , as we will describe below , scores of the form of equation ( 1 ) can be applied to the comparison of sequences where only a subset are related , by adding indicator variables to include or exclude sequences . Log-odds scores for alignment columns immediately suggest substitution scores for aligning two different columns of letters . Specifically , letting be the concatenation of the vectors and , define ( 3 ) These column-column alignment scores may be used consistently in progressive alignment algorithms , which proceed by aligning the most closely related sequences first [38] , [39] , although as will be discussed below problems may arise in the definition of gap scores . They may also be used for profile-profile alignment , a topic of considerable recent interest [40]–[48] . For multiple alignments , perhaps the best approach to defining and calculating is a Bayesian one [31] , [32] . ( An alternative approach , based on pairwise scoring matrices , is described in Text S1 . ) Assume that the letters in a specific column from an accurate alignment of related sequences are generated independently , but with probabilities through that in general differ from the background probabilities . Assume further that it is possible to assign a prior probability distribution to the multinomial distributions associated with columns of related letters . This prior can be derived from a detailed study of related protein or DNA sequences . Although the data associated with a specific column generally have no temporal or other privileged order , assume for convenience that they are observed sequentially , in the order to . Then we may apply Bayes' theorem to transform the prior distribution to a posterior , after the observation of . More generally , each subsequent observation can be seen to transform the prior into a posterior distribution . We may then use the chain rule to write ( 4 ) The individual terms in this product may be calculated by integrating over all possible multinomial distributions : ( 5 ) Finally , combining equations ( 1 ) , ( 2 ) and ( 4 ) yields ( 6 ) We call scores defined in this way Bayesian Integral Log-odds or BILD scores . They can be understood simply as the sum of log-odds scores for the individual letters observed in a column , with the “target frequency” for each letter calculated based upon the prior distribution , and the “previously observed” letters through . Even though , by this formula , the log-odds score for a letter varies with its position in the column , the total column score is nevertheless invariant under permutation of the column's letters . BILD scores have some conceptual connections to star- and entropy-based multiple alignment scoring systems . The simplest generalization of star scores imposes a prior probability distribution on the consensus letter , but still assumes a probabilistic pairwise substitution model . As we describe in Text S1 , this yields a class of log-odds scores we call MELD scores . BILD scores arise , in contrast , by thinking of the “consensus” not as an ancestral letter , but rather as a generative probabilistic model , and by integrating over a prior distribution placed on this model . Given observed and background letter distributions and , entropy scores have been defined variously , and conceptually distinctly , as: i ) , the entropy difference between and ; ii ) , the entropy difference between a uniform distribution on letters and ; and iii ) , the relative entropy of and . Definitions i ) and ii ) differ only by a constant . One may refine any of these definitions by taking to be a posterior letter distribution , derived from a prior and a set of observations . Both BILD and entropy-based scores can be viewed as the sum of scores derived from the probabilities for individual observations . The central distinction is that BILD scores estimate the probability for a given such observation using only “earlier” ones , whereas entropy scores estimate this probability using the complete collection of observations . Although the definition of BILD scores is valid for any prior distribution one wishes to specify , it is in general impractical to calculate the , or the integral in equation ( 5 ) , except when takes the form of a Dirichlet distribution [49] , or a mixture of a finite number of Dirichlet distributions [31] , [32] . In this case , as described below , all the are also Dirichlet distributions , or Dirichlet mixtures , and is easily calculated . Therefore , for mathematical as opposed to biological reasons , we always assume that BILD scores are defined using a Dirichlet or Dirichlet mixture prior . The family of Dirichlet mixtures , however , is rich enough that it can capture well much relevant prior knowledge concerning relationships among the various amino acids or nucleotides . We review here the essentials of Dirichlet distributions . A multinomial distribution on letters is specified by an -dimensional vector , within the simplex defined by , and . The requirement that the sum to 1 renders the space of multinomials dimensional . A Dirichlet distribution , defined over this space , is parametrized by an -dimensional vector with all positive . We shall sometimes refer to such a distribution by its parameters , and we define as the sum of the . The Dirichlet distribution is given by the probability density function ( 7 ) where the normalizing scalar ensures that integrating over its domain yields 1 . Here , is the Gamma function , and for positive integral . The uniform density is a special case that arises when all the are 1 . Dirichlet distributions have two convenient properties . First , the expected frequency of letter implied by is . Second , the posterior distribution yielded by Bayes' theorem , after the observation of the letter , is a Dirichlet distribution with , but with all other parameters equal to those of . To illustrate how to calculate BILD scores using these properties , consider the case of DNA comparison ( with the numbers 1 through 4 identified respectively with the nucleotides A , C , G and T ) , with uniform background probabilities , and a Dirichlet prior given by the parameter vector ( 1 , 1 , 1 , 1 ) . By equation ( 4 ) , the target frequency associated with the alignment column “AATC” is given by . Thus the score for the column is bits . In contrast , for the column “AAAC” , , and the score for this column is bits . The essence of a Dirichlet distribution is perhaps best understood through the alternative parametrization ( ; ) , where , and . Because the must sum to 1 , there are still only independent parameters . The vector describes the center of mass of the distribution , while indicates how concentrated the distribution is about this point . Large values of correspond to distributions with most of their mass near , whereas values of near 0 correspond to distributions with most of their mass near the boundaries of the simplex . It is frequently sensible , although not necessary , to choose a prior whose is identical to the background frequencies . In this case , , and the first summand in equation ( 6 ) is always 0 . In other words , no letter in a column , considered in isolation , carries any information as to whether the column represents a true biological relationship . Single Dirichlet distributions frequently are adequate for capturing prior knowledge concerning “true” alignment columns of related DNA sequences , but this is not the case for proteins . Most simply , distinct regions of multinomial space , representing different collections of amino acids , should have high prior probabilities . In order to address the deficiency of single Dirichlet distributions , Brown [31] proposed the use of Dirichlet mixture priors . A Dirichlet mixture is simply the weighted sum of distinct Dirichlet distributions . It is specified by positive “mixture parameters” through that sum to 1 , and a set of standard Dirichlet parameters , through , for each of the component Dirichlet distributions . ( It will be useful later to define as . ) In all , because of the restriction on the sum of the , a Dirichlet mixture has independent parameters . The Dirichlet components of a mixture generally are thought of as describing various types of positions ( e . g . hydrophobic , charged , aromatic ) typically found in proteins . Bayes' theorem implies that , given a -component Dirichlet mixture as a prior , the posterior distribution after the observation of a single letter is also a -component Dirichlet mixture [31] , [32] . Brown [31] proposed Dirichlet mixture priors in the context of deriving “substitution” scores for aligning amino acids to columns from a multiple protein sequence alignment . This restricted context can be understood as comprehending a single summand from equation ( 6 ) . BILD scores extend Brown 's sequence-profile alignment scores to comprehensive scores for multiple alignment columns . Generalizing the development above , we describe here how to calculate the probability of a particular observation given a Dirichlet mixture prior , and how to calculate the posterior resulting from this observation . First , given a Dirichlet mixture , with parameters and , the probability of observing letter is given simply by ( 8 ) which follows directly from the definition of Dirichlet mixtures , and the result for single Dirichlet distributions . Second , given the observation of letter , and a Dirichlet mixture prior parametrized as above , the parameters and of the posterior distribution may be calculated as follows: ( 9 ) In short , first multiply the mixture parameters by the Bayesian factors and normalize , and then add 1 to each . Mathematics establishing the validity of this procedure appears in [32] . Their development is more complex than we require here , because we modify the Dirichlet mixture parameters only one observation at a time . We note that given the and , it is simple to invert procedure ( 9 ) to determine the and . This is useful for applications such as the Gibbs sampling algorithm discussed below . Many multiple alignment problems involve subsets of sequences that are much more closely related to one another than to the other sequences being considered , and this may yield suboptimal results , because a large number of closely related sequences can “outvote” a few more divergent sequences . One remedy has been to assign each sequence a numerical weight , with closely related sequences down-weighted [50]–[61] . Also , subsumed in such weights may be the recognition that the total number of effective observations represented by an alignment column may be smaller than the number of sequences it comprehends [4] , [62] , [63] . Thus , for certain applications it may be desirable to generalize BILD scores to weighted sequences . To do so , we need to define the concept of the probability of a “fractional observation” of a letter , and describe as well how a posterior distribution is calculated after such a fractional observation . Arguments supporting how this may be done can be extracted from the mathematical development in [32] . Both equation ( 8 ) and the first step of procedure ( 9 ) involve multiplication by the factors . For the fraction of an observation of letter , these factors must be replaced by the alternative factors ( 10 ) Also , in the last step of procedure ( 9 ) , the quantity rather than 1 must be added to each . The factors are identical to the original factors when , and all approach 1 as approaches 0 , as some reflection shows they must . Finally , note that equation ( 10 ) may be applied to as well as , and may be useful even when all observations are unitary . Thus , by aggregating observations , the BILD score for a column containing unique letters may be calculated with summands , rather than the summands of equation ( 6 ) . For a single Dirichlet prior , reduces to the simple formula ( 11 ) where is the count of letter , and is the total count of all residues . Only the numerator inside the product varies from column to column within an alignment , yielding further efficiency for calculation . Only the research team that first proposed Dirichlet mixtures for protein sequence comparison has derived , from analyses of large protein alignment collections , sets of Dirichlet mixture prior parameters [31] , [32] . Twelve such sets , involving various numbers of Dirichlet components , can currently be found at http://compbio . soe . ucsc . edu/dirichlets/index . html . We list five of these in Table 1 , which we refer to as through . Proteins diverged by different degrees of evolutionary change are best studied using pairwise substitution matrices with different relative entropies [30] , and the analogous claim should hold for Dirichlet mixture priors . A Dirichlet mixture prior implies a background amino acid frequency distribution , as well as a symmetric pairwise substitution matrix , by means of the formula . The relative entropies of the substitution matrices implicit in the priors through range from 1 . 44 bits , roughly equivalent to that of the PAM-80 matrix [7] , [8] , which is appropriate for fairly close evolutionary relationships , to 0 . 18 bits , roughly equivalent to that of the PAM-360 matrix , which is appropriate only for extremely distant relationships ( Table 1 ) . As well as , one may calculate the mean relative entropy of the multinomial distributions described by a Dirichlet mixture prior to the background frequencies ( see Text S2 ) . For to , ranges from to bits ( Table 1 ) . That has a much greater value than indicates that on average much more information is available per position from an accurate multiple alignment of many related sequences than from a single sequence . We note that , in lieu of using different priors , the effective relative entropy of a particular Dirichlet mixture may be tuned by scaling the weights of the sequences to which it is applied [43] . Standard pairwise substitution matrices are constructed from sets of proteins with certain background amino acid frequencies , and are non-optimal for the comparison of proteins with compositions that differ greatly from [64] . Similarly , a Dirichlet mixture prior has an implicit background amino acid composition , and should not be optimal when applied to proteins with compositions that differ greatly from . It is possible to adjust standard matrices for use with non-standard compositions [64] , [65] , and we will discuss elsewhere an analogous strategy that can be applied to adjust Dirichlet mixture priors . Single Dirichlet priors may be appropriate for DNA sequence comparison . The uniform density , arising when all ( ) , has frequently been advocated in the absence of prior knowledge , and “Jeffreys' prior” [66] , which is uninformative in a deeper sense , corresponds to all ( ) [33] . When specific prior knowledge concerning an application domain is available , however , there is generally not a strong argument for using uninformative priors . For related DNA sequences , the columns of accurate alignments are sometimes dominated by one or two nucleotides , suggesting that all should be smaller than . Furthermore , it usually makes sense for the to be proportional to the background frequencies . If this is stipulated , the specification of a Dirichlet prior reduces to the specification of . Assuming a uniform nucleotide composition , the values of and implied by from to are given in Table 2 . An empirical study of transcription factor binding sites [67] concludes that , at least for the analysis of such sites , should be or lower . A direct application of multiple alignment log-odds scores is to determining local alignment width . As formulated by Smith and Waterman [1] , an optimal local alignment is one that maximizes an alignment score but is of arbitrary width . Such scores should fall on the log side of the “log-linear phase transition” [68] , which implies that for ungapped local alignments , substitution scores must be of log-odds form [23] , [30] . Equation ( 1 ) explicitly generalizes pairwise log-odds scores to the multiple alignment case . They are positive for some alignment columns , negative for others , and must have negative expected value . Therefore it is appropriate to define an optimal ungapped multiple alignment as one with maximal aggregate log-odds score . This immediately allows one to define the proper width or extent of an ungapped multiple DNA or protein alignment , without resorting to the ad hoc principles frequently required for other scoring systems [69] . Although the Smith-Waterman algorithm can be applied to optimize log-odds-scored local multiple alignments , it is too slow for most purposes . Nevertheless , once relative offsets have been fixed for a set of sequences , it is trivial to determine an optimal ungapped local multiple alignment along the single implied diagonal . The ungapped local multiple alignment problem may be formulated as seeking segments of common width within multiple DNA or protein sequences that , when aligned , optimize a defined objective function . We take this function here to be the aggregate log-odds score for the aligned columns . One way to approach this optimization is by means of a Gibbs sampling strategy , as described by Lawrence [69] . Log-odds scores can be used to adjust dynamically , by applying the Smith-Waterman algorithm to the diagonal implied by a provisional alignment , without the need for an arbitrary parameter or an ad hoc optimization . They may also be used to determine dynamically whether or not a sequence should participate in the multiple alignment at all , for which purpose it is useful first to consider log-odds scores from the perspective of the Minimum Description Length Principle . The Minimum Description Length ( MDL ) Principle provides a criterion for choosing among alternative theories for describing a set of data [33] , [49] . To simplify greatly , it suggests that given a set of alternative theories to describe a set of data , that theory should be chosen which minimizes , defined as the sum of , the description length of the theory , and , the description length of the data given the theory . By convention , description lengths are measured in bits . From information theory [70] , the information associated with an event of probability is bits . Focusing on actual encoding schemes for probabilistic events can unduly complicate MDL analyses . Accordingly , we here follow the approach of section 3 . 2 . 2 of [33] , in which description lengths are allowed to be non-integral , and are identified with negative log probabilities . Thus , if the data can be described probabilistically , . The length of the theory is defined as the number of bits needed to specify the free parameters of , i . e . those that are fitted to the data [33] . For local multiple alignment , the theory that the input sequences are unrelated has only the background probabilities as parameters , whose description length we will call . The data is comprised of sequences , with lengths through , and consisting of the letters . Then . The theory states that segments of width beginning at positions within the various sequences are related , and that the probability of the data within each column of the implied alignment is ; the probability of the rest of the data may be described with the background frequencies . The free parameters are , the vector of starting positions , and . Each may take on one of values , so its description length is approximately , if is not too large compared to . Thus , we have , where is the description length of . ( If all feasible widths are taken to be equally likely , is just . Other encodings have grow slowly with [33] , [49] . ) It is apparent that , where the latter sum is taken only over those letters not participating in the local multiple alignment . Everything simplifies when we consider the difference in the total description lengths of the two theories: ( 12 ) where is simply the log-odds score for the implied alignment . In other words , is preferred whenever exceeds . As described in Text S3 , this prescription is related to the statistical theory for ungapped local alignments [23] . To allow one or more sequences to be excluded from the multiple alignment , we consider not 2 , but theories , distinguished by binary indices , which take on the value 1 to indicate that sequence participates in the alignment , and otherwise . These theories need not be a priori equally likely; if necessary , for from 1 to we can specify prior probabilities that sequence contains a segment related to segments in the other sequences . Let us consider the difference in the description lengths of two theories , and , that differ only in their index . Theory incurs the cost for the prior probability that , and also requires describing the location of the related segment , which costs bits . In contrast , theory incurs only the cost , so costs more bits to describe than . Thus , for to be preferred , the log-odds score of the multiple alignment must increase by at least when the segment from the th sequence is added . If is close to 1 , can be negative , and is if . In short , the greater the prior probability that a given sequence contains a relevant segment , the lower the score of such a segment need be for inclusion in the alignment . The change in the log-odds score with the addition of a segment from the th sequence depends upon which other sequences , and which of their segments , participate in the alignment . Consequently , the values of the indicator variables must be part of the larger optimization , and their selection can be readily incorporated into a Gibbs sampling algorithm . The MDL Principle can also be extended to the case where a single sequence may contain more than one copy of a pattern , and , as previously described [62] , [71] , [72] and discussed in Text S4 , to the clustering of multiple alignments into subfamilies . Although our central concern is to define a new type of multiple alignment substitution score , many important applications require the construction of gapped multiple alignments , and these generally entail scores for insertions and deletions . Multiple alignment gap scores should be defined in a manner consistent with the substitution scores used [73] , so we will consider what type gap scores might fruitfully be paired with BILD scores . Just as the log-odds perspective places pairwise substitution scores in a probabilistic framework [7] , [8] , [23] , [30] , so pairwise gap scores can be viewed as specifying probabilities for insertions and deletions within biologically accurate alignments [74]–[82] . For pairwise alignments , “affine” gap scores , of the form for a gap of length [83]–[85] , are those most commonly used [3] , [4] , although more complex gap scores have frequently been proposed [86]–[89] . When there is an essential asymmetry between the sequences being aligned , differing scores may be assigned to gaps within the two sequences . Furthermore , when substitution and gap scores are properly integrated and both expressed in the units of bits , the two parameters of affine gap scores can be understood to specify jointly the average frequencies and lengths of gaps in the alignments sought [82] . If gaps are to be introduced into the BILD score formalism , an immediate problem is which , if any , letters from individual sequences should be understood as insertions with respect to the “canonical” pattern . In other words , it appears a canonical width for the multiple alignment must somehow be chosen , with respect to which gaps arising in the alignment of individual sequences can be assessed . For simplicity , suppose we have a “canonical” multiple alignment , i . e . one with a specified number of columns , to which we wish to align a single sequence , to produce a new multiple alignment . It is reasonable to define the alignment score of as the pre-existing alignment score for plus the incremental pairwise score for aligning and . This pairwise alignment involves substitutions ( letters from aligned to columns from ) , insertions ( runs of letters from that are not aligned to any columns from ) , and deletions ( runs of columns from that are not aligned to any letters from ) . BILD scores for the columns of arise naturally when one defines the substitution scores for aligning to as incremental BILD scores . It remains then only to define gap scores for insertions and deletions in the alignment of and . There is an essential asymmetry in gap scores for aligning to , relevant in many biological applications . For proteins , the columns of represent canonical positions , present in most sequences of a protein family , and it should accordingly be very costly to delete any of these columns . In contrast , individual proteins often contain long loops not present in the great majority of related sequences [90] , [91] , so even long insertions should not be very costly . Uniform but asymmetric affine insertion and deletion scores can capture this simple idea , and we have implemented them in one program described in the Results section below . These scores can be derived from the average frequencies and lengths [82] of insertions and deletions with respect to canonical protein family multiple alignments . Just as incremental BILD substitution scores change as more sequences are added to a multiple alignment , so it is possible to let insertion and deletion scores change as well , and vary by position . In the context of Hidden Markov Models [76]–[81] , many methods for doing this have been described . Below , we implement one simple procedure that depends only upon the BILD scores of multiple alignment columns , and not upon the relatively sparse gaps observed in any particular alignment . Formula ( 3 ) permits BILD substitution scores to be used for progressive multiple alignment . However , as described above , gaps scores pose a particular problem , because to define insertions and deletions one needs to construct a canonical alignment , and this is difficult for a small number of sequences . For example , when just two proteins are aligned , it is quite possible that gaps in both sequences would ultimately be seen as insertions with respect to a model describing the whole protein family , but there is no obvious way to determine this in advance . ( The problem does not arise when substitution and gap scores are defined using the sum-of-pairs or SP formalism [27] , [28] , for which no canonical alignment is necessary [73] . ) Accordingly , the approach we take below is eschew gaps at first , and thereby construct a canonical multiple alignment whose columns represent positions present in the majority of sequences . Only then do we realign individual sequences to this model , allowing gaps . There has been considerable recent interest in aligning profiles that describe different protein families [40]–[48] . If BILD substitution scores , defined by equation ( 3 ) , are to be used for this purpose , it would seem that we face the same problem for gaps that we do for progressive multiple alignment . Specifically , an insertion with respect to one profile is seen as a deletion with respect to the other , so how may one determine which , if either , perspective to adopt in a model describing both ? However , so long as this goal is only to compare pairs of profiles , and not to proceed further , this problem may be elided . It is consistent to define pairwise gap costs for the alignment of two profiles , just as one would for the alignment of two sequences , without reference to a canonical alignment , and the substitution scores of equation ( 3 ) can be used sensibly with such gap costs . The gap costs chosen may depend upon the profiles being aligned , and may therefore be asymmetric and position specific . We leave for elsewhere the comparative evaluation of profile-profile alignment using substitution scores defined by equation ( 3 ) , and those defined in other ways [40]–[48] .
BILD scores find perhaps their purest application in the ungapped local alignment problem described above , so it is worth studying them in this restricted context . The Gibbs sampling approach to finding optimal local multiple alignments was introduced by Lawrence et al . [69] , and this algorithm can easily be modified to employ BILD scores . Potential advantages are improved sensitivity and the automatic definition of domain boundaries . Evaluation ideally requires a set of proteins with ungapped domains whose correct alignment is structurally validated , but such sets are unfortunately very rare . Nevertheless , the collection of ungapped helix-turn-helix ( HTH ) domains in [69] provides a limited test set for analyzing BILD scores in the absence of gaps . As we describe in Text S5 , with Tables S1 and S2 , BILD scores achieve success on two fronts . First , they have greater average sensitivity than the entropy-based scores proposed by Lawrence et al . [69] , in yielding accurate alignment from fewer sequences; second , they recognize with good precision the extent of the structurally-defined domains , and therefore do not require a prior specification of alignment width . Local multiple alignment programs generally must allow for gaps , either implicitly or explicitly . However , even for aligning gapped domains , the search for ungapped local alignments can be a fruitful first step . BILD scores can play an important role at this stage in defining the common core of a protein family , and can be adapted in subsequent stages to score gapped multiple alignments . As a proof of principle , we here develop a relatively simple algorithm , Program 1 , that uses BILD scores as part of a gapped multiple alignment strategy . We describe this program's architecture and motivation below , and use a standard artificial test set to evaluate its ability to recognize the boundaries of local motifs , and to properly construct gapped local alignments . We then describe in section C how Program 1 may be refined through the consideration of features of protein structure , and illustrate the application of our methods to the delineation of a protein domain family . As mentioned above , real protein domains are subject , on average , to much longer insertions than deletions , and this implies the utility of asymmetric affine gap costs for Program 1 . The particular costs that are best will depend upon the statistical properties of gaps , and a possible refinement of Program 1 would be to adjust gap costs dynamically . From the analysis of a variety of protein families , we have found empirically that reasonable gap scores to use in conjunction with Dirichlet mixture priors are bits for a deletion of motif positions ( corresponding [82] to an initiation frequency per motif position of 0 . 28% , and a mean length of 2 . 0 ) , and bits for an insertion of length into the motif ( corresponding to a frequency of 0 . 87% , and a mean length of ) . Protein structure implies more than an asymmetry between the frequency and length statistics of insertions and deletions . Reflecting the evolution of secondary structure elements and loops , certain motif positions are much less likely to be deleted than others and , similarly , insertions are much less likely to occur between certain pairs of motif positions than others . We describe below an extension of Program 1 to an HMM-based Program 2 that relies only upon column BILD scores to calculate position-specific gap score parameters . We then apply Programs 1 and 2 to the detection of Api-AP2 domains .
We have described a natural generalization of log-odds substitution scores for pairwise alignments to substitution scores for multiple alignment columns . Multiple alignment log-odds scores probably are best derived using a Bayesian approach , yielding what we have called BILD scores . Log-odds scores imply scores for aligning multiple alignment columns to one another , or for aligning multiple alignment columns to single sequences , and it was in this latter context that the Bayesian approach was first formulated by Brown [31] . In conjunction with the Minimum Description Length Principle , log-odds scores provide a means for determining the proper width or extent of a local multiple alignment , and for deciding whether a segment should be included in the alignment . They may also be used to cluster a set of related segments into subclasses; see Text S4 and [62] , [71] , [72] . One may compute rapidly the BILD score for a multiple alignment column , as well as the new score that results from the addition or subtraction of a single letter . This permits BILD scores to be used practically in Gibbs-sampling local multiple alignment programs . They can improve the performance of such programs , and remove the need for specifying the width of a pattern sought . The proper description of protein domains in most cases requires a provision for gaps . We have implemented two relatively simple programs for extending a core ungapped pattern or profile to a gapped local multiple alignment . There are several key elements to our approach . First , the initial maximization of aggregate BILD scores using Gibbs sampling yields a core pattern and pattern length for further refinement . Second , the semi-global alignment of this pattern to the input sequences recognizes the importance of complete occurrences of the pattern . Third , the use of asymmetric affine gap costs ( Program 1 ) recognizes that , with respect to the core pattern , long deletions generally are much more deleterious than long insertions . The placement of gaps can be refined using position-specific gap costs derived from column BILD scores ( Program 2 ) . Fourth , greedy alignment allows multiple instances of a pattern to be found within a single sequence . In conjunction with length-dependent gap costs , it discourages alignments spanning more than one instance of a pattern , but can still uncover long insertions . Fifth , iteration permits the core model to be refined , improving the discrimination of true relationships from chance similarities . This strategy , informed by considerations of protein structure , has proved a rapid and effective method for delineating protein families . Although our programs were developed only for research purposes , with the limited goal of testing the impact of BILD scores , their code is available upon request . We have sought here primarily to describe the construction and potential uses of log-odds scores in the multiple alignment context . However , many avenues for further research , involving the development and benchmarking of complete multiple alignment programs , remain . To what extent can BILD scores improve the accuracy of profile-profile comparison programs ? How does Erickson-Sellers semi-global alignment [92] , with uniform asymmetric affine gap costs , compare to HMM [80] , [81] and other methods [6] in recognizing related sequence in database searches ? We look forward to investigating some of these questions .
|
Multiple sequence alignment is a fundamental tool of biological research , widely used to identify important regions of DNA or protein molecules , to infer their biological functions , to reconstruct ancestries , and in numerous other applications . The effectiveness and accuracy of sequence comparison programs depends crucially upon the quality of the scoring systems they use to measure sequence similarity . To compare pairs of DNA or protein sequences , the best strategy for constructing similarity measures has long been understood , but there has been a lack of consensus about how to measure similarity among multiple ( i . e . more than two ) sequences . In this paper , we describe a natural generalization to multiple alignment of the accepted measure of pairwise similarity . A large variety of methods that are used to compare and analyze DNA or protein molecules , or to model protein domain families , could be rendered more sensitive and precise by adopting this similarity measure . We illustrate how our measure can enhance the recognition of important DNA binding domains .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"computational",
"biology/macromolecular",
"sequence",
"analysis",
"computational",
"biology/protein",
"homology",
"detection",
"computational",
"biology/sequence",
"motif",
"analysis",
"computational",
"biology"
] |
2010
|
The Construction and Use of Log-Odds Substitution Scores for Multiple Sequence Alignment
|
In veterinary parasitology samples are often pooled for a rapid assessment of infection intensity and drug efficacy . Currently , studies evaluating this strategy in large-scale drug administration programs to control human soil-transmitted helminths ( STHs; Ascaris lumbricoides , Trichuris trichiura , and hookworm ) , are absent . Therefore , we developed and evaluated a pooling strategy to assess intensity of STH infections and drug efficacy . Stool samples from 840 children attending 14 primary schools in Jimma , Ethiopia were pooled ( pool sizes of 10 , 20 , and 60 ) to evaluate the infection intensity of STHs . In addition , the efficacy of a single dose of mebendazole ( 500 mg ) in terms of fecal egg count reduction ( FECR; synonym of egg reduction rate ) was evaluated in 600 children from two of these schools . Individual and pooled samples were examined with the McMaster egg counting method . For each of the three STHs , we found a significant positive correlation between mean fecal egg counts ( FECs ) of individual stool samples and FEC of pooled stool samples , ranging from 0 . 62 to 0 . 98 . Only for A . lumbricoides was any significant difference in mean FEC of the individual and pooled samples found . For this STH species , pools of 60 samples resulted in significantly higher FECs . FECR for the different number of samples pooled was comparable in all pool sizes , except for hookworm . For this parasite , pools of 10 and 60 samples provided significantly higher FECR results . This study highlights that pooling stool samples holds promise as a strategy for rapidly assessing infection intensity and efficacy of administered drugs in programs to control human STHs . However , further research is required to determine when and how pooling of stool samples can be cost-effectively applied along a control program , and to verify whether this approach is also applicable to other NTDs .
The soil-transmitted helminths ( STHs ) Ascaris lumbricoides , Trichuris trichiura , and the two hookworm species , Necator americanus and Ancylostoma duodenale , cause the highest burden among all neglected tropical diseases ( NTDs ) , with school-aged children and pregnant women being at highest risk [1]–[3] . Preventive chemotherapy ( PC ) programs , in which albendazole ( 400 mg ) or mebendazole ( 500 mg ) administered in a single dose are the principal means of control of STH infections in school-aged children , has recently received increased political and scientific attention [4] , [5] . The World Health Organization ( WHO ) has devised a roadmap to guide implementation of the policies and strategies set out in a global plan to combat NTDs ( period 2008–2015 ) , and more than 70 pharmaceutical companies , governments , and global health organizations committed to supporting this roadmap [6] in the London Declaration on NTDs in January 2012 by sustaining or expanding drug donation programs [7] . These pledges of drug donations are now in place . However , two factors that might affect the success of these programs have received little attention . First , the therapeutic efficacy of the two benzimidazoles ( albendazole and mebendazole ) differs across STH species [8] . Both drugs are highly efficacious against A . lumbricoides , but albendazole is more efficacious against hookworm , and both drugs are unsatisfactory when used as single regimen against T . trichiura infection , although mebendazole is relatively more efficacious [8] , [9] . Moreover , therapeutic efficacy can vary across levels of infection intensity , albendazole showing a high efficacy when the intensity of T . trichiura is low and poor efficacy when infection levels are high [10] . Second , we are relying on two drugs with the same mode of action , and hence the emergence of anthelmintic resistance as drug donations expand , as substantiated in veterinary medicine , is likely [11]–[13] . For these reasons , it is important to seek for alternative strategies ( i ) to ensure appropriate choice of drug and regimen , ( ii ) to monitor anthelmintic resistance , and ( iii ) to assess the long-term impact of PC programs . Traditionally , both the assessment of infection intensity and drug efficacy are based on the examination of individual stool samples . However , this strategy impedes the up-scale of epidemiological surveys that is required to support health care decision makers to further maximize the efficiency of PC at nationwide level . A possible alternative to individual stool examination is the examination of pooled stool samples . Pooling samples ( e . g . , stool , serum , and urine ) of the same individual has been found valuable for diagnosis of various pathogens , including Giardia [14] , Chlamydia [15] , Salmonella [16] , and HIV [17] . Studies validating a pooling strategy for human STHs are lacking . In animal health it has been shown that pooling stool samples allows for a rapid assessment of infection intensity and drug efficacy . Pools of up to 10 animals provided estimates of intensity of helminth infections by means of fecal egg counts ( FECs ) comparable to those obtained by examination of individual stool samples [18] , [19] . However , it has also been suggested that pooling of animal stool samples may not be recommended when infections become more aggregated [19] . The effect of the number of samples pooled has not yet been examined . The main objective of the study reported in this paper was therefore to develop and to evaluate a sampling strategy based on pooling of stool samples . To this end , we assessed the intensity of STH infections across varying epidemiological settings and the efficacy of a single dose of mebendazole ( 500 mg ) on both individual samples and pooled samples ( pool sizes of 10 , 20 , and 60 individual stool samples ) . The ultimate aim was to facilitate rapid identification of STH infections in epidemiological studies and drug efficacy assessment .
Ethical approval was obtained from Ghent University ( 2011/374 ) , Belgium , and Jimma University ( RPGC/09/2011 ) , Ethiopia . The efficacy trial was registered under Clinical Trials . gov identifier B670201111554 . The school authorities , teachers , parents , and the children were informed about the purpose and procedures of the study . The written consent form was prepared in two commonly used local languages ( Afaan Oromo and Amharic ) and handed over to the children's parents/guardians . Only those children ( i ) who were willing to participate and ( ii ) whose parents or guardians signed the written informed consent form were included in the study . Moreover , an additional separate written informed consent form for children older than 12 years was prepared , read , and handed over to them and their additional written informed consent obtained . The study was conducted in Jimma Town , Ethiopia , located approximately 350 km southwest of the capital , Addis Ababa . Jimma Town is situated at a latitude and longitude of 7°40′N36°50′E , and is characterized by a semi-arid type climate with an average annual rainfall of 800–2 , 500 mm . The mean daily temperature is 19°C , and ranges from 12 to 30°C . It is located 1 , 720–2 , 010 m above see level . Our study focused on schoolchildren from age 5 to age 18 , across all eight grades . In total , there were 24 primary schools hosting a total of 23 , 492 children of all age groups of interest . The female/male ratio across the different schools was approximately 1∶1 ( Report Document 2011/2012 of Jimma Education Bureau ) . STH infections have been documented in Jimma Town , but at present no PC program to control STHs in school-aged children has been implemented . All stool samples were individually processed by the McMaster egg counting method , as described elsewhere [20] . McMaster is a flotation technique that is commonly used in veterinary parasitology both to assess intensity of gastro-intestinal parasite infections and to evaluate drug efficacy against these parasites . For the diagnosis and enumeration of STHs in public health , it has been found to be user-friendly ( vs . FLOTAC [23] ) , robust ( vs . Kato-Katz thick smear ) and accurate for enumeration of STHs , but less sensitive when intensity of infection is low ( vs . Kato-Katz and FLOTAC ) [20] . Briefly , 2 g of stool were suspended in 30 ml of saturated salt ( NaCl ) solution at room temperature ( density: 1 . 2 ) . The fecal suspension was poured three times through a tea sieve to remove large debris . After thorough mixing 10 times , 0 . 5 ml aliquots were added to each side of a McMaster slide chamber . Both chambers were examined under a light microscope using 100× magnification and the FEC , expressed as EPG for each helminth species , was obtained by multiplying the total number of eggs counted under the microscope by a factor 50 . A detailed tutorial can be found on http://www . youtube . com/watch ? v=UZ8tzswA3tc . In addition , in both studies , a subset of the stool samples was pooled in pools of 10 , 20 , and 60 individual samples . We considered pooling a rapid alternative to individual stool examination if at least 10 samples were pooled . Pools of 60 individual samples allowed for pooling all stool samples of one school in the study assessing infection intensity . For uniformity across the two studies , pooling of 60 stool samples was also applied for the evaluation of the drug efficacy . Pools of 20 samples were considered as an intermediate between pools of 10 and 60 . The procedure for pooling individual samples is illustrated in Figure 3 , and is discussed in more detail below . A visualized tutorial can be found on http://www . youtube . com/watch ? v=IUZijtBABn0 . At first , 60 individual samples were randomly organized in 6 rows of 10 individual stool samples . From each row 1 g of each of the 10 individual stool samples was transferred into a new pre-labeled plastic beaker ( resulting in a total of 6 pools of 10 individual stool samples ) . After homogenization , 5 g from 2 plastic beakers representing pools of 10 individual samples were transferred into another new pre-labeled plastic beaker , resulting in a total of 3 pools of 20 individual samples . Next , 3 . 33 g was transferred from the 3 vials of pools 20 into new pre-labeled plastic beaker , resulting in 1 pool of 60 individual stool samples . Finally , each of the pools was processed by the McMaster egg counting method as done for individual samples . Homogenization was standardized by means of stirring the stool until homogenized . Stools from different subjects have different colors . We stopped stirring the pooled stool when the pool had one homogeneous color . This pooling procedure has two important advantages . First , the cascade procedure applied ( e . g . we pooled pools of 10 to make pools of 20 ) allowed for pooling samples into different pool sizes with only 1 g per individual sample . Second , it avoids the homogenization of too large quantities of stool . For example , for pools of 60 we only had to homogenize 10 g of stools ( 3 . 33 g of three pools of 20 individual stool samples ) , whereas this would have been 60 g if we had pooled 60 times 1 g of individual samples . For the assessment of infection intensity , samples were randomized according to age class ( 2 rows of 10 samples per age class ) . For the efficacy trial , small deviations from the aforementioned procedure should be noted . Although samples were randomly pooled at both baseline and follow-up , unforeseen dropouts meant that pools at baseline did not always match pools at follow-up , when pooling of exactly 10 , 20 , or 60 samples was not always possible . In addition , not all subjects were included at both baseline and follow-up . Quality of the parasitological examination was ensured by ( i ) analyzing the samples within an average of 4 hours , ( ii ) verification of density of the NaCl solution , ( iii ) verification of the sensitivity of the scale weighing the fecal material , ( iv ) supervision of the McMaster and pooling procedures , and ( v ) re-examination of 10% of the McMaster slides by a senior researcher . The total numbers of the individual samples and pooled samples across the assessment of the infection intensity and the efficacy trial are provided in Figures 1 and 2 , respectively .
The prevalence of STHs in the 14 primary schools was 52% . T . trichiura was the predominant species ( 39% ) , followed by A . lumbricoides ( 24% ) and hookworm ( 11% ) . The arithmetic mean FEC was 2 , 411 EPG ( 0–176 , 000 ) , 295 EPG ( 0–19 , 350 ) and 35 EPG ( 0–950 ) for A . lumbricoides , T . trichiura and hookworm , respectively . Across schools , there was large variation both in prevalence of any STH infection ( 11% to 73% ) and in prevalence of each of the three STHs ( 6% to 58% for T . trichiura; 0% to 43% for A . lumbricoides; and 0% to 30% for hookworm ) . Across the three age classes there was little variation in prevalence of STHs ( age class A: 50% to age class C: 49% ) , of T . trichiura ( age class A: 40% to age class C: 36% ) , and of A . lumbricoides ( age class A: 28% to age class C: 21% ) , but substantial variation for hookworm ( age class A: 8 . 0% to age class C: 16% ) . Overall , there was a significant positive correlation between mean FEC of individual samples and the FEC of the pooled samples for each of the three STH species ( RsA . lumbricoides = 0 . 91 , p<0 . 01; RsT . trichiura = 0 . 82 , p<0 . 01; Rshookworm = 0 . 68 , p<0 . 01 ) . As illustrated in Figure 4 , there was also a significant positive correlation between mean FEC of individual samples and the FEC of the pooled samples for each of the three pool sizes ( A . lumbricoides: Rs = 0 . 91–0 . 98 , p<0 . 01; T . trichiura: 0 . 75–0 . 85 , p<0 . 01; hookworm: Rs = 0 . 62–0 . 92 , p<0 . 01 ) . Table 1 summarizes the mean FEC for both individual and pooled samples . Overall , there were no significant differences in FEC between individual and pooled samples across the three STHs . Only for A . lumbricoides was a significant difference in FEC observed when pool sizes increased up to 60 samples , resulting in higher FECs ( FEC60 = 3 , 321 EPG vs . FECindividual = 2 , 411 EPG , p<0 . 01 ) . For the remaining two STHs , no significant difference across pool sizes was observed ( p>0 . 05 ) . The mean FEC and FECR for each of the STHs based on examination of individual and pooled samples are described in Table 2 . Based on individual samples , FECR was high for A . lumbricoides ( 97 . 2% ) , but only moderate for T . trichiura ( 60 . 9% ) and low for hookworm ( 44 . 2% ) . Pooled samples provided comparable FECR results for A . lumbricoides and T . trichiura . However , for hookworm , a significant statistical difference was found for pools of 10 ( p = 0 . 03 ) and 60 individual samples ( p = 0 . 02 ) .
Given the recent pledges of continuing donations of anthelmintic drugs [7] , and hence prospects of increasing drug pressure on parasite populations , cost-effective tools to guide healthcare decision makers on how to optimize treatment strategies and on how to monitor the control of STHs are urgently needed . In analogy with studies conducted in animal health , our results show that pooling stool samples also holds promise as a rapid strategy in public health ( i ) to assess infection intensity , ( ii ) to ensure appropriate choice of drug and regimen , ( iii ) to monitor anthelmintic resistance , and ( iv ) to assess the long-term impact of the ongoing PC programs to control STHs . However , before we can provide specific recommendations , further research is required to gain additional insights into how and when to apply pooling as described here . First , it is essential to assess the effect of pool size , sample size ( number of pools ) , the detection limit of the diagnostic method for quantifying infection intensity by means of FEC ( FEC method ) , and level of aggregation and intensity of infections on the precision and the accuracy of FEC and FECR results . This last assessment is particularly essential when level of infection and aggregation of STH infections change across different rounds of PC , and hence demand a different pool and sample size and FEC method . This effect of level of infection and aggregation of STH infections on factors inherent to the study design ( pool size , sample size and detection limit of the FEC method ) became already apparent in the present study , where pooling samples to assess drug efficacy using the McMaster egg counting method worked for A . lumbricoides and T . trichiura , but not for hookworm , for which the level of FEC was low and FECs were highly aggregated . Because it is impossible to thoroughly evaluate the impact of pool size , sample size , the detection limit of the FEC method , and level of aggregation and intensity of infections by field or laboratory experiments , a simulation study is in place . This approach will allow us to theoretically assess the accuracy and precision of FEC and FECR across different scenarios of pool size , sample size , detection limit of the FEC method , and level of aggregation and intensity of infections . Recently , such a simulation study has been performed for FECR based on individual stool samples [19] , [24] , which can be easily adapted for a pooling strategy . Second , a detailed cost-effectiveness analysis is highly recommended [25] . Examination strategies resulting in a comparable level of accuracy and precision on FEC or FECR may still require a different level of technical and financial support . The present study was not designed to verify the cost-effectiveness of our pooling strategy . However , from the current results obtained in a region endemic to STHs , pooling of stool samples might well be cost-effective . For example , in the present study we were able to reduce the samples examined by a factor of 10 without a significant loss in accuracy of the FEC results ( only for A . lumbricoides was a significant difference in FEC observed between individual samples and pools of 60 ) . We estimate that processing and reading a McMaster requires approximately 5 min [23] . By pooling 10 individual samples we would save 270 min per day ( = 60 individual samples ×5 min – 6 pools of 10×5 min ) . Of course , pooling samples requires some additional time , and pooling will be hard to justify as cost-effective when the pooling procedure requires more than 270 min . In the strategy used for the study described here we had to weigh a certain quantity of stool 60 times ( 60 individual samples to make 6 pools of 10 ) and to homogenize 6 pools . If we conservatively assume that homogenization of 1 pool demands 5 min , the pooling strategy will still be cost-effective when the quantity of stools can be measured within 240 min ( 270 – 6 pools of 10×5 min ) or 4 min per step of measuring stool ( 240/60 ) . Given that McMaster can be applied in 5 min and comprises weighing of 2 g of stool , homogenization in a flotation solution , and filling and reading of the McMaster slide , it is clear that the 4 min available to transfer a fixed quantity is also quite conservative . Third , various strategies for pooling stool samples should be evaluated . In the present study we pooled samples in a cascade: pools of 10 were made by pooling individual samples , but rather than repeating this procedure of pooling individual samples for the other pool sizes , we used the pools of 10 to make pools of 20 , and subsequently the pools of 20 to make the pools of 60 . This procedure provided an equal amount of stool pooled for each pool , in casu 10 g , and an elegant way to assess different pool sizes without too much additional work . However , it remains to be established that this procedure does not itself introduce any bias , particularly when the contribution of each subject decreases over pool size . For pools of 10 , each individual in our study contributed 1 g , whereas this was 0 . 5 g and 0 . 15 g for pools of 20 and 60 , respectively . Moreover , pools were homogenized by simple stirring . Homogenization in a liquid phase prior to examination , however , should be recommended , as this facilitates homogenization of pools . It is particularly important when eggs are not equally distributed among stool samples [26] . Homogenization of stool remains a crucial step in the most important FEC methods applied in veterinary parasitology , including McMaster , FECPAK ( www . fecpak . com ) and ( mini- ) FLOTAC [27; unpublished data] . An additional point is that homogenization will reduce drying of the stool while pooling samples . Due to our cascade procedure larger pool sizes were made at the end , but this probably resulted in an increase of evaporation of water from the stool samples . As a consequence of this increase , the mass of the stool decreased over time , whereas the number of eggs in the stool remained unchanged , hence the observed trend of increasing FECs over pool sizes . This potential bias in FECs could have been overcome by homogenizing the pools immediately in the flotation solution . Finally , to further simplify procedures under field conditions , it would be worth evaluating pooling based on a fixed volume rather than pooling a fixed amount of feces . Fourth , notwithstanding the comparable FECR results between individual and pooled stool samples for A . lumbricoides and T . trichiura in our study , assessing drug efficacy based on pooled samples remains a delicate matter . Subjects who are not infected ( truly or apparently ) at baseline cannot be excluded from the analysis and there is no perfect match of pools before and after drug administration due to drop out . Therefore , both analyses need to be validated independently . Fifth , we focused only on STH infections , but since the advocacy to integrate NTD control measures , this pooling strategy should also be validated for other NTDs , such as schistosomiasis . Finally , although pooling samples does not provide prevalence data , various models have been developed for other pathogens to estimate prevalence based on pooled samples . Validation of these models for STHs is required [28] . In conclusion , this study highlights that pooling stool samples is a rapid procedure that holds promise as a cost-effective strategy for assessing the intensity of STH infection and for monitoring PC programs . However , further research is required ( i ) to gain additional insights into the impact of pool size , sample size , detection limit of the FEC method , intensity , and aggregation of infections on the validity of pooling stool samples , ( ii ) to verify the cost-effectiveness of pooling , ( iii ) to optimize the methodology of pooling stool samples , and ( iv ) to validate models to estimate prevalence based on pooled samples .
|
Since the last decade , growing awareness of the control of neglected tropical diseases ( NTDs ) has resulted in worldwide increased pledges of drug donations . However , health care decision makers have a limited repertoire of strategies for a rapid assessment of infection intensity and for checking of drug resistance development . Therefore , we verified whether examination of pooled stool samples provide estimates of intestinal worm infection intensity and drug efficacy comparable to those obtained by examination of individual stool samples . Overall , the results showed that pooled samples provide comparable levels of infection intensity and drug efficacy . We conclude that pooling stool samples holds promise as a means of rapidly appraising the intensity of intestinal worm infections on a population level and of monitoring the efficacy of donated drugs . However , this study was conducted in an endemic region . Further research is required to determine when and how pooling of stool samples can be cost-effectively applied in a control program that is reducing the transmission of disease , and to verify whether this approach is also applicable to NTDs other than studied in this paper .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"infectious",
"diseases",
"global",
"health",
"epidemiology",
"drugs",
"and",
"devices",
"public",
"health"
] |
2013
|
Comparison of Individual and Pooled Stool Samples for the Assessment of Soil-Transmitted Helminth Infection Intensity and Drug Efficacy
|
Telomeric repeats preserve genome integrity by stabilizing chromosomes , a function that appears to be important for both cancer and aging . In view of this critical role in genomic integrity , the telomere's own integrity should be of paramount importance to the cell . Ultraviolet light ( UV ) , the preeminent risk factor in skin cancer development , induces mainly cyclobutane pyrimidine dimers ( CPD ) which are both mutagenic and lethal . The human telomeric repeat unit ( 5′TTAGGG/CCCTAA3′ ) is nearly optimal for acquiring UV-induced CPD , which form at dipyrimidine sites . We developed a ChIP–based technique , immunoprecipitation of DNA damage ( IPoD ) , to simultaneously study DNA damage and repair in the telomere and in the coding regions of p53 , 28S rDNA , and mitochondrial DNA . We find that human telomeres in vivo are 7-fold hypersensitive to UV-induced DNA damage . In double-stranded oligonucleotides , this hypersensitivity is a property of both telomeric and non-telomeric repeats; in a series of telomeric repeat oligonucleotides , a phase change conferring UV-sensitivity occurs above 4 repeats . Furthermore , CPD removal in the telomere is almost absent , matching the rate in mitochondria known to lack nucleotide excision repair . Cells containing persistent high levels of telomeric CPDs nevertheless proliferate , and chronic UV irradiation of cells does not accelerate telomere shortening . Telomeres are therefore unique in at least three respects: their biophysical UV sensitivity , their prevention of excision repair , and their tolerance of unrepaired lesions . Utilizing a lesion-tolerance strategy rather than repair would prevent double-strand breaks at closely-opposed excision repair sites on opposite strands of a damage-hypersensitive repeat .
Telomeric DNA consists , in all eukaryotes examined to date , of a tandemly repeated sequence located at each end of each chromosome . In humans , it is constituted of 5–10 kb of a repeated hexamer ( 5′TTAGGG/5′CCCTAA ) . Telomeres are required for chromosomal stability and integrity ( reviewed in [1] ) . Telomeres are hypersensitive to single-strand DNA damage induced by oxidative stress . This is thought to be due to the fact that sequences containing guanine triplets are highly sensitive to oxidation [2] , [3] . When inserted in a plasmid , telomere sequence is 7-fold more sensitive to Fe2+/H2O2-induced strand breakage than bulk sequence [2] . Moreover , breaks induced in telomeres are repaired significantly more slowly than in other sequences , including interstitial guanine rich repetitive sequence tracts; repair is still incomplete after 19 days compared to complete repair at 1 day elsewhere [4] . In addition , the oxidation of telomeric DNA contributes to their premature shortening . The frequency of oxidative DNA damage at the telomere correlates with the amount of telomere lost during subsequent rounds of DNA replication [5] . It was proposed that the telomere enters DNA replication with greater oxidative DNA damage than the rest of the genome and this elevated damage contributes to telomere shortening [6] . Contrasting with this hypothesis , however , it has been shown that telomere shortening induced by oxidative DNA damage can be replication independent [3] . Ultraviolet light-induced DNA damage has been used for decades as a model to study DNA damage induction and repair . It is biologically relevant because UV is a complete carcinogen , requiring no additional treatments for tumor development , and is the preeminent risk factor in skin cancer development . The vast majority ( >80% ) of UV-induced damage in B-form DNA consists of cyclobutane pyrimidine dimers ( CPD ) [7] , [8] . CPDs are intra-strand DNA lesions formed when two adjacent pyrimidines are joined across their 5–6 double bonds due to UV-excitation of one of them . The most frequent is the TT cyclobutane dimer [9] . These photoproducts are repaired by the nucleotide excision repair ( NER ) pathway , which nicks the DNA backbone and excises the damaged segment . Theoretically , the telomere sequence constitutes a perfect target for UV-induced DNA damage . First , the TT on the G-rich strand is repeated thousands of times in each chromosome . On the other strand , the 5′CCCTAA3′ would nominally generate low frequency CC and CT CPD , but two factors supervene: tracts of adjacent pyrimidines tend to generate multiple CPDs on the same molecule , due to cooperative denaturation of the helix by each successive CPD [10] and A:T tracts tend to transfer energy down the base stack until depositing it at a G:C pair [11] , [12] . These potential CCCT dimer tracts are again repeated thousands of times in each chromosome . These considerations suggested that this sequence might constitute a hotspot for UV-induced damage . The presence of potential hotspots on both telomeric strands then raises the following spectre: if the cell attempts to simultaneously repair two nearby CPDs on opposite strands , the twin incision nicks would mimic a double-strand DNA break [13]–[15] , triggering a DNA damage response and chromosome aberrations [16] , [17] . Studying DNA damage induction and repair in the telomere is challenging . The vast majority of the techniques used to study DNA damage induction and repair in a specific part of the genome are PCR based [18] . Because telomeres are constituted of repeated sequences , there are no unique PCR-primer sites . Mismatch primers have been developed to analyze human telomere length by quantitative PCR [19] . However , since those primers can bind to any repeat element of the telomere sequence , they cannot be used in standard techniques to study DNA damage induction and repair , which rely on having one or two known DNA ends . An older study used a single-enzyme modification of the telomere restriction fragment technique ( TRF ) to study UV-induced CPD in telomeres [20] . However , it is now known that the TRF technique does not provide information on the true length of telomeres [21]: restriction enzymes used to cleave non-telomeric DNA ( HinfI or RsaI ) give TRF lengths that depend on the site of restriction in the pre-telomeric region . The situation is exacerbated by the fact that achieving complete digestion of genomic DNA using a single restriction enzyme is challenging . Thus studying the induction of DNA damage using the TRF technique does not provide information exclusively about telomeres but about a mixture of telomeric and pre-telomeric DNA . Pre-telomeric DNA is now known to be one of the most rapidly-repaired regions of the genome [22] , skewing lesion measurements if this region is included . We developed a novel method , based on the chromatin immunoprecipitation technique ( ChIP ) , to study single-strand DNA damage . This technique , “immunoprecipitation of DNA damage” ( IPoD ) , allows the separation of damaged DNA from undamaged . The result is two fractions that can each be quantitated by PCR using primers specific for the gene under study . Previously developed primers specific for the human telomeric sequence [19] can be used in this technique , allowing the study of single-strand DNA damage induction and repair in this region . Using the IPoD technique , we have studied UV-induced CPD induction and repair in the telomere as well as in the p53 tumor suppressor gene , in 28S ribosomal DNA , and in a portion of mitochondrial DNA . We find that the telomere sequence is highly sensitive to the induction of CPD by UV light . Moreover , we show that the repair of those UV-induced CPD in telomeres is nearly absent .
IPoD is based on the ChIP technique [23] . Instead of immunoprecipitating a protein covalently cross-linked to DNA , IPoD directly immunoprecipitates DNA fragments containing a DNA structural alteration . Here we use the IPoD technique to study the CPD damage induced on a DNA strand by UV radiation [9] . The technique is schematized in Figure 1A . As the level of DNA damage in a specific region of the genome increases , the number of immunoprecipitated fragments from this region will increase . UV-irradiated DNA , but not unirradiated DNA , yielded an IP fraction using antibody against CPD but not with antibody to Bcl-xL protein or with antibody omitted ( Figure S1 ) . UVC has been used in this study to minimize the introduction of photosensitized oxidative DNA damage that accompanies UVB . The quantity of specific genomic DNA fragments present in the IP fraction was measured , after removing CPDs using photolyase , by PCR amplification using primers specific for the p53 tumor suppressor gene , the 28S ribosomal RNA repeat region , the CYTB gene of mitochondrial DNA , and telomeric DNA . The telomere sequence is composed of a 6-mer concatenated to greater than 5 kb , complicating the design of PCR amplification primers . A 5′ 21-mer primer composed of telomeric repeats is certain to have a complement on the 3′ primer , so primers will anneal together instead of annealing to the telomeric DNA target . Cawthon [19] describes telomeric primers containing mismatches that prevent primers from annealing to each other , thus achieving preferential annealing to telomeric DNA . Because the particular site at which any primer binds on the telomere sequence is random , the resulting PCR product is not a sharp band but a smear . For the exponential PCR process to be used quantitatively , it must contain an internal control , as in real-time PCR , or be carried out so that all samples have been amplified by the same factor of 2n , that is , with all samples lying on the log-linear part of the amplification curve so that they can be compared to a calibration curve . No internal control is possible with IP , so we adjusted the amount of starting DNA material and the number of PCR cycles to achieve log-linearity for each primer . Figure 1B ( upper two panels ) shows that the signal from the PCR amplified IPoD-immunoprecipitated DNA is proportional to the UVC dose for 3 different genomic regions . Each genomic region's signal is normalized to that region's signal at 20 J/m2 . The signal was linear up to 30 J/m2 UVC ( Figure S2 ) . Above this dose range , the slope decreased . Doses above 20 J/m2 UVC are lethal so the present experiments did not enter that range . The high-dose slope reduction could be due to sustaining more than 1 CPD per DNA fragment , saturating the anti-CPD antibody with CPDs , or depleting PCR reagents . Linearity at doses below 30 J/m2 UVC indicates that: a ) CPDs are not missed because they occur in DNA segments that will already be IPd due to another CPD; b ) the many telomere copies do not saturate the PCR reaction; and c ) CPDs or ( 6-4 ) photoproducts remaining in the fragment during PCR do not cause a dose-dependent dropout of sample . To confirm the last point , we also amplified the IP fraction without first reversing remaining CPD with photolyase ( Figure 1B , lower panel ) . When normalized to the signal at 20 J/m2 , the shape and slope of the dose-response curve were unchanged for both single-copy and repeat genes . Because PCR-blockage is sometimes used as a relatively insensitive lesion assay , this might seem paradoxical . But the goal of the PCR blockage assay is to determine whether the extent of amplification is reduced compared to undamaged DNA , by measuring the percentage of fragments that have no lesions between the PCR primers . In contrast , IPoD has already identified the CPD-containing fragments via the IP step , so the CPDs can lie outside the PCRd region . The IPoD amplification serves only to make visible a particular set of CPD-containing fragments present in the IP sample . Even when some photoproducts are present , as in the absence of photolyase , the signal is nearly normal: a ) 60% of the ∼750 bp sheared fragment lies outside the ∼300 bp PCR fragment; thus , even if photoproducts are a complete block to PCR , the PCR primers are assaying a CPD-target region external to the PCR primers rather than internal plus external . b ) Diminution of a gene's PCR signal due to a photoproduct internal to the primers is equal between genes , on average , because every IPd molecule has by definition at least one cyclobutane dimer and , at the UV doses used , typically no more than one dimer per molecule . c ) PCR inhibition is only partial because i ) Taq polymerase can slowly bypass CPDs [24] and ii ) partially-extended fragments will , in the next PCR cycle , anneal to a different partner and extend further; thus the internal region is sampled as well . To compare the level of UV-induced CPDs in telomeres with the level in other genomic regions , we calculated for each region the absolute percentage of the input that was IPd ( IP/Input ) . This absolute number circumvents differences in PCR efficiency or copy number . The IP fraction was amplified using primers for p53 , 28S rDNA , and the telomere after removing CPD with photolyase . For the corresponding Input DNA , various dilutions were amplified and a calibration curve of PCR signal vs dilution was constructed . The PCR signal from the IP was compared with the curve to determine the dilution factor matching the IP signal , and thus the ratio IP/Input . At 20 J/m2 , 14% of the telomeric DNA fragments were damaged ( Figure 2 ) , whereas approximately 2% of fragments from the p53 or 28S rDNA genes were damaged at the same dose . The same ratios were obtained whether or not remaining CPD were reversed with photolyase prior to PCR amplification ( not shown ) . Therefore , the telomeric region is 7 times more sensitive than two other regions of the genome . To determine whether one of the telomeric DNA strands was responsible for this sensitivity , we examined the strands separately . Because each strand of the telomere contains only 3 of the 4 possible nucleotides ( only GAT for the 5′TTAGGG strand and only ATC for the 5′CCCTAA strand ) , we performed a strand-specific amplification of the telomere by omitting one nucleotide from the reaction . In addition , an initial linear amplification using only one of the 2 primers and 3 of the 4 nucleotides was performed for 30 cycles . Linear amplification was followed by a standard PCR amplification of the linear-amplified DNA ( see Materials and Methods ) . Each strand was more sensitive than p53 or rDNA ( Figure 2 ) , with 16% of the 5′CCCTAA strand fragments being damaged at 20 J/m2 UVC and 6% of the 5′TTAGGG strand . The telomeric sensitivity was not due to a difference in the frequency of dipyrimidine sites ( the site of formation of cyclobutane dimers ) . This frequency was 29 . 5 dipyrimidine sites per 100 nucleotide in the p53 fragment , 28 . 9 in the rDNA fragment , and 33 . 3 in the telomere . We also examined possible artifactual explanations for the telomeric sensitivity . First , repeated DNA at the ends of chromosomes might sonicate differently , producing more-readily IPd fragments . A Southern blot showed that the sizes of sheared telomere and p53 DNA are the same ( Figure S3A ) . Second , telomeric DNA might have a conformation more accessible to antibody or enzymes . A similar Southern blot experiment revealed that photolyase could completely reverse cyclobutane dimers in both telomeres and p53 , suggesting that , at least in naked DNA , accessibility differences do not play a role ( Figure S3B ) . Thirdly , the large number of telomeric repeats might create shorter PCR fragments , which would PCR more efficiently . But the number-average molecular weight of the telomere smear is 250–500 bp , the same range as the ∼300 bp p53 and 28S bands . Finally , we considered that more ‘copies’ of the ( diluted ) telomeric repeat are present in the PCR reaction than are p53 copies , but this is also true for its pre-IP control . To confirm the UV hypersensitivity of telomeres independently of IPoD , and to test whether the telomere's hypersensitivity was due to its DNA sequence independent of telomere-bound proteins such as shelterins or chromatin-induced DNA conformation , we examined synthetic oligonucleotides . Four different double-stranded 102-mer oligonucleotides were constructed in which the central 60 bp were varied to include either: 10 repeats of the telomere sequence ( 5′TTAGGG/CCCTAA ) ( “Telomere” ) , 10 repeats of a modified 6-mer ( 5′TTCAGG/CCTGAA ) having the same number of potential UV photoproduct sites ( dipyrimidine sites ) ( “Repeats” ) , or a single random sequence containing the same number of potential UV photoproduct sites ( two examples , “Equi-diPyr #1” and “Equi-diPyr #2” ) . Each 102-mer was irradiated with either 100 or 500 J/m2 UVC ( 0 . 1 – 0 . 5 CPD per molecule ) . The irradiated double strand oligonucleotides were directly applied onto a nylon membrane ( without PCR amplification ) using a dot-blot apparatus and CPD-containing DNA was detected using a CPD-specific antibody ( Figure 3 ) . The quantification shows that the telomere repeat was 5 times more sensitive to UVC-induction of CPD than either of the non-repeated sequences . Surprisingly , the non-telomeric 6-mer repeat ( 5′TTCAGG ) was 3 times more sensitive than the random ( non-repeated ) sequences . This result suggests that repeatedness per se renders dipyrimidine-containing oligonucleotides more sensitive to UV , with telomeric sequences being particularly sensitive . To determine the number of repeats needed to confer sensitivity to CPD formation , we designed 102-mer double-strand oligonucleotides having increasing numbers of telomeric repeats ( “Telo” series ) . As control , oligonucleotides were designed to have a dipyrimidine-containing region of the same length as the corresponding telomeric repeats but not arranged as repeats ( “Equi” series ) . Outside the repeated region or the corresponding dipyrimidine-containing region , the oligonucleotide lacks dipyrimidine sites . For the “Equi” series , increasing the length of the dipyrimidine-containing region linearly increased the number of CPDs induced , as expected ( Figure 4 ) . The Telo series behaved similarly to the Equi oligonucleotides up to 4 repeats . Strikingly , a positive effect of repeats on UV induction of CPD appeared around 5 repeats , as if the DNA had undergone a phase transition . The oligonucleotide containing 5 telomere repeats was 3 times more sensitive than the non-repeated oligo . At 7 repeats , a plateau was reached at which sensitivity to CPD formation was 4–5 times greater in the oligonucleotide containing repeats than in the non-repeated oligo . Limitations on synthesizing longer telomeric oligonucleotides prevented us from determining whether the UV-susceptibility of repeats continues to increase with repeat number – with the plateau merely reflecting the fact that double-strandedness is partially lost at DNA ends – or truly plateaus due to complete acquisition of an altered conformation . To examine photoproduct repair , sub-confluent human diploid lung fibroblasts ( WI38 ) were irradiated with a minimally lethal dose of UVC ( 10 J/m2 ) and harvested at different time points 0–48 hours post-irradiation . Photoproduct-containing DNA was then isolated using IPoD and , after photoreversing CPD , amplified using primers specific for the telomere region ( “Telomere” ) , mitochondrial DNA ( “mtDNA” ) , the gene for the RNA component of ribosomal subunit 28S ( “28S” ) , and tumor suppressor gene p53 ( “p53” ) . p53 was used as a positive control for fast repair by the transcription-coupled NER system ( TCNER ) [25] because p53 is actively transcribed in human cells , especially after a genotoxic stress such as UV irradiation . The repair rate observed here will reflect both DNA strands and thus will be an average of TCNER on the transcribed strand and slower global genomic NER ( GGNER ) on the non-transcribed strand . CPDs in the 28S gene of mammalian cells are known to be repaired only by GGNER and not by TCNER [26]–[28] , so it serves as a positive control for normal GGNER . In contrast , NER proteins are not present in mitochondria and CPD are not repaired in mtDNA [29]–[31]; thus mtDNA serves a negative control for repair and would indicate any apparent photoproduct loss due to cell dilution during replication . We found that , 48 hours post-UVC , approximately 70% , 40% and 10% of CPD were removed from p53 , 28S and mtDNA DNA regions , respectively ( Figure 5A ) . Repair of CPD in the telomere region was comparable to or less than that seen in the mtDNA negative control , less than 10% after 48 hr , indicating that the NER system is ineffective in telomeres . To ensure that the lack of repair in the telomere region was not specific to the cell line used , the growth stage , or the UV dose , the experiment was repeated in confluent ( quiescent ) skin fibroblasts at 20 J/m2 UVC , with the same result ( Figure 5B ) . Cyclobutane pyrimidine dimers are profound blocks to DNA replication forks in mammalian cells , triggering cell cycle arrest and DNA damage responses through the ATR pathway [32] , [33] . Oxidative damage at telomeres interferes with maintenance of the D loop and induces telomere shortening [3] , [6] , [34] . To determine how the elevated and persistent levels of CPD affect telomere maintenance , we investigated UV-induced telomere shortening . Human diploid fibroblasts were chronically irradiated with minimally-lethal doses of UVB , receiving 0 to 200 J/m2 UVB 1 day after each passage ( approximately every 5 days ) . After 16 passages , DNA was isolated and approximate telomere length was measured using the telomere restriction fragment ( TRF ) technique [35] . At passage 12 ( “X12” ) , the mean telomere length of un-irradiated cells was approximately 12 kb . At passage 28 ( “X28” ) , the telomere length was approximately 8 kb ( Figure 6 ) , corresponding to the expected telomere shortening with increasing passage level . Irradiating cells with 10 to 200 J/m2 of UVB 16 times did not increase the rate of telomere shortening . Therefore , a ) normal telomere shortening is not accelerated by unrepaired CPD and b ) unrepaired CPDs are not removed by telomere shortening . Evidently , the telomere possesses an efficient tolerance mechanism for cyclobutane pyrimidine dimers .
Telomeres were found to be 7 times more sensitive to UV-induced CPD than other DNA regions ( Figure 2 ) . This observation was made in a cellular context , so the proteins and secondary structure of the chromatin might be involved in this hypersensitivity . To distinguish these possibilities , we tested the sensitivity of telomere sequence inserted in a 102-mer oligonucleotide . Because this oligonucleotide was irradiated in vitro , it was free of any cellular context . This oligonucleotide showed hypersensitivity comparable to the telomere DNA sequence in vivo ( Figure 3 ) . Thus the cellular context is not the major contributor to the UV hypersensitivity of the telomere . What , then , can explain it ? We tested different oligonucleotides for UV sensitivity and found that short repeats , like the telomeric sequence , render those oligonucleotides more sensitive . An oligonucleotide containing 10×6-mer repeats was ∼5-times more sensitive than an oligonucleotide containing the same frequency of dipyrimidine sites but randomly distributed ( not in repeats ) . Surprisingly , the sensitivity of the telomeric repeat underwent a sudden transition at 5 repeats , suggestive of a structural phase change ( Figure 4 ) . This result means that the expected sensitivity based on DNA sequence is not the entire source of UV sensitivity . The biophysical nature of this transition , and its effect on the distribution of DNA photoproducts , will require biophysical investigation . G-rich single strands undergo a variety of interactions such as Hoogsteen base pairing and G-G stacking . These can create G quadruplexes , parallel-stranded helixes , A- and Z-form DNA , hairpins , and local melting . In telomeric and trinucleotide repeats , the stability of the various structures depends on the number of repeats [36] , [37] . The behavior in double-stranded DNA is less studied . A region of the genome so critical to cell survival and genomic integrity would be expected to preserves its own integrity after a genotoxic stress . Yet little is known about how telomeric DNA does this . The finding that telomeres are hypersensitive to UV-induced DNA damage prompted the expectation that repair of this DNA damage would be rapid , to prevent DNA damage accumulation in this region . What we found was the contrary . Repair was almost absent in telomere regions , proceeding as slowly as in mitochondrial DNA where NER proteins are absent ( Figure 5 ) . Two days after UV irradiation , CPD were still present in telomeres but had been half removed from coding regions ( p53 or 28S genes ) and probably entirely removed from the transcribed p53 strand . The repair defect could be active or passive . In the passive category , compaction of telomeric heterochromatin may prevent access of repair proteins [38] . Also , telomeric DNA has been reported to have partial A-DNA character [39] , which predisposes to trans- rather than cis-isomers of CPD [8] . Little is known about the repair of trans-isomers of CPD and they may be more difficult for the NER system to recognize or remove . In the active category , some of the many protein factors bound to telomeres ( reviewed in [40] ) may inhibit the repair system in this region . There are two reasons suppression of excision repair can be desirable . The high frequency of CPDs in the telomere , together with the telomere's repeat nature , may generate multiply damaged sites ( MDS ) . MDS are sites where DNA lesions are closer than ∼20 bp on opposite strands [41] . After the incision nicking that is the first step in excision repair , multiply damaged sites result in double-strand DNA breaks . This has been observed for UV-irradiated DNA containing halogenated nucleotide analogs in close proximity [41] . Double-strand breaks , in turn , are clastogenic and lethal events . At an MDS , displacement of the lesion-containing oligonucleotides during the second step of excision repair will also create overlapping daughter strand gaps [13]–[15] . This event increases the permissible distance between CPDs . In unique-sequence DNA , such MDSs would be rare , but in repeats they could be the rule when photoproduct frequency is high . The absence of telomere shortening after chronic UV irradiation ( Figure 6 ) indicates that , in fact , such double-strand breaks have been avoided . The fact that cell proliferation was unhindered by chronic UV irradiation , despite the presence of CPD in their telomeres , raises a new question: how can a cell tolerate DNA damage in its telomeres ? During mammalian DNA replication , a bulky lesion such as a CPD typically blocks replication fork progression [42] , [43] . This blockage leads to single-stranded DNA that activates ATR-dependent stress responses such as G2/M arrest and apoptosis [44] . To avoid these events at unrepaired CPDs , the replication mechanism uses DNA polymerases capable of bypassing CPD . In E . coli , the SOS response activates polV to a translesion synthesis polymerase by transferring RecA-ATP to it from a RecA filament [45] . In human cells , the XPV gene ( defective in the xeroderma pigmentosum variant complementation group ) codes for pol eta , a polymerase able to bypass CPD by incorporating A opposite a T or C in a CPD ( reviewed in [46] ) . Correspondingly , cells from a squamous cell carcinoma from an XPV patient were found to generate recurrent chromosome abnormalities as they were passaged in vitro . These were dicentric chromosomes , particularly telomere–telomere bridges , indicative of telomeric damage [16] . It seems likely , then , that CPD accumulating in the telomere are especially reliant on bypass to avoid replication gaps . In the absence of bypass , these replication gaps would be frequent enough to trigger telomeric double-strand breaks and telomere–telomere bridges , the same kinds of genetic catastrophes that repair suppression aims to avoid .
Each experiment was performed with two different primary human fibroblast cell strains . The first strain was derived from breast reduction tissue from a healthy 25-year old female [47] . The other strain was the commercially available WI38 , derived from lung tissue of a male foetus ( ATCC , Manassas , VA ) . Cells were grown in high-glucose DMEM ( Gibco Invitrogen ) supplemented with 10% FBS and 1% penicillin/streptomycin . Cells were UV-irradiated at room temperature after replacing the medium with cold sterile phosphate buffered saline ( PBS ) . The two cell strains have different UVC sensitivities ( Figure S4 ) . The UVC source was a germicidal lamp emitting at 254 nm . Using UVC rather than UVB avoids potential complications from photosensitized oxygen radical formation . For the telomere shortening experiment , UVB was used to maximize the likehood of telomere shortening; the source consisted of three fluorescent tubes ( FS20T12/UVB/BP , Philips ) filtered through a sheet of cellulose acetate to eliminate wavelengths below 290 nm ( Kodacel TA-407 clear , 0 . 015 inch thickness; Eastman-Kodak Co . ) . Dose rate was measured prior to each experiment using a UVX UV-meter ( UV Products , Upland , CA ) . Purification of the DNA was performed using DNeasy Tissue Kit ( Qiagen , Valencia , CA ) , according to the manufacturer's protocol . Purified DNA was sonicated to 500–1000 bp fragments ( Branson sonifier 250 , microtip , at 30% power , 3×15 sec on ice ) , precipitated with NaCl/ethanol , and resuspended in resuspension buffer ( 0 . 01% SDS , 1 . 1% Triton X 100 , 1 . 2 mM EDTA , 16 . 7 mM Tris-Cl pH 8 . 1 , 167 mM NaCl ) . DNA was denatured by boiling 10 min , incubated with the CPD-specific antibody ( D194-1 , MBL , Woburn , MA ) [48] overnight at 4°C and then with a rabbit anti-mouse secondary antibody for 1 hour . The anti-CPD antibody was used in molar excess to CPD to ensure that each damaged dipyrimidine was pulled down regardless of its local sequence or slight variations in the binding affinity of the antibody to each dipyrimidine type . Molar excess is indicated by the linearity of the dose-response with respect to substrate ( Figure 1B ) . Antibody-bound DNA was pulled down using Staph A beads ( Calbiochem ) . The bead/DNA complexes were washed 2 times with wash buffer 1 ( 2 mM EDTA , 50 mM Tris-Cl pH 8 . 0 ) and 4 times with wash buffer 2 ( 100 mM Tris-Cl pH 8 . 0 , 500 mM LiCl , 1% NP40 , 1% deoxycholic acid ) . DNA was eluted from the staph A beads with elution buffer ( 50 mM NaHCO3 , 1% SDS ) and the eluted DNA was cleaned using a Qiagen PCR purification kit to remove salts and SDS prior to PCR . In the indicated experiments , CPD were removed before the PCR reaction ( but after the IP step ) using cloned E . coli CPD photolyase ( kindly provided by Drs . C . Selby and A . Sancar ) . The CPD photoreactivation mix ( 10 mM Tris-HCl pH 7 . 6 , 10 mM NaCl , 2 mM EDTA , 20 mM DTT , 0 . 2 mg/mL BSA , 0 . 1 µL CPD photolyase ) was added to the DNA and exposed for 1 h to UVA light from eight F20T12BL lamps ( Spectra Mini , Daavlin Co . , Bryan , OH ) passed through filters to remove UVB and UVC . The DNA was then cleaned using a PCR purification kit ( Qiagen ) . For PCR reactions , 20 cycles of amplification were performed on a Biometra TGradient thermal cycler with Taq polymerase in 10 mM Tris/HCl , 1 . 5 mM MgCl2 , 50 mM KCl , pH 8 . 3 and 200 µM each dNTP ( Roche Molecular Biochemicals , Indianapolis , IN ) . A test run of PCR using different amounts of starting material was done on each sample and on each primer set to ensure the amplification lay in the exponential portion of the amplification reaction . The following primers were used: For the telomere sequence: 5′GGTTTTTGAGGGTGAGGGTGAGGGTGAGGGTGAGGGT and 5′TCCCGACTATCCCTATCCCTATCCCTATCCCTATCCCTA [19] . The underlined bases are mismatched with respect to the telomere sequence . For the p53 gene: 5′CTGCCTCTTGCTTCTCTTTTCC and 5′GGTTTCTTCTTTGGCTGGG , giving a PCR product of 309 bp . For 28S ribosomal DNA: 5′GTAGAATAAGTGGGAGGCCCCCGG and 5′AGGCCCCGCTTTCACGGTCTGTATTCG , giving a PCR product of 368 bp . For the CYTB gene in mitochondrial DNA: 5′CCCTAGCCAACCCCTTAAAC and 5′TTGGCTTAGTGGGCGAAATA , giving a PCR product of 297 bp . The agarose gel was scanned and quantification was done using ImageQuant 5 . 0 software ( Molecular Dynamics ) . For p53 , 28S and mtDNA , the band was simply quantified and the background was subtracted from the signal . For the telomere sequence , the PCR primers can anneal varying distances apart on the telomeric repeat , so the PCR product is not a single-size product but rather an assortment of DNA fragments over a size range . We therefore ran telomeric samples on the agarose gel for a few minutes ( to let the DNA enter the gel and to separate the PCR product from the primers ) , making the smear band-like . The entire smear was quantified using the same technique as for coding regions . Oligonucleotides used in dot-blot experiments are depicted in Table 1 . 400 ng of each double-strand oligo was irradiated with the indicated UVC doses using a 254 nm source . The irradiated DNA was denatured and applied onto a nitrocellulose membrane using a dot-blot apparatus . CPD-containing DNA on the membrane was visualized using a CPD-specific antibody ( D194-1 , MBL , Woburn , MA ) [48] followed by a secondary anti-mouse-HRP antibody ( Santa Cruz Biotechnology , Santa Cruz CA ) and revealed by chemiluminescence ( Denville , Metuchen , NJ ) . Different film exposures were scanned and quantification was done using ImageQuant 5 . 0 software ( Molecular Dynamics ) . Cells were irradiated with different UVB doses ( 0 , 10 , 50 , 100 and 200 J/m2 ) . After the irradiation , cells from each condition were allow to grow until they reached full confluency . When cells from every exposure condition reached 100% confluency , they were all passaged 1∶4 . This precaution was taken to assure that UV-irradiated cells did not undergo fewer population doublings than unirradiated ones at the same passage number . ( A disadvantage of this design is that mortality at the higher UV doses would cause more divisions of the remaining living cells to compensate , possibly leading to faster telomere shortening at these doses . However , because telomere shortening was not seen , this absence is conclusive . ) UV-induced telomere shortening would be obscured if UV also reduced the number of cell doublings by decreasing the cell density at confluence . This effect would reduce the extent of normal , replication-related , telomere shortening . The cell density reduction apparently did not occur here . Because each cell lineage was split at the same ratio , a 25% reduction in cell density of treated cells compared to untreated would result in a ( 0 . 75 ) 16 = 100-fold difference in cell number after 16 passages ( from X12 to X28 ) . But no difference in the final amount of DNA harvested was observed between any of the UV doses . Terminal restriction fragment length measurements were obtained using the Telo TTAGGG telomere length assay kit ( Roche Molecular Biochemicals , Indianapolis , IN ) as done previously [47] . Briefly , 2 mg of HinfI/RsaI-digested genomic DNA were separated on 0 . 8% agarose gels and Southern blotted onto a Hybond-N+ nylon membrane ( Amersham Biosciences , Piscataway , NJ ) . After UV-fixation of DNA fragments onto the membrane , membranes were hybridized with digoxigenin-labeled telomere-specific probe ( TTAGGG ) 4 . After washing out non-bound probe , membranes were incubated with a digoxigenin-specific antibody covalently coupled to alkaline phosphatase . Finally , the telomere fragments were visualized by a chemiluminescent substrate ( CDP-star , Roche Molecular Biochemicals , Indianapolis , IN ) . TRF lengths were determined by comparing the signals relative to a standard molecular weight using ImageQuant 5 . 0 software ( Molecular Dynamics ) . All lanes were divided into 75 intervals , and the mean TRF length was defined as S ( ODi ) /S ( ODi/Li ) , in which ODi is the chemiluminescent signal and Li is the length of the TRF fragment at position I [49] . Although TRF fragments have one terminus in the pre-telomeric region , changes in TRF length reflect changes in telomere length .
|
Telomeres consist of a repeated sequence located at each end of each chromosome . This repeated sequence is required for chromosomal stability and integrity , a function important for both cancer and aging . The DNA sequence of human telomeres is 5–10 kb of a repeated double-strand hexamer ( 5′TTAGGG/5′CCCTAA ) . In theory , this sequence is nearly optimal for acquiring UV-induced DNA damage . We developed a novel technique , the immunoprecipitation of DNA damage ( IPoD ) , to study DNA damage induction and repair in the telomere and in coding regions ( p53 , 28S rDNA , and mitochondrial DNA ) . We find that human telomeres are hypersensitive to UV-induced DNA photoproducts and that the removal of those DNA photoproducts is almost absent . Cells containing persistent high levels of telomeric DNA damage nevertheless proliferate and chronic UV irradiation of cells does not accelerate telomere shortening . Telomeres are therefore unique in at least three respects: their biophysical UV sensitivity , their prevention of excision repair , and their tolerance of unrepaired lesions .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"molecular",
"biology/dna",
"replication",
"cell",
"biology/cellular",
"death",
"and",
"stress",
"responses",
"dermatology/skin",
"cancers,",
"including",
"melanoma",
"and",
"lymphoma",
"molecular",
"biology/chromosome",
"structure",
"genetics",
"and",
"genomics/chromosome",
"biology",
"genetics",
"and",
"genomics/cancer",
"genetics",
"molecular",
"biology/chromatin",
"structure",
"molecular",
"biology/dna",
"repair"
] |
2010
|
Human Telomeres Are Hypersensitive to UV-Induced DNA Damage and Refractory to Repair
|
Studies of the evolution of development characterize the way in which gene regulatory dynamics during ontogeny constructs and channels phenotypic variation . These studies have identified a number of evolutionary regularities: ( 1 ) phenotypes occupy only a small subspace of possible phenotypes , ( 2 ) the influence of mutation is not uniform and is often canalized , and ( 3 ) a great deal of morphological variation evolved early in the history of multicellular life . An important implication of these studies is that diversity is largely the outcome of the evolution of gene regulation rather than the emergence of new , structural genes . Using a simple model that considers a generic property of developmental maps—the interaction between multiple genetic elements and the nonlinearity of gene interaction in shaping phenotypic traits—we are able to recover many of these empirical regularities . We show that visible phenotypes represent only a small fraction of possibilities . Epistasis ensures that phenotypes are highly clustered in morphospace and that the most frequent phenotypes are the most similar . We perform phylogenetic analyses on an evolving , developmental model and find that species become more alike through time , whereas higher-level grades have a tendency to diverge . Ancestral phenotypes , produced by early developmental programs with a low level of gene interaction , are found to span a significantly greater volume of the total phenotypic space than derived taxa . We suggest that early and late evolution have a different character that we classify into micro- and macroevolutionary configurations . These findings complement the view of development as a key component in the production of endless forms and highlight the crucial role of development in constraining biotic diversity and evolutionary trajectories .
The tremendous diversity of shapes and forms observed in nature is truly remarkable and yet it represents only a small fraction of the ‘space’ of the possible . One reason for this is that the space of possible genotypes has been incompletely sampled over the course of the history of life on earth . If we consider the astronomical volume of the genotypic space , then the set of all DNA strands that were ever produced during earth history constitute a tiny fraction of the total sequence space . Moreover , the genotypes that have existed are the result of an evolutionary process—descent with modification from a common ancestor—which is a locally-delimited generative process . Phenotypic diversity is further constrained by another process , one intrinsic to the manufacture of adaptive varieties , the developmental mechanisms that determine the mapping of genotypes into phenotypes . Development induces a non-linear and highly degenerate mapping from gene-space to phenotype space , whereby many genotypes produce similar ( or identical ) phenotypes , and concomitantly , ensuring that there are many phenotypes that cannot be generated by any genotype . This arises from both neutral genetic properties of the developmental dynamic , and from the evolution of robustness mechanisms which seek to preserve functional phenotypes in the face of environmental and genetic variation [1] , [2] . Degeneracy has the effect of hiding genotypes from the selective process and rendering a large portion of potential phenotypes inaccessible . This is an architectural constraint that limits available variation and adaptive capacity , with potentially dramatic effects on the trajectory of the evolutionary process . Whereas evolutionary search over the space of frequently generated phenotypes is in strict accordance with neo-darwinian theory ( population genetics for example ) , the sparse distribution of the phenotypic space has implications for large scale patterns of evolutionary change , and this can only be appreciated through the introduction into the evolutionary dynamic of a suitable model of development . Developmental mappings are generally extremely complex . This complexity derives from a combination of hierarchical regulation , multi-gene control , epistasis , and pleiotropy . A large body of work examines the statistical and dynamical properties of developmental maps in simple systems , focusing on neutrality and neutral networks of RNA [3]–[5] and on gene regulatory networks in multicellular development [6]–[8] . These studies have generated interest among paleontologists inquiring into the origin and diversification of body plans [9]–[11] and have lead to the suggestion that morphological variation is extensive early in the history of multicellular life [9] , [12] , that phenotypes are sparsely distributed in the space of ‘potential’ phenotypes [13] , and that diversity is better predicted by variation in the structure of gene regulation networks than variation in the presence and absence of structural genes [14] . Here , we consider a very generic property of complex developmental maps—the interaction between multiple genetic elements and the non-linearity of gene interaction—in shaping various aspects of a phenotype . On the mechanistic genetic level , this is usually referred to as epistasis and pleiotropy , but the same generic constraint principle might also apply to many other biological mappings , ranging from the physical interactions between amino acids in the production of protein structures , to the interactions between tissues and their effects on gross morphology . We wish to show that a basic geometric property of development provides a null model able to account for the bias and nonuniformity of phenotype distributions . The model is constructed as a generic representation , capturing the way multiple genetic inputs combinatorially interact to influence multiple phenotypic traits , and does not assume selection . One natural interpretation is that of cis-regulatory architecture and gene interaction [14]–[16] . For convenience , we use terms related to this interpretation throughout the paper . We use the model to examine a number of statistical regularities of the developmental map that it induces . In particular , we derive the fraction of visible phenotypes generated during development and the dependence of this fraction on the level of interaction between genetic elements . We characterize the distances among visible and frequently occurring phenotypes and the influence of development on phylogenetic relationships . We demonstrate that many of the empirical , developmental and paleontological regularities summarized above can be recovered using this null model .
Genotypes and phenotypes are represented as binary vectors of lengths r and k . Generally speaking , genotypes represent the presence/absence of r genetic elements ( e . g . , genes , alleles , etc . ) , and phenotypes represent the presence/absence of k phenotypic traits . An interpretation in terms of cis-regulatory dynamics posits that genotypes represent the expression pattern of a set of r transcription factors ( TFs ) and that phenotypes denote the expression pattern of k target genes regulated by these TFs . In this sense , genotypes and phenotypes in our model may be viewed as representing certain aspects of the cell transcriptional state . In the following , we refer to r as the regulatory dimension and to k as the phenotypic dimension . A developmental plan maps genotypes to phenotypes . We define a developmental plan as a matrix , D , of size k×r . Each entry in this matrix is either +1 or −1 with equal probability ( using real numbers drawn from a uniform or Gaussian distribution with mean 0 does not qualitatively change the results presented in this paper ) . Given a genotype , , the phenotype to which it maps is calculated by , where H denotes the heaviside function ( i . e . , the unit step function centered at zero ) . In the regulatory interpretation , Dij describes properties of the binding site for transcription factor j in the promoter of gene i ( Figure 1 , and see [15] , [17] ) . The heaviside function can alternatively represent a switching mechanism , producing a signal only if inputs exceeds a threshold value . In the analysis presented throughout the paper , we enumerated all 2 r possible genotypes and used a fixed , randomly generated developmental plan D to map these genotype onto the corresponding 2 r phenotypes . To obtain large-scale statistics for the distribution of visible phenotypes and their relationships , we repeated this process , using numerous developmental plans . The distinction between a ‘structural’ part of the genome ( which is allowed to vary ) and a developmental part ( that remains fixed ) , is motivated by our attempt to explore the implications of a given plan on the distribution of phenotypes , and by the suggestion that developmental plans form a mechanical basis for phylogenetic grades [7] ( see also Discussion ) . Previous studies on the evolution of development , have considered a dynamical recurrent model of gene regulation . In these models , the resulting ‘phenotype’ ( or pattern of gene expression ) is fed back into the regulatory plan , until the system reaches steady state [15] , [17] . These models aim to capture multilayered plans , where the output of one layer forms the input to the next ( these recurrent models are a simple case where the same plan is applied to each layer ) . By contrast , the model described above employs a single layer architecture . To examine the effect of such multilayered plans we extend our model by allowing it to include multiple regulatory layers . Formally , we define , where t denotes the regulatory layer number and the initial phenotype , , corresponds to a given genotype ( as described in our original model ) . We iterate this model repeatedly , starting with a collection of all possible 2 r genotypes , and record the phenotype distribution at each layer . Our basic model assumes a ‘fully connected’ regulatory plan , wherein all entries in the developmental matrix D are nonzero: every gene in the genotype affects every element in the phenotypic vector . Previous models have considered sparser interaction plans and examined the effect of varying the density of the regulatory interactions . Specifically , the density of the regulatory plan has been shown to have important effects on the consequences of gene duplications [18] , epigenetic stability [15] , the evolution of canalization [17] , and robustness [19] . We therefore further extend our model by introducing an additional parameter , c , which denotes the density of the matrix D ( i . e . , the probability for each entry in the matrix to be attributed with a nonzero value; c = 1 corresponds to a fully connected plan ) , and examine its effect on the distribution of phenotypes .
Consider a developmental model with regulatory dimension r and phenotypic dimension k . There are 2 r genotypes which could produce a maximum of 2 r phenotypes . However , the developmental plan maps several genotypes into the same phenotype ( giving rise to degeneracy ) , and consequently generates a much smaller number of distinct phenotypes . We refer to the set of phenotypes produced by a given developmental plan as visible phenotypes , and examine the number of visible phenotypes and the number of potential phenotypes as a function of r ( Figure 2A ) . We find that while the number of visible phenotypes increases with the regulatory dimension , r , their fraction , out of the number of potential phenotypes , rapidly declines , with around 5% of the potential phenotypes remaining visible ( Figure 2B ) when r = k . In other words , the expansion of the genotypic space , which also promotes an expansion in the number of possible genotypic configurations , also brings about an increased canalization , masking the expansion in the number of new visible phenotypes . This is further exemplified by the marginal contribution of each genetic element ( e . g . , each transcription factor ) , measured as the relative increase in the number of visible phenotypes obtained by adding a new genetic element . This declines from about two-fold for the first few elements , to less than 1 . 4 as r reaches k ( Figure 2C ) . As per the mathematical analysis below , this can be attributed to the effect of an unbalanced sample of D entries and the multiplicative effect of the nonuniform distribution of each element in the phentype . Interestingly , the function describing the fraction of visible phenotypes ( Figure 2B ) is sigmoidal , with the greatest change in the fraction of visible phenotypes occurring at regulatory dimensions on the order of half the phenotypic dimension . Thus for smaller regulatory circuits , a large fraction of potential phenotypes remain visible , whereas for larger regulatory circuits , the greater fraction of the phenotypic space is hidden and inaccessible to selection and evolutionary transformation . Having established that large regulatory networks lead to a small number of visible phenotypes , we turn to the statistical characteristics of the visible , phenotypic subspace . We focus on developmental plans for which r = k ( e . g . , the number of TFs matches the number of target genes ) . As demonstrated above , these developmental plans produce the most restricted set of visible phenotypes . Unless otherwise indicated , we set r = k = 14 to allow for the complete enumeration of all genotypes . We consider the distribution of frequency levels among the visible phenotypes . We calculate for each phenotype j , a degeneracy level , nj , denoting the number of different genotypes that produce it ( visible phenotypes correspond to those phenotypes for which nj>0 ) . The distribution of degeneracy levels fits a generalized power-law distribution ( Figure 3A ) , implying that there are a few very common ( frequent ) phenotypes and many rare ones . These findings replicate those on the highly nonuniform frequencies of folding geometries in the RNA secondary structure genotype/phenotype map [20] . Are the visible phenotypes uniformly distributed across the phenotypic space or clumped in a nonuniform , subspace of closely related phenotypes ? To answer this question , we compute the gain function of a given developmental plan . This function describes the distribution of Hamming distances among phenotypes ( using the average pairwise dissimilarity [21] ) whose origins are genotypes a certain Hamming distance apart . In other words , the gain function measures how the magnitude of a perturbation in the genotype space maps onto perturbations in the phenotype space . The resulting gain function suggests that the developmental plan induces a significant degree of canalization; a large perturbation in the genotype space ( measured as the Hamming distance between the original and perturbed genotypes ) produces , on average , a significantly smaller perturbation in the phenotype space ( Figure 3B ) . This canalization compresses the image of the genotypes in the phenotype space , and promotes a patchy subspace . This property of regulatory networks has been adduced as evidence for the incremental evolution of developmental robustness [19] . To further characterize the patchiness of the visible , phenotypic subspace , we calculate the distribution of pairwise Hamming distances between visible phenotypes , this time not conditioning on the distance between their genotypes ( Figure 3C ) . Here we plot the frequency distributions of pair-wise Hamming distances under three different conditions: ( 1 ) randomly drawn phenotypes from the space of potential phenotypes , ( 2 ) randomly drawn phenotypes from the set of distinct , visible , phenotypes , and ( 3 ) randomly drawn phenotypes from the set of visible phenotypes including all occurrences of each phenotype . In other words , each distinct , visible phenotype is sampled with a probability proportional to its frequency . We find that the condition involving the visible phenotypes , tends to generate phenotypes more similar than expected by chance ( random phenotype distribution ) . Moreover , when controlling for frequency ( the degeneracy level ) of the visible phenotypes , we find that the distribution is further skewed toward smaller Hamming distances . This suggests that the most frequent phenotypes span a smaller subspace than the total visible phenotypes , and are located towards the center of the visible phenotype set . To examine this observation in greater detail we measure the average Hamming distance between visible phenotypes as a function of their frequency and represent them on a frequency-rank versus distance plot . The highest ranked phenotypes are presented as the lowest rank values . As shown in Figure 4A , the distance between the most frequent phenotypes is significantly smaller than the average distance ( which in this case is ∼6 ) , and increases as more visible phenotypes ( with lower frequencies ) are considered . Considering the case where all the visible phenotypes are included in this analysis , the average distance is still smaller than that expected by chance . We find that the top 5% most frequent phenotypes are very similar ( average Hamming distance is smaller than 4 ) yet cover approximately 50% of all the visible phenotypes ( 4A inset ) . An additional illustration of this patchiness can be observed in Figure 4B , plotting the one mutant-neighbor network of all the visible phenotypes . Here we observe that the nodes that represent the most frequent phenotypes tend to be separated in most cases by a single edge . While an exact mathematical derivation for the nonuniform distribution of degeneracy levels and fraction of hidden phenotypes is hard to obtain , we consider an approximate , statistical approach in order to provide an intuition for their origin . We first examine the expected statistical properties of a single trait element . Let pj denote the jth element of the phenotype . We consider complex , non linear mappings of the form: , where H denotes the heaviside function , denotes the jth row of the developmental matrix D , and denotes a given genotype . The binary vector , selects elements of for summation . It follows that Pr ( pj = 1 ) is the probability that the sum of the elements in a subset of elements is greater than zero . Each element in is either +1 or −1 with equal probability . Let zj denote the number of +1 elements in . zj follows a binomial distribution B ( r , 0 . 5 ) , where r is the regulatory dimension—the number of elements in . Let sg denote the number of nonzero elements in the genotype . Pr ( pj = 1 ) is the probability that a subset of size sg drawn without replacement from a set of zj number of +1 elements and r−zj number of −1 elements , contains more +1 elements than −1 elements . This probability is given by , ( 1 ) where f denotes the hypergeometric probability mass function , . Furthermore , since in our model we consider all genotypes ( all possible subsets of r choose sg ) to be occupied by 1 or a 0 with equal probability , we can multiply our previous expression by the binomial probabilities for each element of the genome , to derive an average probability for each trait value : ( 2 ) Figure 5 illustrates that Pr ( pj = 1 ) is a sigmoidal function of zj . If pj had been determined by only one , randomly drawn , element of , Pr ( pj = 1 ) would be proportional ( linearly ) to the fraction of +1 elements in . However , since pj is determined by a random subset , the consequences of a larger fraction of +1 elements is a combinatorial amplification . For example consider the case where is comprised mostly of −1's with only very few +1 elements . A subset of will typically have many more −1's than +1's , as there are exponentially many more ways to choose −1 elements than the +1 elements . We argue that this strong dependence of the phenotypic element on the number of +1 elements in the corresponding developmental matrix row is the source for the nonuniform distribution of degeneracy levels . We next consider the entire phenotypic vector , rather than a single trait element j . Clearly , Pr ( pj = 1 ) and Pr ( pl = 1 ) , the probabilities of producing 1 in the jth and lth elements of the phenotype , are not independent . When mapping a genotype to a phenotype , we use the same columns of D ( as defined by ) to construct the summed subset in each row . Let's just assume that each trait element is independent which can be stated through the following identity: ( 3 ) where k indexes the phenotypic dimension . Note that the expected value of zj is E ( zj ) = r/2 and from Equation 2 we get . If all rows of D possess an equal number of +1's and −1's we find for every phenotype . This generates a uniform distribution of degeneracy levels ( and no hidden phenotypes ) . Because zj is sampled from a binomial distribution the number of +1's in each row can diverge from r/2 , and consequently , as illustrated in Figure 5 , bias the probability distribution of phenotypes . Consider the case where several rows of the developmental matrix have zj>r/2 . The probability of producing 0's in the phenotype elements that correspond to these rows is very small ( note again Equation 2 and the sigmoidal shape in Figure 5 ) . Consequently , producing phenotypes with 0's in all these elements is extremely unlikely ( see Equation 3 ) and these phenotypes are expected to be hidden . This intuition can also help us to understand the similarity of high frequency phenotypes and the patchiness of the visible phenotype space . Assume that zj≈r/2 only in the first and third rows , and zj>r/2 in all others . Since the phenotypes are biased toward 1's in all elements apart from the first and the third , all phenotypes of the form [– , 1 , – , 1 , 1 , … , 1] ( where ‘–’ denotes either 0 or 1 ) are likely to be highly degenerate and will form a dense patch of high frequency phenotypes . We further confirm this intuition numerically ( see Text S1 ) . We applied the mathematical formulation and drew a sample set of zj's from a binomial distribution B ( r , 0 . 5 ) . We calculated the probability of obtaining certain phenotypes and showed that the distribution of degeneracy levels is comparable to that obtained with our model . We also demonstrated that the degeneracy levels of neighboring phenotypes are strongly correlated ( Text S1 ) . The model we have utilized employs a single layer architecture and might be thought to limit the scope of possible regulatory schemes . Computationally this is the case , as at least two-layers ( input layer plus a hidden layer ) are required to produce a perceptron that is a universal Turing machine ( or universal function approximator ) , as proven for the Cybenko theorem [22] , able to achieve linear separability of inputs as is required by , for example , the XOR function . However , since we show that a single regularity layer compresses the image of the genotypes in the phenotype space , introducing additional layers only produces further canalization and strengthens our findings . We quantify this effect by using an extended multilayered developmental model ( see Models ) and record the number of unique visible phenotypes and the phenotype distribution obtained after each iteration . First we consider the simple , recurrent , or recursive model , where Dt = D for every t , and examine the effect of introducing up to 50 regulatory layers . For this recursive scheme , at each additional layer there is a reduction in the number of visible phenotypes and an increase in canalization ( Figure 6A ) . Moreover , the number of visible phenotype reaches an ( extremely low ) asymptotic value , which is not influences by additional regulatory layers , suggesting that a steady state has been reached . In order to examine these findings in detail we allow the developmental process to iterate indefinitely until a steady state is reached . This extended model is also strictly comparable with the recurrent models introduced in [15] , [17] . As we consider the entire set of possible initial genotypes ( rather than a single , predefines , initial phenotype ) , we apply a slightly more stringent condition for ascertaining the steady state , and require that the set of unique , visible phenotypes does not change ( note that this condition also accommodates limit cycle equilibria ) . Considering 10 , 000 different developmental plans , we find that on average the number of visible phenotypes at steady state is only 10 . 3±7 ( 0 . 063% of all possible phenotypes ) , and that this steady state is reached after 17 . 7±7 layers . Figure 6B further illustrates the distribution for the number of unique visible phenotypes at steady state . We now examine the behavior of a more general , developmental model , where each regulatory layer can incorporate a different developmental plan . This is closer to natural regulation where we observe multiple layers of post-transcriptional control . As illustrated in Figure 6A , this model yields a more complete reduction in the number of visible phenotypes ( as a function of the number of layers ) . It is enough to note that the ‘all zeros’ phenotype always maps onto itself , and that each layer can map some fraction of the remaining visible phenotypes to this zero-class . This establishes why this model is asymptotically destined to reach a steady state with a single , ‘all zeros’ , visible phenotype . This raises an intriguing question as to the optimum number of regulatory layers . Two or more layers offer greater computational power , but at a cost of reduced phenotypic variability . Finally , we observe a change in the distribution of degeneracy levels as more regulatory layers are introduced . The reduction in the number of visible phenotypes is accompanied by an increase in the probability of highly degenerate phenotypes , and by an overall increase in the extent of degeneracy in the system ( Figure S1 ) . This abundance of highly canalized phenotypes further strengthens the conclusions obtained for the original model . The fully connected model analyzed above represents , to some extent , a worst-case scenario in terms of gene interactions , pleiotropy , and epistasis . Here , we examine whether the regulatory bias observed in our model holds when the regulatory plan is less dense , and how the density of the plan influences this bias . There are two competing possibilities to be considered . On the one hand , if the developmental matrix is very parse , there may be entries in the phenotype vector that are never activated . This would further reduce the number of visible phenotypes . On the other hand , for sparse matrices , each phenotypic element is influenced by only a few genes ( low epistatsis ) , making the contribution of each gene to the state of the phenotypic elements higher . Changing one gene in the genotype could change a corresponding element in the phenotype ( inducing a steeper gain function - see Figure 3B ) . In the extreme case of the unit matrix , every change in the genotype induces a change in the phenotype . This would decrease the level of neutrality ( degeneracy ) of the genotype-phenotype map and consequently produce more visible phenotypes . We find that sparse matrices generate a smaller fraction of visible phenotypes ( Figure 7A ) . For example , in comparison to the 8 . 2% visible phenotypes obtained for a fully connected plan ( c = 1 ) , only 3 . 4% of the phenotypes are visible for a matrix with c = 0 . 25 and only 0 . 6% are visible for c = 0 . 1 ( see Models ) . It also appears that the maximum number of visible phenotypes ( which is still only 8 . 7% of the total number of potential phenotypes ) is produced for an intermediate value of c≈0 . 85 . This could be the outcome of a trade-off between the two competing effects discussed above . We note , however , that the influence of an increase in matrix density on the fraction of visible phenotypes diminishes for c>0 . 5 . To disentangle the influence of a sparse matrix density on the fraction of visible phenotypes derived from varying levels of gene interactions ( epistatsis ) , from the effect of reduced phenotypic activation , we control for the number of potentially activated phenotypic elements under each plan . We measure the number of variable phenotypic elements ( traits ) , ν , obtained for a range of values of c ( Figure 7B ) . A variable trait is defined as a phenotypic elements that can be activated ( i . e . , assume a value of 1 ) in at least one of the visible phenotypes produced by a given developmental matrix . As expected , sparser plans result in a lower number of variable traits . If only ν traits are variable , the potential number of phenotypes is bounded by 2 ν ( rather than 2 r ) , and it is reasonable to measure the fraction of visible phenotypes in relation to this lower limit . Examining the fraction of visible phenotypes out of the achievble ( 2 ν ) phenotypes , the effect of the reduced interaction level is revealed ( Figure 7C ) . Sparser matrices produce a higher fraction of visible phenotypes ( reaching almost 80% on average for very sparse plans ) , owing to the higher marginal contribution each gene makes in determining the state of a phenotypic element . Thus lower epistasis in the sparse matrix allows for a greater per locus contribution to the phenotype . We further confirm that the statistical properties of the visible phenotype distribution still holds for sparse matrices . We first compare the distribution of degeneracy levels obtained for varying values of c ( Figure S2 ) . Although sparse matrices ( e . g . , c = 0 . 1 ) produce a more variable distributions , a clear power-law distribution can already be observed for matrices with c = 0 . 25 or higher . The patchiness of the visible phenotypic space is also confirmed for sparse matrices by examining the distribution of pairwise hamming distances among randomly selected phenotypes ( Figure 8 ) . Sparse matrices induce an even more pronounced patchy phenotypic space , largely as a result of a reduction in the number of visible phenotypes these plans produce ( see Figure 7A ) . In summary , we find that reducing the level of regulatory interactions in the developmental plan produces two competing effects . The first effect is to reduce the number of visible phenotypes and increase the patchiness of the visible phenotypic space . These are the result of an increase in inactivation for a number of phenotypic elements . The second effect is an increase in the number of visible phenotypes . This is a result of the higher marginal contribution of each gene to determining the state of each associated phenotypic element . Both these effects are prominent for sparse matrices , but become negligible for density values in the range c>0 . 5 . This might suggest that the phenotypic bias generated by the fully connected matrix remains relevant for typical , empirically derived networks , for which density values below or close to 0 . 5 have been observed [23] . Finally , we consider the effects of the developmental map on phylogenetic regularities . Since we are focusing on the evolution of development bearing on phenotypic diversity and disparity , we do not consider the evolution of the structural genes , but only regulatory interactions . We assume in the following treatment that developmental plans evolve incrementally and neutrally by addition of new genetic regulatory elements into existing regulatory networks . Consider , for example , an ancestral developmental plan that possesses ra transcription factors , controlling k target genes . Descendant developmental plans acquire rb>ra transcription factors ( still controlling the same k genes ) , where all descendant plans share an identical regulatory wiring for the ancestral ra transcription factors , and differ in the wiring of the derived factors ( Figure 9 ) . Following findings in the previous section , we focus only on the most frequent phenotypes produced by each plan as evolutionarily representative of the complete , visible phenotype set . By focusing on the most frequent phenotypes , we are considering those phenotypes most likely to be observed . We are interested in the phylogenetic distribution of phenotypes generated by the evolutionary sequence of developmental plans . We observe that the phenotypes comprising a single developmental plan , become more similar throughout the evolutionary process , whereas disparity among members of different plans increases ( Figure 10A ) . This process relates to an increase in the regulatory dimension of the genome , and hence illustrates how regulatory evolution promotes increasing phyletic disparity while decreasing phenotypic disparity . To illustrate the similarities and relationships among phenotypes , specifically between current phenotypes and ancestral phenotypes , we perform a phylogenetic analysis . We follow the evolutionary process described above ( see also Figure 9 ) , starting with an ancestral group that embodies a developmental plan with r = 4 and k = 14 . A first branching event results in two intermediate groups , each with r = 9 and k = 14 . A second branching event results in four groups , each with r = 14 and k = 14 . We consider a collection of phenotypes comprising the most frequent visible phenotypes in the most derived groups , the intermediate groups , and the ancestral group , and reconstruct a phylogenetic tree relating these phenotypes ( Figure 10B ) . This tree is exact as we preserve the complete evolutionary history of each lineage . The resulting tree not only clusters the derived groups correctly , but also demonstrates that intermediate and ancestral groups span the same phenotypic space as their descendants . Note , in particular , that phenotypes in the ancestral group cover ( though , more sparsely ) most of the space covered by the derived groups . A similar pattern can be observed by means of a principal components analysis of the phenotypic set ( Figure 10C ) .
The distribution of phenotypic degeneracy levels recalls results from the genotype/phenotype map induced by RNA secondary structure [20] , where it has been shown that frequencies of planar structures are highly nonuniform ( following a generalized form of Zipf's law ) resulting in few common structures and many rare ones . There are two important differences between simple genotype/phenotype maps and our results . First , whereas the RNA genotype/phenotype map is the outcome of physical interactions between base pairs , the mapping presented in this paper is the result of a developmental scheme , representing interactions among multiple transcripts . Second , for RNA secondary structure , the space of potential shapes is considerably smaller than the sequence space . RNA studies focus on the distribution of visible phenotypes and on the organization of the visible phenotypic neutral networks . We consider the size and structure of the space not covered by neutral networks . The molecular study of developmental maps in multicellular lineages has tended to focus on changes over a small number of generations , typified by studies of homeotic mutants . Paleontologists have become interested in the macroevolutionary implications of developmental evolution , in particular , the production of features associated with higher taxonomic levels [11] . The benchmark example of what we might call ‘developmental macroevolution’ is the Cambrian radiation associated with a rapid proliferation of highly disparate , multicellular animals [12] . The putative causes of this radiation include the accumulation of atmospheric oxygen [10] , a snowball earth scenario [29] , as well as a variety of putative developmental innovations including the emergence of Hox cluster of genes [6] , and the co-opting of regulatory networks for new structures and functions [30] . Whatever factors might have lead to the original ‘explosion’ of varieties , we are able to show with a suitable model for development , that simple , low dimensional ancestral regulatory networks will tend to produce a higher disparity among the set of most frequent phenotypes than is the case for , derived , high-dimensional networks . This is because the ancestral programs are less constrained by regulatory epistasis . Moreover , developmental evolution generates anisotropic phenotypic variation , towards an increasingly clustered occupancy of phenotypic subspaces . These results agree with prior studies showing a tendency towards a clustering of phenotypes and a deceleration of diversification in abstract morphospaces that arise through branching random walks [13] at levels above individuals , or through random rates of speciation and extinction imposed on a background rate of discrete anagenesis [31] . It has been suggested that developmental plans constitute a mechanical explanation and justification for phylogenetic grades [7] . These results support this hypothesis , as each developmental plan represents a conserved core responsible for imposing a shared pattern of expression on a lineage of organisms . Critically , these organisms can share the bulk of their genes and yet remains significantly different when these genes are expressed through their unique developmental programs . It remains to be determined why these programs remain relatively uniform through time . One possibility is that changes to these programs are more deleterious than changes to the non-regulatory quotient of the genome [7] . Another possibility , is that since selection acts only indirectly on the genetic program but directly on the traits that it generates , the selective pressure on the plan is weak , and when coupled to the canalizing effects of the plan , severely decelerates the evolutionary process . In an important sense , it is this property of variation in structural genes compared to invariance of the developmental plan that allows for the emergence of high level grades . If this constraint is relaxed , phenotypes are more uniformly distributed , making the concept of , for example , phyla an arbitrarily placed epiphenomenon of phylogenetic trees . The role of development in generating , or constraining , biotic diversity has been one of the most active debates in evolutionary biology [32]–[34] . The roots of this debate go back to the study of homologies and questions over physico-chemical verses genetically-selected rules of growth . One merit of simple developmental models is to illustrate how these two positions reflect necessary , complementary properties of generic developmental programs . Regulatory epistasis introduces non-linearities into development , allowing similar genotypes to generate significant divergence among phenotypes , whereas degeneracy tends to contract the occupancy of morphospace and bias phenotypic samples . Of great interest is how these structural properties of development have themselves been modified over the course of evolutionary time , potentially changing the tempo and mode of the evolutionary process . One of the paradoxical implications of this study has been to show how innovations in development ( arising through increasing regulatory dimensions ) that lead to an increase in the volume of accessible phenotypes , can lead to a reduction in selective variance ( through increasing regulatory epistasis ) , so whereas the potential for novel phenotypes increases , the fraction of space these phenotypes occupies tends to contract . Hence the evolutionary process moves from a macro-configuration , sampling distant regions of space sparsely , to a micro configuration , sampling local regions of space at high resolution . This is analogous to an annealing process , whereby as an optimization process proceeds , the solutions become more frequent and more densely localized around the putative solution points .
|
At the very end of his On the Origin of Species , Charles Darwin wrote , “from so simple a beginning endless forms most beautiful and most wonderful have been , and are being , evolved . ” Nature truly displays a bewildering variety of shapes and forms . Yet , with all its magnificence , this diversity still represents only a tiny fraction of the endless “space” of possibilities; research on the evolution of development has revealed that observed common morphologies and body plans ( or , more generally , phenotypes ) occupy only small , dense patches in the abstract phenotypic space . In this paper , we introduce a simple model of evolving gene regulation and show that these empirically identified patterns can be attributed , at least in part , to interaction between genes ( epistasis ) in the developmental network . Our model further predicts that early developmental programs with low levels of interaction would span most of the variation found in extant species . The theory presented in our paper complements the view of development as a key component in the production of endless forms and highlights the crucial role of development in constraining ( as well as generating ) biotic diversity .
|
[
"Abstract",
"Introduction",
"Models",
"Results",
"Discussion"
] |
[
"computational",
"biology/transcriptional",
"regulation",
"computational",
"biology/evolutionary",
"modeling",
"evolutionary",
"biology/developmental",
"evolution"
] |
2008
|
An End to Endless Forms: Epistasis, Phenotype Distribution Bias, and
Nonuniform Evolution
|
A hallmark of the G1/S transition in budding yeast cell cycle is the proteolytic degradation of the B-type cyclin-Cdk stoichiometric inhibitor Sic1 . Deleting SIC1 or altering Sic1 degradation dynamics increases genomic instability . Certain key facts about the parts of the G1/S circuitry are established: phosphorylation of Sic1 on multiple sites is necessary for its destruction , and both the upstream kinase Cln1/2-Cdk1 and the downstream kinase Clb5/6-Cdk1 can phosphorylate Sic1 in vitro with varied specificity , cooperativity , and processivity . However , how the system works as a whole is still controversial due to discrepancies between in vitro , in vivo , and theoretical studies . Here , by monitoring Sic1 destruction in real time in individual cells under various perturbations to the system , we provide a clear picture of how the circuitry functions as a switch in vivo . We show that Cln1/2-Cdk1 sets the proper timing of Sic1 destruction , but does not contribute to its destruction speed; thus , it acts only as a trigger . Sic1's inhibition target Clb5/6-Cdk1 controls the speed of Sic1 destruction through a double-negative feedback loop , ensuring a robust all-or-none transition for Clb5/6-Cdk1 activity . Furthermore , we demonstrate that the degradation of a single-phosphosite mutant of Sic1 is rapid and switch-like , just as the wild-type form . Our mathematical model confirms our understanding of the circuit and demonstrates that the substrate sharing between the two kinases is not a redundancy but a part of the design to overcome the trade-off between the timing and sharpness of Sic1 degradation . Our study provides direct mechanistic insight into the design features underlying the yeast G1/S switch .
In the cell cycle of the budding yeast Saccharomyces cerevisiae , DNA replication initiation is driven by a sharp rise of Clb5/6-Cdk1 activity [1]–[3] . As the cell passes Start [4] , [5] , the commitment to the next round of cell division , the transcriptional inhibitor Whi5 , is phosphorylated and then excluded from the nucleus [6] , [7] , and two transcription factor complexes Swi4/Swi6 ( SBF ) [8] and Mbp1/Swi6 ( MBF ) [9] are activated . SBF and MBF transcribe about 200 G1/S genes including the G1 cyclins ( CLN1 and CLN2 ) and the S cyclins ( CLB5 and CLB6 ) [10] , [11] . Unlike Cln1/2-Cdk1 , which is the major driver of cell cycle progression in the late G1 phase by phosphorylating many G1/S targets [12] , Clb5/6-Cdk1 is rendered inactive throughout G1 phase by the inhibitor Sic1 until Sic1 is phosphorylated and degraded ( Figure 1 ) [13] , [14] . The timing and the speed of Sic1 destruction determine the activation profile of Clb5/6-Cdk1 activity . Strains with either SIC1 deleted or altered Sic1 degradation dynamics show a significant increase in genomic instability ( Figure S4 ) , underlining the importance of Sic1 for a proper S-phase entry [15] , [16] . Sic1 has nine consensus CDK phosphorylation sites , and phosphorylation on multiple sites is required for the SCF ubiquitin ligase F-box protein Cdc4 to efficiently recognize Sic1 , thereby targeting it for degradation by the 26S proteasome [14] , [15] , [17] . Both Cln2-Cdk1 and Clb5-Cdk1 can phosphorylate Sic1 efficiently in vitro [18] , [19] , and a recent biochemical study suggested the multisite phosphorylation is carried out in a processive cascade fashion [18] . Given the in vitro observation that Cln2-Cdk1 and Clb5-Cdk1 have different specificity on different phosphorylation sites of Sic1 , it was proposed that Sic1 is first phosphorylated by Cln2-Cdk1 on a certain site ( priming site ) , facilitating its subsequent phosphorylation by Clb5-Cdk1 , eventually leading to its degradation after phosphorylation of the phosphodegrons [18] . However , this interpretation is clouded by extensive circumstantial in vivo evidence . Strains lacking CLN1 and CLN2 are highly sensitive to SIC1 gene dosage [20] , the DNA replication delay in strains lacking CLN1 and CLN2 is Sic1 dependent [21] , premature Clb5 expression from the GAL1 promoter does not advance the onset of S phase [2] , and deleting CLB5 and CLB6 does not affect the normal timing of Sic1 turnover [14] , [22] . All in vivo observations seem to suggest that it is Cln1/2-Cdk1 , but not Clb5/6-Cdk1 , who is responsible for the physiological Sic1 destruction . It was even thought that the only nonredundant essential function of the Cln1/2-Cdk1 is to inactivate Sic1 [23] . To date , the only supporting evidence for Clb5/6-Cdk1's contribution to Sic1 destruction in vivo is that Sic1 is stable with B-type cyclin inhibition [18] . However , the same experiment performed by another group earlier reached the opposite conclusion [14] . Therefore , whether or not Clb5/6-Cdk1 phosphorylates Sic1 in vivo remains a mystery . In addition , previous studies suggested that the requirement of multisite phosphorylation sets a threshold for CDK activity for the onset of Sic1 degradation , and thus is responsible for a rapid , switch-like destruction of Sic1 [15] , [24] . However , a theoretical study showed that while multisite phosphorylation can establish a threshold for kinase activity , the response beyond the threshold is not necessarily switch-like [25] . The same study reported that a switch-like response requires disparities of several orders of magnitude in catalytic efficiencies at different phosphorylation sites . This does not seem to be the case for Sic1 [18] . In this article , by monitoring Sic1 destruction dynamics directly in single cells and in real time under a variety of systemic and environmental perturbations , we resolve the discrepancies mentioned above and provide a clear picture of the G1/S transition in yeast . We investigate the role various components in the G1/S regulatory circuitry play in Sic1 destruction dynamics and dissect the G1/S switch at the core of this circuitry to reveal its underlying design principles .
To quantitatively monitor Sic1 degradation dynamics in individual cells and in real time , we tagged the endogenous SIC1 at the C terminus with a green fluorescent protein ( GFP ) and used live-cell fluorescence microscopy ( Figure 2B ) . We verified that Sic1-GFP half-life is the same as the endogenous Sic1 ( Figure S1A–C ) . For each cell , we were able to measure the concentration of Sic1 as a function of time and thereby obtained the half-life of Sic1 by fitting our data to an exponential decay function ( Figure 2C , Text S1 ) . We observed considerable cell-to-cell variability in Sic1 half-life ( Figure 2F ) , though there is no systematic difference between mother and daughter cells ( Figure 2D , E , Figure S1F , Figure S2F–H , and Table S1 ) . We first deleted one by one each of the components in the circuitry that precedes Clb5/6-Cdk1 activation ( blue box in Figure 2A ) . Perturbations on different components led to the various familiar phenotypes: For example , swi4Δ and cln1Δcln2Δ prolonged the cell cycle , whi5Δ had a smaller cell size , and cln1Δcln2Δ delayed budding . However , none of these has a statistically significant effect on Sic1 half-life ( Figure 2F; Tables S1 and S2 ) . Surprisingly , even deletion of both CLN1 and CLN2 had no effect on the speed of Sic1 destruction ( Figure 2D–F; Tables S1 and S2 ) . These findings suggest that none of these components , including Cln1/2 , contributes significantly to the speed and variability of Sic1 destruction . This is in stark contrast with the current model suggested by earlier studies , in which only Cln2-Cdk1 is responsible for switch-like destruction of Sic1 [15] , [24] . We next deleted CLB5 and CLB6 . In this case , large effects were observed on both the median and the variability of Sic1 half-life ( Figure 2D–F; Tables S1 and S2 ) . The median value increased to τclb5Δ = 6 . 35 min and τclb5Δclb6Δ = 7 . 14 min in clb5Δ and clb5Δclb6Δ strains , respectively , compared to τWT = 3 . 93 min . The variability also increased significantly . These results suggest that Clb5/6-Cdk1 plays a critical role in controlling the speed of Sic1 destruction . The effects of the various gene deletions on Sic1 degradation dynamics were quantitatively captured in a stochastic model of the entire G1/S circuitry ( Text S2 , Figure 2G , Figure S1G ) . An important point worth noting is that the slow degradation we observed in clb5Δclb6Δ cells suggests that Cln1/2-Cdk1 mediated phosphorylation does not lead to Sic1's fast destruction . This finding argues against the current model , in which Cln1/2-Cdk1 is solely responsible for rapid destruction of Sic1 [15] . The result above is in line with the in vitro study that showed that Clb5-Cdk1 is more potent than Cln2-Cdk1 on more phosphorylation sites of Sic1 [18] . Furthermore , it provides direct evidence that a double-negative feedback loop between Clb5/6-Cdk1 and Sic1 is in action ( Figure 1 and red box in Figure 2A ) . Positive feedback loops are capable of generating sharp transitions and are widely implemented in cell fate circuitries [26]–[30] . The observed fast degradation of Sic1 attributed to Clb5/6-Cdk1 can either be due to a higher potency of Clb5/6-Cdk1 on Sic1 phosphorylation or the double-negative feedback loop , or both . Thereby , we next sought to disentangle the contribution of the feedback loop from that of the kinase on Sic1 destruction . To investigate the function of the feedback loop , we constructed a Sic1 reporter , designated Sic1* , by fusing the regulatory domain of Sic1 ( including all nine CDK phosphorylation sites ) [31] to a fluorescent protein ( mCherry ) and placing it under the control of the ADH1 constitutive promoter ( Figure 3A and 3B ) . We verified that Sic1* has the same subcellular localization and the same degradation dynamics as the endogenous Sic1 ( Figure 3C , Figure S2A ) and that Sic1* does not inhibit Clb5/6-Cdk1 ( Figure S2B ) . Thus Sic1* can serve as a reporter of Sic1 destruction dynamics , but due to its lacking of the CDK binding domain , it cannot inhibit Clb5/6-Cdk1 . With the reporter Sic1* , we were able to study the role of the other components in the feedback loop on Sic1 destruction dynamics ( Figure 3A and D–F ) . Interestingly , unlike the deletion of CLB5/6 ( Figure 2F ) , disabling other components of the double-negative feedback loop resulted in minimal impact on Sic1 degradation speed ( compare pairs of data in black and colored dots highlighted by the small rectangles in Figure 3D–F ) , suggesting that it is Clb5/6-Cdk1 , not the feedback loop , that is responsible for Sic1's fast destruction . However , when the cell is subject to genetic or environmental perturbations and stress , we observed a very different behavior: the absence of the loop resulted in large variations in the degradation speed , suggesting that the feedback loop is necessary for robustly fast Sic1 destruction under perturbations ( Figure 3D–F , Tables S3 , S4 , S5 , S6 , S7 ) . These genetic and environmental perturbations led to noisy transcriptional activities in CLN2 and CLB5 ( Figure 3G and 3H , Table S8 ) , and some perturbations increased the cell cycle period more than 2-fold ( Figure S2C ) . ( Notably , in the absence of the loop , the degree of the variability in Sic1 half-life reflects the degree of variability in transcriptional activity of CLN2 and CLB5 ( Figures 3D–H ) . ) Remarkably , the double-negative feedback loop can buffer the extrinsic noise ( cell-to-cell variability ) and fluctuations to ensure a robust fast destruction of Sic1 . So far we have established the important role of Clb5/6-Cdk1 and the double-negative feedback loop in Sic1 destruction . On the other hand , the upstream kinase Cln1/2-Cdk1 can phosphorylate Sic1 in vitro [14] and , presumably , also contributes to Sic1 degradation in vivo as suggested both by previous work [20] , [21] and by our experiment with the clb5Δclb6Δ strain . We have shown that Cln1/2-Cdk1 is not essential for Sic1's fast destruction . What is then the in vivo role , if any , of Cln1/2-Cdk1 here ? It is known that in cln2Δ strains the onset of the S phase is significantly delayed [21] , and recent biochemical study found that Cln2-Cdk1 has a preferred phosphorylation site on Sic1 that could function as a priming site [18] . To investigate the role of Cln1/2-Cdk1 on Sic1 destruction , we further tagged the transcriptional repressor Whi5 with a red fluorescent protein ( mCherry FP ) , which enabled us to measure the time interval between the Start and the S phase entry . The exclusion of Whi5 from the nucleus marks the point of Start transition at which the transcriptional activity of SBF and MBF is turned on [29] . We define “timing” of Sic1 destruction as the time from Whi5 nuclear exclusion to the point of the fastest Sic1 degradation ( Figures 1 and 4B , Figure S3B ) . Proper timing of Sic1 destruction is essential for a proper S phase entry ( Figure S4 ) . We found that the timing of Sic1 destruction was not significantly affected in clb5Δclb6Δ strains . In contrast , it was much more variable in cln1Δcln2Δ strain ( Figure 4D and 4E ) . The median timing was shorter in cln1Δcln2Δ strain , presumably because Clb5/6-Cdk1 significantly contributed to Whi5 nuclear exclusion when Cln1/2-Cdk1 was absent ( Figure S3C ) . Considering that the dynamics of Whi5 nuclear exclusion is highly affected in cln1Δcln2Δ strain ( due to the absence of the first feedback loop; Figure 4A ) , rendering our Start reference point unreliable , we further introduced a Sic1 variant with a mutation on the Cln1/2-Cdk1 binding site ( Figure 4C ) . This mutation disables Cln1/2-Cdk1–mediated phosphorylation of Sic1 [18] , [32] without deleting CLN1/2 , and thus the Whi5 dynamics is unaffected . The destruction timing of this Sic1 mutant was delayed and much more variable than the wild-type ( Figure 4E ) , while the speed of its destruction was as fast as the wild-type ( Figure 4F ) . Taken together , these observations imply that while Clb5/6-Cdk1 is critical for Sic1 fast destruction , Cln1/2-Cdk1 is responsible for setting a robust timing of Sic1 destruction . To demonstrate how population-level studies can sometimes be misleading , we averaged the Sic1 concentration profile over many individual cells ( Figure 4G ) . The plot shows an apparent more significant effect of Cln1/2-Cdk1 than Clb5/6-Cdk1 on Sic1 degradation , which is clearly due to the more variable timing of the Cln1/2 mutants . This result explains the discrepancy between our single-cell experiments and the earlier population-level studies that reported that deleting CLB5 and CLB6 does not affect the rate of Sic1 turnover ( compare the black and red lines in Figure 4G ) [14] , [22] . An important feature of this circuitry is that Sic1 has to be phosphorylated multiple times before it can be degraded [14] , [15] , [17] . A study based on Western blotting of Sic1 in synchronized cell populations suggested that the destruction of wild-type Sic1 is more switch-like than a single-phosphosite mutant , Sic1CPD [15] . Sic1CPD lacks all nine endogenous CDK phosphorylation sites of Sic1 but incorporates a single CDK site with a high-affinity Cdc4 binding motif ( LLTPP ) in place of Ser 76 as shown in Figure 5A [15] . This could imply that the requirement of multisite phosphorylation for Sic1's degradation is responsible for its switch-like destruction [15] , [24] . We placed SIC1CPD-GFP at the chromosomal SIC1 locus under the control of the endogenous SIC1 promoter and investigated its destruction dynamics in single cells . We found no obvious difference in half-life between Sic1 and Sic1CPD ( Figure 5B and 5C ) . However , compared to Sic1 , a much more variable timing of Sic1CPD destruction was observed ( Figure 5D ) . Considering that clb5Δclb6Δ mutations increase the half-life of Sic1 and that Sic1CPD mutant degrades as fast as wild-type , we conclude that Sic1's fast destruction is not a result of a threshold set for Cln1/2-Cdk1 activity by multisite phosphorylation , suggested by the current model [15] . When we averaged Sic1 and Sic1CPD concentration profiles over many individual cells , we saw that Sic1CPD does appear to degrade slower , consistent with the earlier population-level study [15] . Our result demonstrates that slow Sic1CPD degradation observed in [15] is mainly due to the large cell-to-cell timing variability of Sic1 degradation ( Figure 5E ) . To further elaborate on the design principles of the G1/S switch and to place our experimental findings on a more general footing , we also carried out a computational study . We constructed an analogous ODE model of the switch consisting of three components ( named Kinase1 , Kinase2 , and Inhibitor ) to identify the contributions of the components of the switch to the sharpness and timing of the transition and their respective robustness under cell-to-cell variability on the inputs of the switch ( Figures 6A–C ) [33] . Essentially , we incorporated extrinsic noise on Kinase1 and Kinase2 production and degradation by varying the synthesis and degradation rates from realization to realization ( Text S3 ) . We then optimized the phosphorylation rates to obtain either the sharpest or most timely Kinase2 activation ( Figures 6A–C and S5A ) . We did not incorporate noise in any other reactions and kept the Inhibitor level constant . Note that even though only extrinsic noise was incorporated in this simplified model , the results are applicable to cases where intrinsic noise is also present . ( The simplified model is different than the stochastic model of the entire G1/S circuitry , whose results are presented in Figure 2G . ) First , we compared the linear circuit ( suggested by earlier studies ) and the double-negative feedback to discern the roles of multisite phosphorylation and topology on the sharpness of the Inhibitor destruction . We optimized the catalytic efficiencies of the kinases in both circuits to obtain the sharpest Kinase2 activation , as a measure of the decisiveness of the switch ( Figure 6D ) . The double-negative feedback yielded sharper Inhibitor destruction on the average than the linear circuit , confirming that the topology is the dominant factor in determining sharpness and its robustness ( i . e . , decisiveness ) , and not multisite phosphorylation . Double-negative feedback is more effective in generating sharp Inhibitor destruction [34] because in the linear topology case , speed of degradation is directly affected by the variability in Kinase1 level . The feedback loop ensures that the Inhibitor degradation does not begin before sufficient amount of Kinase2 is accumulated; therefore , it buffers the variability . The disadvantage of the double-negative feedback circuit is that the timing is highly variable for the same reason ( last panel in Figure 6D and 6F ) . The activation occurs when Kinase2 level exceeds the Inhibitor level . The time of the activation varies highly , because Kinase2 approaches to its steady-state level asymptotically , and we assume that the initial level of the Inhibitor and the steady-state level of Kinase2 are comparable , and Kinase2 approaches to that level asymptotically as illustrated in Figure S5D . ( The assumption is due to the fact that lower levels of Inhibitor yield timing variabilities that do not match the timing variability of Cln1/2 deletion , as seen in right panel of Figure 6G and Figure S5B . ) This is a trade-off that underlies the design of the switch: increasing the Inhibitor level increases the sharpness of the Kinase2 activation and Inhibitor degradation; however , it also increases the variability of the timing ( Figure S5D and Figure 6G ) . An interesting fact about the linear circuit is that it generates robust timing , even though it was optimized for sharpness . We next used this insight to see if timing variability of the double-negative feedback loop circuit could be improved by adding an upstream trigger . We picked a double-negative feedback circuit from the sharpness-optimized set ( Figure 6D ) and attached Kinase1 to it . We then optimized this circuit for timing by only mutating the Kinase1 phosphorylation rates on Inhibitor . Indeed , addition of the trigger reduced the standard deviation of the timing distribution by more than 65% ( Figure 6E ) . This is because Kinase1 enforces timely activation of Kinase2 by initiating Inhibitor degradation ( Figure 6F ) . However , the trade-off is seen here as well—this modification reduced the sharpness of the transition and increased its variability , since Kinase2 activations occurred at a lower level of the Inhibitor on the average . This result is consistent with the Cln2-Cdk1 binding-site mutant of Sic1 that yielded slightly sharper transitions with a lower variability compared to wild-type ( Figure 4F ) . The model predictions qualitatively agree with the clb5Δclb6Δand Cln2-Cdk1 binding-site mutant experiments ( Figure S5B ) . In light of the mathematical model , it is clear that Sic1 plays two roles in the G1/S switch by setting a threshold for Clb5/6-Cdk1 activity . First , it prevents precocious activation of Clb5/6-Cdk1 . Second , it allows accumulation of a large stockpile of Clb5/6-Cdk1 prior to activation . This stockpile is responsible for the consistently fast ( i . e . , switch-like ) destruction of Sic1 , leading to a sharp rise in Clb5/6-Cdk1 activity , buffering the variability in Clb5/6 level . It is possible to drive the transition only by Clb5/6-Cdk1 [35] . In this case , Sic1 degradation begins when Clb5/6-Cdk1 exceeds the Sic1 level , given that Sic1 inhibits Clb5/6-Cdk1 tightly [34] . Our model shows that this event will occur with a large timing variability if the rate of increase in Clb5/6 slows down as it approaches to Sic1 level . Cln1/2-Cdk1 corrects the timing of the double-negative feedback loop by initiating Sic1 degradation to compensate for the asymptotic approach of Clb5/6 to its steady-state level . It seems like timing variability can also be reduced by increasing the steady-state level of Clb5/6-Cdk1 . Why nature chose to use Cln1/2-Cdk1 instead is an open question .
Given its essential role in guarding genome integrity , Sic1 destruction has been studied extensively within the last two decades . However , a clear understanding of the dynamic nature of the S phase entry did not emerge due to the discrepancies between the studies and lack of single-cell experiments . In this article , we monitored the dynamics of Sic1 destruction in real time and in single cells , and provided a dynamic picture of the G1/S transition in yeast . . Our experiments show that both Cln1/2-Cdk1 and Clb5/6-Cdk1 contribute to Sic1 destruction in vivo . The role of Cln1/2-Cdk1 is to set the proper timing of Sic1 destruction , whereas the double negative feedback loop between Sic1 and Clb5/6-Cdk1 ensures the robustness of the destruction speed . The double-negative feedback loop functions as a noise filter , and is essential for the robust S-phase entry under genetic and environmental perturbations ( Figure 7A ) . Interaction between Sic1 and Clb5/6-Cdk1 is analogous to the interaction between Rum1 and Cdc13-Cdc2 studied in [36] , [37] . At first glance , involvement of two kinases in the phosphorylation of Sic1 seems like a redundancy . Sic1-Clb5/6-Cdk1 double-negative feedback loop generates irreversible and robustly sharp Sic1 destruction [38] , and Clb5/6-Cdk1 is capable of triggering its own activation [34] , [35] . Timing of the transition ( i . e . , when Clb5/6-Cdk1 level exceeds the Sic1 level ) can be tuned by adjusting the synthesis/degradation rates of Clb5/6 . Then why does Cln1/2-Cdk1 initiate the Sic1 degradation ? A possible reason could be that Cln1/2-Cdk1 transcription initiation serves as a reference for the Start transition , which can be used to time the subsequent cell-cycle events . However , it is hard to imagine why Clb5/6 cannot serve as a reference since Cln1/2 and Clb5/6 are both on the G1/S regulon—that is , their transcription is coupled [39] . Our model suggests that Clb5/6-Cdk1 is not used to time its own activation because its asymptotic approach to steady state makes timing unreliable ( second panel in Figure 6F and Figure S5D ) . Timing variability is reduced dramatically when Sic1 destruction is initiated by Cln1/2-Cdk1 as Clb5/6 rises . Collaboration of two kinases brings robustness in timing and speed of destruction of Sic1 . Sharing substrates between CDKs is almost a signature of the cell cycle control system . Here in this case , we provide an important functional implication for this apparent redundancy . Indeed , the regulatory motif we study here , a trigger coupled to a double-negative ( or positive ) feedback loop , occurs in the regulation of virtually every cell cycle transition ( Figure 7B ) [38] . Multisite phosphorylation of Sic1 by Cln1/2-Cdk1 was thought to be responsible for the switch-like destruction of Sic1 at the S phase entry [15] , [24] . In principle , multisite phosphorylation can set a threshold for CDK activity and generate an ultrasensitive response . However , our experiments show that neither Cln1/2-Cdk1 nor a multisite phosphorylation scheme is necessary for switch-like destruction of Sic1 . First , it is Clb5/6-Cdk1-mediated phosphorylation ( but not Cln1/2-Cdk1 ) that leads to a rapid destruction of Sic1 . Second , the threshold for Clb5/6-Cdk1 activity is set by Sic1 inhibition , not by multisite phosphorylation . Third , a single , optimized phosphosite on Sic1 can generate just as a rapid destruction as the wild-type . Our results are consistent with the recent theoretical studies , which suggest that while multisite phosphorylation can effectively establish a threshold , the response beyond the threshold may not be switch-like [25] , and a sharp transition that relies on one-step “linear” ultrasensitivity alone may be prone to cellular noise [40] . We emphasize that , on wild-type Sic1 , phosphorylation of multiple sites is required for a rapid destruction ( see the docking network shown in figure 3 in [18] ) . In other words , no single site on wild-type Sic1 has a high enough affinity for Cdc4 to promote a rapid destruction . The mutant we use , Sic1CPD , has an optimized phosphosite specifically for this purpose . The result that Sic1CPD degrades as fast as the wild-type does not suggest that multisite phosphorylation is redundant . It is likely that multisite phosphorylation of Sic1 integrates Cln1/2-Cdk1 and Clb5/6-Cdk1 signals to optimize the response of the switch to its inputs . Besides the possibility of helping with the switching dynamics , it has been reported that some phosphorylation sites are important for cells to respond to stress [41] . This suggests that a benefit of having multiple phosphorylation sites is to interpret signals from other signaling pathways . The requirement for multisite phosphorylation could potentially allow Sic1 to ignore low levels of Cln-CDK activity in early G1 phase and thus , in principle , could also help with the timing . Indeed , it seems the multisite phosphorylation does help with the timing in view of the Sic1CPD case . However , given the fact that the phosphorylation site of Sic1CPD is located on S76 , which is a very Clb5-Cdk1 favored site [18] , we cannot rule out the possibility that the large timing variability observed there is simply due to the lack of Cln1/2-Cdk specificity; and a single phosphorylation site may be able to achieve robust timing with some tuning . Most previous in vivo studies of Sic1 destruction dynamics have mainly relied on Western blotting of synchronized yeast populations , which , due to large cell-to-cell variability and intrinsic noise , do not accurately report the actual Sic1 concentration as a function of time in individual cells . As demonstrated in this work and some other works as well [29] , some dynamic features of a system are easily discernible at the level of individual cells but may be hard to detect at the population level , and population average of protein dynamics sometimes can lead to erroneous conclusions . Besides Clb5 and Clb6 , there are four other B-type cyclins in budding yeast . They all can phosphorylate Sic1 efficiently in vitro . Clb1–4 do not contribute to the DNA replication onset in wild-type , suggesting that they do not play a role in Sic1 destruction under normal conditions . However , since Clb1–4 initiates DNA replication in the absence of Clb5 and Clb6 , we cannot rule out the possibility that Clb1–4 contribute to the Sic1 destruction in absence of Clb5 and Clb6 . If that is the case , the Sic1 destruction with Cln1/2-Cdk1 activity only could be even slower than we observed in clb5Δclb6Δ cells . Some previous experiments suggested that Cdh1 is phosphorylated by Clb5 in vivo [42] , which could imply that Cdh1 may be capable of forming another double negative feedback loop with Clb5/6 at G1/S transition . Therefore , we also investigated the potential function of Cdh1 in Sic1 destruction dynamics . We first monitored the Sic1* half life in cdh1Δ , and we found that it is similar to wild-type ( Figure S5 ) . Then we further perturbed the system by either deleting CLN1 and CLN2 in the cdh1Δ strain , or subjecting the cdh1Δ cells to stress ( 1 M KCl ) . In all cases , Sic1* half-life is similar to wild-type ( Figure S2D , E ) , suggesting that Cdh1 does not contribute significantly to Sic1destruction dynamics . In conclusion , our study reveals the design principles of the biochemical switch at the G1/S transition: ( 1 ) the feedback loop between Clb5/6-Cdk1 and Sic1 is responsible for generating a sharp Clb5/6-Cdk1 activation robustly , and ( 2 ) Cln1/2-Cdk1 initiates Sic1 degradation and sets the timing of the activation ( Figure 6F and Figure S5C ) . It is known that cyclin-CDKs often share their substrates in cell-cycle regulation . In this case we see that the apparent “redundancy” of Sic1 being a substrate of both Cln1/2-Cdk1 and Clb5/6-Cdk1 is an important part of the design of the G1/S switch . The regulatory motif we study here , a trigger coupled to a double-negative ( or positive ) feedback loop activated through multisite phosphorylation , occurs in the regulation of virtually every cell cycle transition ( Figure 7B ) as well as in other pathways [43] , highlighting its utility in achieving both robust timing and decisive switching .
Standard methods were used throughout . Most strains used in this study were congenic S288c , except the SIC1-0P strain from Frederick R . Cross [44] . All SIC1 mutants were introduced at the chromosomal SIC1 locus under the control of the endogenous SIC1 promoter . The KanMX/NAT/LEU2 fragments , flanking with homologous sequence to the target gene ( ∼40–50 bp ) , were used to delete genes . Genotypes of deletion strains were tested by PCR . All single mutant strains were characterized by sequencing PCR products . The plasmid pCT03 ( pGREG506-ADH1pr-mCherry ) was obtained by subcloning ADH1 promoter ( starting from 713 bp upstream of the start codon of gene ADH1 , and PCR from the genome ) into the SacI-NotI site and inserting the NotI-SalI fragment containing mCherry with a 11 amino acid linker ( Ala-Ala-Ala-Gly-Asp-Gly-Ala-Gly-Leu-Ile-Asn- ) [45] at the N terminal ( PCR from pNT11 , a kind gift of Jonathan Weissman ( UCSF ) ) . Plasmids pCT04 , pCT05 , and pCT06 were constructed by inserting NotI-NotI fragments containing the regulatory domain of Sic1 ( 1–220 aa , PCR from the genome ) , MCM marker ( PCR from pML103 , a kind gift of Joachim Li ( UCSF ) [46] ) , and HTB2 ( PCR from the genome ) , respectively . The plasmid pCT07 was constructed by inserting GFP ( PCR from the plasmid pNT10 , a kind gift of Jonathan Weissman ( UCSF ) ) at SpeI-SalI , and inserting the PEST sequence of CLN2 [47] ( PCR from the genome ) at SalI of the plasmid pGREG533 . Then plasmids pCT08 , pCT09 , and pCT10 were constructed by inserting CLN2 promoter ( starting from 656 bp upstream of the start codon of the gene CLN2 , PCR from the genome ) , CLB5 promoter ( starting from 800 bp upstream of the start codon of the gene CLB5 , PCR from the genome ) , and SWI4 promoter ( starting from 1 , 068 bp upstream of the start codon of the gene SWI4 , PCR from the genome ) . All plasmids were characterized by sequencing . Cells growing exponentially in synthetic liquid medium were seeded onto thin 1 . 5%–2% agrose slabs of the same medium . Multiple different positions were followed simultaneously . Images were acquired every 1 min for Sic1 half-life experiments , 3 min for timing experiments , and 7 min for promoter experiments . For endogenous Sic1 experiments , fluorescence and phase microscopy were performed using a TE2000E Inverted Microscope ( Nikon ) with perfect Focus , Apo 60×/1 . 40 NA oil-immersion objective , a Cascade II 512 CCD ( Photometrics ) , and NIS-Elements Advanced Research software . The microscopy experiments of Sic1 reporter and promoters were performed in the University of California San Francisco Nikon Imaging Center using a TE2000E Inverted Microscope ( Nikon ) with perfect Focus , Apo 60×/1 . 20 NA water-immersion objective , a Coolsnap HQ2 Camera ( Photometrics ) , and NIS-Elements Advanced Research software ( http://nic . ucsf . edu/timelapse . html ) . Potential toxicity of fluorescence illumination was tested with pre-experiments . Image segmentation and fluorescence quantification were performed using custom Matlab software and ImageJ . Custom Matlab software was used to fit an exponential function to the fluorescence data to obtain the half-life of the endogenous and reporter Sic1 in various strains . Fluorescence signals from endogenous Sic1 or Sic1 reporter/mutants were extracted from the images . We fitted an exponential function to the signal intensity data to obtain the half-life of the protein ( Figure 2C , Text S1 ) . For Sic1 reporter , we fitted the function directly onto the mean reporter fluorescence signal from the maximum-intensity 5×5 square . For endogenous Sic1 , we first divided the mean nuclear Sic1 intensity by the mean nuclear label intensity and fitted the function onto the resulting values . The function we fitted to the fluorescence signal is ( 1 ) where y is the time series of fluorescence signal . However , since this function will cause a bias towards larger values , we also fit the signal to this form of the function: ( 2 ) To fit using Eq . 2 , we first found the constant C that gives the straightest line in the log domain when subtracted from the data . Then we fitted a straight line to the log subtracted data and the absolute value of the slope is the degradation rate . The chemical Langevin equation [48] was used to simulate the intrinsic noise . Extrinsic noise was simulated by randomizing the model parameters around their nominal values within a certain percentage range . Sic1 half-life was measured in many realizations of the stochastic simulation . See Text S2 for details . All simulations start with low concentrations of kinases and a high concentration of Inhibitor . Kinase1 and Kinase2 have identical synthesis and degradation rates , but their profiles vary as the rates are subject to extrinsic noise ( Figures 6B , F and S5C ) . Inhibitor level does not vary . We define timing of the transition as the time when Kinase2 level exceeds that of Inhibitor , and sharpness as the slope of free Kinase2 at its half-maximum level . We start the simulation with 1 , 000 copies of the switch , each assigned a random set of Kinase1 or Kinase2 catalytic efficiencies ( i . e . , phosphorylation rates ) at each phosphosite . We run each copy and calculate a fitness score based on timing or sharpness of the transition . We eliminate half of the population with lowest scores and duplicate the survivors to make up the deficit . Lastly , we mutate the catalytic efficiencies for each phosphosite in each duplicate by a small rate . We repeat this process for 1 , 000 generations . We assume that phosphorylation is ordered—that is , site 1 is phosphorylated first , site 2 is phosphorylated second , etc . See Text S3 for details .
|
In eukaryotic organisms , genome replication starts simultaneously from many sites on the DNA , called origins of replication . In budding yeast , these origins are activated by a kinase , Clb5/6-Cdk1 . Until the start of S-phase , when the replication origins are activated , this kinase is kept inactive by an inhibitor , Sic1 , which has multiple phosphorylation sites . Sic1 phosphorylation at the onset of S-phase leads to its rapid destruction , unleashing a stockpile of Clb5/6-Cdk1 . Here , we show using live-cell fluorescent microscopy that Clb5/6-Cdk1 phosphorylation of Sic1 creates a feedback loop that functions as a switch . Our experiments reveal that the feedback loop shields Sic1 destruction from molecular fluctuations and environmental variability , ensuring that the switch flips decisively . We also demonstrate that a multisite phosphorylation scheme is not required for rapid Sic1 destruction . Sic1 can also be phosphorylated by another kinase , called Cln1/2-Cdk1 . We demonstrate that this seemingly redundant interaction is responsible for robust timing of Sic1 destruction . Our experiments and mathematical model identify the contribution of each component to the function of this biochemical circuit .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[] |
2013
|
Design Principles of the Yeast G1/S Switch
|
From computational simulations of a serotonin 2A receptor ( 5-HT2AR ) model complexed with pharmacologically and structurally diverse ligands we identify different conformational states and dynamics adopted by the receptor bound to the full agonist 5-HT , the partial agonist LSD , and the inverse agonist Ketanserin . The results from the unbiased all-atom molecular dynamics ( MD ) simulations show that the three ligands affect differently the known GPCR activation elements including the toggle switch at W6 . 48 , the changes in the ionic lock between E6 . 30 and R3 . 50 of the DRY motif in TM3 , and the dynamics of the NPxxY motif in TM7 . The computational results uncover a sequence of steps connecting these experimentally-identified elements of GPCR activation . The differences among the properties of the receptor molecule interacting with the ligands correlate with their distinct pharmacological properties . Combining these results with quantitative analysis of membrane deformation obtained with our new method ( Mondal et al , Biophysical Journal 2011 ) , we show that distinct conformational rearrangements produced by the three ligands also elicit different responses in the surrounding membrane . The differential reorganization of the receptor environment is reflected in ( i ) -the involvement of cholesterol in the activation of the 5-HT2AR , and ( ii ) -different extents and patterns of membrane deformations . These findings are discussed in the context of their likely functional consequences and a predicted mechanism of ligand-specific GPCR oligomerization .
Serotonin 2A receptors ( 5-HT2AR ) are a very well characterized family of G-protein coupled receptors ( GPCRs ) in the amine sub-class of rhodopsin-like class A GPCRs [1] , [2] . The 5-HT2ARs are targeted by chemically and pharmacologically distinct classes of ligands which include antidepressants , anxiolytics , antiemetics , antipsychotics and anti-migraine agents . Notably , some agonists exhibit hallucinogenic properties [2] , [3] that have been attributed to specific manners of activation of these receptors [4] , [5] . Even when they share key structural features , such as the indole moiety of the non-hallucinogen 5-HT and the hallucinogen LSD , the 5-HT2AR ligands have been shown to be able to bind differently to the receptor molecule , and to exhibit different pharmacological properties [2] , [6] , [7] , [8] . Understanding the relation between the different modes of binding of structurally diverse compounds in the 5-HT2AR binding site , and the pharmacological responses they elicit , has therefore been of great interest in the quest for understanding the function of the 5-HT2AR and especially its role in hallucinogenesis [5] . Important clues came from in vivo studies demonstrating that behavioral responses to different 5-HT2AR ligands correlate with distinct transcriptome fingerprints for the ligands [4] . However , while it remains unclear how ligand binding induces distinct conformational states of the 5-HT2AR , and how this can result in different pharmacological outcomes [5] , the significant variability in receptor conformations that can be induced by different ligands has recently been demonstrated for the cognate β2-adrenergic receptor [9] . Structural evidence for differential effects of the GPCR ligands in relation to receptor function should be reflected in the variability of rearrangements in the key structural elements involved in the various activation states of the receptors , e . g . , the structural motifs/functional microdomains ( SM/FMs ) [10] ( see Figure 1A ) that characterize GPCR activation [5] , [11] , [12] , [13] . Specific SM/FMs have been reported from studies of a large variety of GPCRs [10] , [14] , [15] , [16] , [17] , and their dynamic signatures include ( i ) -the flipping of the toggle switch W6 . 48 ( Trp336 , identified here by the Ballesteros-Weinstein generic numbering [18] ) in the cluster of conserved aromatic residues in TM5 and TM6 , ( ii ) -the opening/closing of the ionic lock between the DRY motif ( D3 . 49–R3 . 50–Y3 . 51 ) and E6 . 30 , involved in the movement of the intracellular ( IC ) end of TM6 away from TM3 , and ( iii ) -the dynamics of the conserved NPxxY motif at the IC end of TM7 that connects as well to H8 . These are elements of activation common to many GPCRs ( see [5] , [10] , [11] , [13] , [14] , [15] , [16] , [17] , [19] ) , and their status in the X-ray structures of various GPCRs has been evaluated [12] , [20] , [21] , [22] , [23] , [24] , [25] . It is still unclear , however , how the binding of different ligands affects these elements of GPCR activation and how they connect to the mechanisms of the ligand-driven receptor oligomerization that has been shown to be critical for GPCR function [26] , [27] , [28] , [29] , [30] , [31] . To shed new light on these central mechanistic questions from the perspective of ligand-dependent conformational states involved in the activation and oligomerization of GPCRs in their membrane environment , we performed large-scale molecular dynamics ( MD ) simulations of 5-HT2AR in complex with ligands exhibiting different pharmacological properties: the full agonist 5-HT , the partial agonist LSD , and the inverse agonist Ketanserin ( KET ) ( Figure 2A ) . The simulation results show that the three ligands affect differently the dynamics of SM/FMs monitored in the simulations ( Figure 1B ) , which achieve distinct conformations that are consistent with the pharmacological classification of these ligands . Moreover , the simulations show that the ligand-bound GPCRs produce differential responses in the lipid membrane surrounding the receptor , as reflected in the spatial pattern of the remodeling of membrane thickness . These trajectories reveal as well the modes and effects of direct receptor-cholesterol interaction . Recently we have described the development and implementation of a new method , CTMD ( Combined conTinuum and Molecular Dynamics ) , for quantitative analysis of the membrane remodeling pattern based on MD trajectories [32] . With this method we account for both the membrane remodeling energy and the energy cost of any partial ( incomplete ) alleviation of the hydrophobic mismatch by this remodeling of the membrane . From the quantitative analysis with CTMD of the simulation results for the monomeric 5-HT2AR we identified ligand-specific local membrane perturbations that can produce different patterns of 5-HT2AR oligomerization driven by hydrophobic mismatch [32] . Our results lead to the prediction that the dimerization interfaces for 5-HT2AR oligomers will be different when the receptor binds ligands with different pharmacological properties ( inverse agonist , partial agonist , or agonist ) , as suggested earlier [27] . Notably , the extent of membrane-driven oligomerization of a 5-HT2AR in the inverse agonist-bound state is predicted to be larger than in the agonist-bound state . These predictions are consistent with previous experimental findings on cognate GPCRs [27] , [28] , [31] , supporting the link we identify here between ligand-dependent conformational changes in GPCRs and differences in local membrane perturbations .
The main dynamic rearrangements observed in the simulations of the 5-HT2AR when it binds each of the ligands , are described below with reference to the SM/FMs ( Figure 1A ) identified in this family of GPCRs [5] . The sequential order of the description is determined by the order in which these changes appear in the simulation trajectories of the 5-HT2AR bound to the full agonist 5-HT ( Figure 1B ) . Comparison of results in Figure 4 with Figure 3 brings to light the differences among the dynamic mechanisms connected with the binding of the three different ligands to the 5-HT2AR , as detailed below . The nature of similarities and differences observed in the dynamics of the 5-HT2AR when it binds each of the three ligands was further evaluated with Combined Essential Dynamics ( Comb-ED , see Methods ) [42] performed on concatenated trajectories for 5-HT&LSD , 5-HT&KET , and LSD&KET , each combining the last 100 ns of the individual trajectories for the pair . The comparison of such combined trajectories by their projection along their first and second eigenvectors is shown in Figure 6A , which illustrates the differences in the conformational spaces sampled by the 5-HT2AR bound to different ligands . Clearly , along the first eigenvector , the conformational spaces sampled by the 5-HT-bound and LSD-bound receptor are seen to be more similar to each other than either one is to the space sampled by KET-bound 5-HT2AR ( note that the first and second eigenvectors are different in each plot because the concatenated trajectories differ , so that the sampled spaces shown in the plots for any one ligand-bound receptor appear at different positions ) . The comparison in Figure 6B–C shows the differences in a structural context by indicating where the largest differences occur , as monitored by the magnitudes of the projections on the first eigenvectors ( color coded from red , green to blue representing magnitudes from large , median to small , respectively ) . Also evident in this figure is the greater similarity between the dynamics of the 5-HT and LSD-bound receptors ( Figure 6B–C , top panel ) . Comb-ED analysis identifies only insignificant differences between the agonist- vs . partial agonist-bound states of the receptor , with some variations in the positioning of the juxta-membrane H8 and in TM4 ( Figure 6B–C , top panel ) . However , the structure of 5-HT2AR in complex with either 5-HT or LSD is clearly distinct from that with KET bound , as seen in Figure 6B–C where the Comb-ED detects differences in TM5–6 ( linked by IL3 ) and TM4 in the 5-HT vs . KET comparison ( middle panel ) , and LSD vs . KET ( bottom panel ) . Differences between 5-HT2AR complexes with the inverse agonist , and those with the agonists 5-HT or LSD , are apparent as well for TM1 , TM3 and H8 ( Figure 6B–C , middle and bottom panels ) . Thus , in the KET-bound receptor , Comb-ED identifies the movement of TM5 and TM6 toward TM3 at the IC end , consistent with the observed closing of the ionic lock in the inverse agonist state ( Figure 4D , F , right panel ) . Furthermore , differences are evident at the EC end of TM6 between KET- and 5-HT-induced conformations , in agreement with the different level of kink in TM6 around the P6 . 50 in the two systems ( compare Figures 3A and 4A ) . In addition , in line with the observed differences in the dynamics of NPxxY motif ( Figures 3F and 4E ) , the Comb-ED analysis in the KET-bound receptor detects the motion of H8 toward TM7 to close the angle between them , consistent with earlier studies of cognate GPCRs [22] , [37] , [43] . Based on the Comb-ED results suggesting structural differences as well in TM1 and TM4 between the states of 5-HT2AR stabilized by the three ligands ( Figure 6B–C ) , we found different levels of tilt in TM1 and TM4 in the three states of the receptor . Thus , in 5-HT , LSD , and KET trajectories TM4 forms angles of 12° , 16° and 22° , respectively , with the membrane normal z axis; TM1 tilts so that in KET-bound compared to 5-HT-bound receptors its EC end is 3 Å closer to TM7 and its IC side is 1 . 5 Å farther from TM7 . The differences in conformational changes of TM1 are consistent with the available X-ray structures of the activated GPCR , where a repositioning of the IC end of TM7 towards TM1 is reported in active β2AR [23] and opsin structures [20] , [21] . As discussed below , these tilt differences in TM1 and TM4 are reflected in the response of the membrane to the interaction with the protein , and thereby can affect the ligand-regulated oligomerization of the 5-HT2AR . The nature of the changes occurring in the transition from the “activated” 5-HT-bound state of the receptor , to the KET-bound “inactivated” state , is evidenced by the application of Comb-ED analysis to combined trajectories involving the KET-substituted simulation ( started from an equilibrated 5-HT-bound receptor ) ( Figure 5B ) . Separately , two Comb-ED analysis were performed: One comparing the last 100 ns from the KET-substituted and the original KET-bound simulations , and the other comparing the KET-substituted and the 5-HT-bound simulation . The projections along the first eigenvector of these combined trajectories ( Figure 5B ) reveal the internal consistency of the results and show that , upon KET substitution , the 5-HT2AR structure deviated from the 5-HT-stabilized conformation and became similar to that stabilized by KET in our earlier simulation , with TM4 and TM6 helices changing the most . Consistent with the results in Figure 6 , in the KET-substituted simulation the IC end of TM6 moved towards TM3 , and TM4 became tilted . In addition to Comb-ED analysis of pair-wise concatenated trajectories , we applied Comb-ED as well to all four trajectories ( 5-HT , LSD , KET , KET-substituted ) concatenated together . The results ( Figure S3 in Text S1 ) clearly show that KET-substitution transitions the receptor from the conformational states visited by 5-HT to those most visited when KET is bound in the receptor . From the results of the comparative simulations we have identified two mechanisms of membrane re-organization in response to the conformational changes associated with the dynamics of the ligand-bound receptor: ( i ) -the direct interactions of the receptor with the Cholesterol ( Chol ) constituent of the membrane , and ( ii ) -the deformation of the membrane around the GPCR , which modulates the local thickness of the bilayer and the hydrophobic mismatch that can drive oligomerization of the 5-HT2AR [32] . The distinct conformational changes in the receptor produced by the binding of the different ligands ( see above ) produce different patterns of bilayer deformations around the receptor protein in complex with the different ligands ( Figure 8 ) . This difference is a result of the tendency of the lipids to minimize the hydrophobic mismatch at various TMs , i . e . , the difference in the hydrophobic lengths presented to the membrane by the corresponding TMs in the different receptor complexes ( see detailed discussion in [32] ) . Therefore , hydrophobic thickness profiles of membranes around 5-HT2AR in the simulated complexes with 5-HT , LSD , and KET , shown in Figure 8 , reveal remarkable differences in the membrane organization around individual TMs in the three systems . For example , the membrane appears thinner around TM4 and TM6 in 5-HT ( left panel ) than in the KET simulation ( right panel ) , whereas at TM1 the bilayer is thicker in the LSD ( middle panel ) than in the complexes with 5-HT or the KET . We have developed a quantitative method ( CTMD ) , for the analysis of such membrane deformations and the significant residual exposure to unfavorable hydrophobic-hydrophilic interactions at specific TMs that results from an incomplete alleviation of the hydrophobic mismatch [32] . When applied to the 5-HT2AR complexes discussed here , residual exposure [32] was found at TM4 for all three complexes , although the values were different possibly because the TM4 tilt is different in the KET , LSD and 5-HT trajectories ( see above ) . Because the extent of the hydrophobic mismatch around the TM helices is considered to be a driving force for oligomerization [32] , , we had compared the residual exposure energies at all TMs in the simulation results for the three complexes . At TM1 it was found to be substantial only in the KET simulation , consistent with the conformational changes we observed for TM1 in different systems ( see above ) , and at TM5 it appeared to be relatively similar in all three complexes , but somewhat more pronounced in the 5-HT-bound structure; lastly , the residual exposure at TM6 is largest as well in the 5-HT trajectory , possibly due to the relatively straighter configuration of this helix in the 5-HT simulation ( Figures 3–4 ) . One possible mechanism to reduce the energy penalty for this residual exposure in the membrane-embedded receptor conformation produced by the binding of a particular ligand , is to bring together the TM domains where the residual exposure is largest . Therefore , we proposed [32] that this represents a membrane-determined energy drive for the association of the proteins in the membrane . Consequently , our data in Table 2 of [32] suggests that if the hydrophobic mismatch is the driving force for receptor oligomerization , then the contact interfaces for oligomerization of the 5-HT2AR will be different in the complexes with 5-HT , LSD , or KET . According to this mechanism , ligands will not only regulate the extent of GPCR oligomerization , but will also influence which TM domains constitute the oligomerization interface . Thus , a comparison of residual surface area values at different TMs in 5-HT , LSD , and KET simulations implicates TM1 , TM4 and TM5 as likely participants in the oligomerization interface of 5-HT2AR in complex with KET , TM4 and TM5 in the oligomerization interface of LSD-bound receptors , and TM5 ( and possibly TM6 , TM4 and TM2 as well ) as the most likely participants in the oligomerization of 5-HT-bound serotonin receptor . In addition , the results in Table 2 of [32] for the 5-HT and KET simulations imply that overall the inverse agonist KET will promote more extensive hydrophobic mismatch-driven oligomerization , since the residual surface area value summed over all TMs is about 90 Å2 higher for KET-bound 5-HT2AR than it is for 5-HT-bound receptor . This prediction is in excellent agreement with the experimental data on ligand-regulated oligomerization on β2AR [31] , where in comparison to the agonist isoproterenol , the binding of an inverse agonist was suggested to promote tighter packing on β2AR protomers and/or to result in formation of higher-order oligomeric structures . With regard to the validation of the ligand-dependent dynamic properties , it is important to note that similar residual exposure is observed in the two KET-bound simulations starting from very different initial conformations . Thus , the trend of large residual exposures at TM1 , TM4 , and TM5 of the KET system is also observed in the KET-substituted system ( Table S3 in Text S1 ) . Moreover , near the TMs where the hydrophobic mismatch is alleviated by the membrane remodeling ( e . g . , TM6 ) , the membrane has similar thickness in both the KET and KET-substituted system ( Figure S6 in Text S1 ) .
The MD simulations of the 5-HT- , LSD- and KET-bound 5-HT2AR reported here provide the first molecular representation of the different effects that pharmacologically distinct ligands can have on the 5-HT2AR . The concepts of “functional selectivity” [49] , [50] and “receptor bias” [51] are frequently being used to explain the increasingly common observation of differential responses elicited by different ligands from the same receptor ( e . g . , for 5-HT2AR see [4] , [52] ) . However , no structural context had been identified for the distinct effects on the dynamics produced in the same GPCR by the binding of pharmacologically different ligands . Here we simulated the dynamics of the 5-HT2AR binding of such pharmacologically distinct ligands , and identified different effects on the SM/FMs of the receptor . These effects were shown to lead to different rearrangements that correspond to the different levels of activation known to be produced by these ligands . Notably , the differential effects were shown to be consonant with the pharmacological characterization of the three ligands as a full , partial and inverse agonist , respectively . To our knowledge , such inferences were obtained for the first time here from unbiased atomic MD simulations , but they are in line with the increasingly detailed experimental evidence about ligand-related functional selectivity [49] , [50] , [51] , [53] , [54] , [55] , [56] , [57] , [58] , [59] , [60] , [61] , [62] , [63] , [64] , [65] , [66] , with the proposals of ligand-selective conformations in the 5-HT2AR [67] and the D2R [68] , and with structural data indicating that GPCRs such as β2AR are stabilized in distinct conformational states by inverse , partial , or full agonists - respectively [12] , [13] . In the current simulations , structural changes associated with SM/FM characteristics of an “activated state” of the 5-HT2AR appear in sub-microsecond trajectories . In contrast , experimentally determined GPCR activation timescales generally vary from microseconds ( photoactivation of rhodopsin [69] ) to seconds ( β2AR in living cells [70] ) . We emphasize that the conclusions reached here do not require the simulations to have converged to an “active state structure” of the kind claimed for the constructs determined crystallographically . Indeed , a number of modes of activation proposed from experiment share this characteristic and can also be significantly faster [71] , [72] , [73] . But in general , there are many reasons for the time scale differences between our results and functional measurements . In particular , the simulated system is an idealized construct in that all interaction components are placed in optimal positions to be at or near their targets . Titratable groups are also assigned their final charge states , e . g . , when the D3 . 49 and E6 . 30 are in the protonated form in some of the constructs . Interestingly , the specific protonation form does not determine whether the ionic lock is formed or not ( see Figures 3–4 , and Figure S5 in Text S1 ) ; rather , the determinant factor is seen from our results to be the nature of the dynamics induced by the binding of a specific ligand . But considering that inactive GPCR ( β2AR ) may be pre-coupled to G-protein Gs [31] and the protonation of E3 . 49 in rhodopsin ( an activation step ) depends on transducin [74] , the degree of precoupling will likely play a role in the activation time . Moreover , the simulation conditions ( such as pH , salt , lipid composition , and crowding ) certainly do not mimic completely those surrounding the receptor in living cells ( e . g . , it is known that the highly flexible DHA chain of SDPC , included in the lipid mixture used here , facilitates GPCR activation [75] ) , and similar time-scale differences have been observed between computer simulations and experiments for other GPCRs [76] , [77] . The response of the membrane environment to the different ligand-induced structural re-arrangements produces a reorganization of the membrane around the receptor . This is reflected in ( i ) -the involvement of Chol in direct interactions with the protein [43] , [78] , that was shown here to affect the dynamics of the SM/FMs , and ( ii ) -the membrane deformations around the TM bundle of a GPCR [48] , [79] , described here with the use of the CTMD method [32] . Because the different ligand-determined conformational changes in 5-HT2AR establish different patterns of local perturbations in membrane structure around the receptor complex , they were suggested to promote different ligand-dependent receptor oligomerization patterns through the hydrophobic mismatch between the TMs and the surrounding membrane [32] . This is supported by observations in the literature that: ( i ) - oligomeric associations of the dopamine D2R [27] , 5-HT2CR [28] , and the β2AR [31] is ligand-sensitive; and ( ii ) - GPCR self-assembly is regulated by the mismatch between the hydrophobic length of the TM segments of GPCRs and the hydrophobic thickness of the lipid bilayer , as suggested by both experimental results [80] and computational studies for rhodopsin [32] , [48] , [79] . Along these lines , the results presented here suggest that the dimerization interfaces of 5-HT2AR oligomers will be different for inverse agonist- , partial agonist- , or agonist-bound complexes , and moreover that the inverse agonist KET would promote more extensive 5-HT2AR oligomerization than the full agonist ( 5-HT ) . We note that these experimentally testable predictions regarding possible oligomerization interfaces were obtained by analyzing monomeric GPCRs in complex with different ligands , without the need to simulate the dimers or higher oligomers .
Several model systems of the serotonin 5-HT2AR receptor were studied with all-atom MD simulations in explicit models of the hydrated lipid membrane environment . The 5-HT2AR was simulated in complex with three ligands known to exhibit different pharmacological efficacies: the full agonist 5-HT , the partial agonist LSD , and the inverse agonist KET ( Figure 2A ) . The parameters for 5-HT were taken from an earlier study [7] . For LSD and KET , the results of geometry optimization and electrostatic potentials obtained from quantum mechanical calculations with the Gaussian program ( Gaussian , Inc . , Wallingford , CT ) were used as input to the Restrained-ElectroStatic-Potential fit method [102] implemented in Antechamber [103] to derive charges . CHARMM topology and parameter files were then prepared with Antechamber using Restrained-ElectroStatic-Potential charges and GAFF force field . Force field parameter files for 5-HT , LSD and KET are included in Text S1 . For protein , PALM , lipids etc . , the all-atom CHARMM27 force field with CMAP corrections [100] was utilized ( this approach is similar to a procedure used successfully in previous studies [104] , [105] ) . All MD simulations were performed with the NAnoscale Molecular Dynamics ( NAMD ) suite [106] . As established in similar studies in the lab ( e . g . , see [107] ) , the simulations were conducted under constant temperature and pressure conditions with anisotropic pressure coupling , and utilized PME for long-range electrostatics [108] . The Nose-Hoover Langevin piston method [106] was used to control the target pressure with the LangevinPistonPeriod set to 100 fs and LangevinPistonDecay set to 50 fs . All MD simulations were performed with rigidBonds allowing 2 fs time step . All the simulated systems were equilibrated following a procedure described recently [109] . According to this protocol , the 5-HT2AR backbones and the heavy atoms of the ligands were initially fixed and then harmonically constrained , and water was prevented from penetrating the protein-lipid interface . Constraints were released gradually in four 300 ps-step MD simulations with decreasing force constants of 1 , 0 . 5 , 0 . 1 and 0 . 01 kcal/ ( mol·Å2 ) , respectively . Following this equilibration phase , all three GPCR-membrane complexes were simulated for 350 ns . The stability of the simulated complexes was monitored from the backbone RMSDs of the TMs in 5-HT2AR using the following definitions for TMs: L1 . 29–L1 . 59 , A2 . 38–Y2 . 67 , L3 . 24–N3 . 56 , S4 . 38–V4 . 62 , D5 . 35–K5 . 67 , N6 . 29–I6 . 60 , G7 . 32–F7 . 56 and K7 . 58–I7 . 68 . As illustrated in Figure 1E , after initial equilibration , the RMSDs in all the three systems were stable and fluctuated around or below 2 Å . In all three simulations the ligands maintained key interactions with the receptor ( Figure 1B–E ) , consistent with previous experimental data [2] , [6] , [7] , [91] , [92] . To quantify the changes in protein structure produced by the simulations we used various analysis tools . Analysis of protein structural data was carried out with Ptraj in AMBER 9 [110] and other tools discussed below . To quantify helix distortion parameters in the simulations , we used the Prokink package [111] implemented in Simulaid [112] . The correlation analysis on the time-dependent data of different variables , such as helix bend angles , face-shifts , as well as Chol-protein distances , was conducted following the procedure described in [43] . Briefly , the correlation analysis was carried out on two separate sets of dynamic variables . In the first , we followed the time-sequence of m = 8 selected variables that included proline kink and face-shift angles in TM6 and TM7 , the minimum distances between the Chol at the EC end of TM6 and the residues on TM6 ( I6 . 53 , M6 . 57 , I6 . 60 , C6 . 61 ) . In the second set , m = 12 dynamic variables were selected that included proline kink and face-shift angles in TM6 and TM7 , the minimum distances between the Chol at the IC end of TM6–7 and the residues on TM6 and TM7 ( K6 . 35 , I6 . 39 , F6 . 42 , V6 . 46 , L7 . 44 , V7 . 48 , V7 . 52 , L7 . 55 , F7 . 56 ) . For each set , we first studied pair-wise correlations between different variables , and constructed the matrix of coefficients of determination , R2 ( Figure 7D of the main text ) using Spearman's rank correlation test ( see for instance Ref . [113] ) . In this method , given Nframes pairs of observations , ( xi , yi ) , first the xi and yi values separately are assigned a rank , and then the corresponding difference , di between the xi and yi ranks is found for each pair . The R2 is then defined as: ( 1 ) Because it uses rankings , Spearman's method eliminates the sensitivity of the correlation test to the function linking the pairs of values and thus is preferred over parametric tests when no a priori knowledge exists on the functional relationship between xi and yi pairs . To compare the conformational spaces of 5-HT2AR stabilized by the different ligands ( i . e . , 5-HT , LSD and KET ) , a Combined Essential Dynamics analysis [42] , [114] was performed on Cα atoms of the protein using Gromacs 3 . 3 [115] . Essential dynamics analysis separates the configurational space into an essential subspace with a few degrees of freedom which describe overall motions of the protein that are likely to be relevant to its function , and a physically constrained subspace describing local fluctuations . The method is based on the diagonalization of the covariance matrix of atomic fluctuations defined by: ( 2 ) where xi are the three Cartesian coordinates of the carbon atoms Cα of the molecule under study , and the angular brackets denote averages over an ensemble of configurations and over the simulation time . The diagonalization of Eq . ( 3 ) yields eigenvectors that describe the directions of correlated positional changes in the molecule , whereas the eigenvalues indicate the total mean square fluctuation along these directions . In the Comb-ED , the covariance matrix is calculated for two or more concatenated trajectories , which are fitted on the same reference structure . Given this construct , the eigenvectors resulting from Comb-ED do not represent the essential degrees of motion of the molecules , but rather reveal differences and/or similarities in the dynamical and structural characteristics of the compared simulated structures . To identify structural differences between 5-HT2AR stabilized by the three ligands , Comb-ED analysis was performed on 3 concatenated trajectories obtained by combining the trajectories for the pairs 5-HT-LSD , 5-HT-KET , and LSD-KET , each for the last 100 ns , respectively .
|
The 5-HT2A receptor for the neurotransmitter serotonin ( 5-HT ) belongs to family A ( rhodopsin-like ) G-protein coupled receptors ( GPCRs ) , one of the most important classes of membrane proteins that are targeted by an extensive and diverse collection of external stimuli . Recently we learned that different ligands targeting the same GPCR can elicit different biological responses , but the mechanisms remain unknown . We address this fundamental question for the serotonin 5-HT2A receptor , because it is known to respond to the binding of structurally diverse ligands by producing similar stimuli in the cell , and to the binding of quite similar ligands with dramatically different responses . Molecular dynamics simulations of molecular models of the serotonin 5-HT2A receptor in complex with pharmacologically distinct ligands show the dynamic rearrangements of the receptor molecule to be different for these ligands , and the nature and extents of the rearrangements reflect the known pharmacological properties of the ligands as full , partial or inverse activators of the receptor . The different rearrangements of the receptor molecule are shown to produce different rearrangements of the surrounding membrane , a remodeling of the environment that can have differential ligand-determined effects on receptor function and association in the cell's membrane .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"mechanisms",
"of",
"signal",
"transduction",
"membrane",
"proteins",
"membrane",
"structures",
"biophysics",
"simulations",
"cell",
"membrane",
"cytochemistry",
"proteins",
"biophysics",
"theory",
"biology",
"biophysics",
"molecular",
"biology",
"biochemistry",
"signal",
"transduction",
"transmembrane",
"proteins",
"molecular",
"cell",
"biology"
] |
2012
|
Ligand-Dependent Conformations and Dynamics of the Serotonin 5-HT2A Receptor Determine Its Activation and Membrane-Driven Oligomerization Properties
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.